Child Sexual Abuse: Empirical Research On Understanding and Helping Victims and Offenders

CHILD SEXUAL ABUSE: EMPIRICAL
RESEARCH ON UNDERSTANDING AND

HELPING VICTIMS AND OFFENDERS
EDITED BY : Noora Ellonen, Robert Horselenberg, Cristina Izura,

Taina Laajasalo and Nadia Marie Wager
PUBLISHED IN : Frontiers in Psychology and Frontiers in Sociology
Frontiers eBook Copyright Statement
About Frontiers
The copyright in the text of
individual articles in this eBook is the Frontiers is more than just an open-access publisher of scholarly articles: it is a
property of their respective authors
or their respective institutions or pioneering approach to the world of academia, radically improving the way scholarly
funders. The copyright in graphics research is managed. The grand vision of Frontiers is a world where all people have
and images within each article may
be subject to copyright of other an equal opportunity to seek, share and generate knowledge. Frontiers provides
parties. In both cases this is subject immediate and permanent online open access to all its publications, but this alone
to a license granted to Frontiers.
The compilation of articles
is not enough to realize our grand goals.
constituting this eBook is the
property of Frontiers.
Frontiers Journal Series
Each article within this eBook, and
the eBook itself, are published under The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access,
the most recent version of the
Creative Commons CC-BY licence.
online journals, promising a paradigm shift from the current review, selection and
The version current at the date of dissemination processes in academic publishing. All Frontiers journals are driven
publication of this eBook is
CC-BY 4.0. If the CC-BY licence is by researchers for researchers; therefore, they constitute a service to the scholarly
updated, the licence granted by community. At the same time, the Frontiers Journal Series operates on a revolutionary
Frontiers is automatically updated to
the new version. invention, the tiered publishing system, initially addressing specific communities of
When exercising any right under the scholars, and gradually climbing up to broader public understanding, thus serving
CC-BY licence, Frontiers must be
the interests of the lay society, too.
attributed as the original publisher
of the article or eBook, as
applicable.
Dedication to Quality
Authors have the responsibility of
ensuring that any graphics or other Each Frontiers article is a landmark of the highest quality, thanks to genuinely
materials which are the property of
others may be included in the
collaborative interactions between authors and review editors, who include some
CC-BY licence, but this should be of the world’s best academicians. Research must be certified by peers before entering
checked before relying on the
CC-BY licence to reproduce those
a stream of knowledge that may eventually reach the public - and shape society;
materials. Any copyright notices therefore, Frontiers only applies the most rigorous and unbiased reviews.
relating to those materials must be
complied with. Frontiers revolutionizes research publishing by freely delivering the most outstanding
Copyright and source research, evaluated with no bias from both the academic and social point of view.
acknowledgement notices may not By applying the most advanced information technologies, Frontiers is catapulting
be removed and must be displayed
in any copy, derivative work or scholarly publishing into a new generation.
partial copy which includes the
elements in question.
All copyright, and all rights therein,
What are Frontiers Research Topics?
are protected by national and
international copyright laws. The
Frontiers Research Topics are very popular trademarks of the Frontiers Journals
above represents a summary only. Series: they are collections of at least ten articles, all centered on a particular subject.
For further information please read
Frontiers’ Conditions for Website
With their unique mix of varied contributions from Original Research to Review
Use and Copyright Statement, and Articles, Frontiers Research Topics unify the most influential researchers, the latest
the applicable CC-BY licence.
key findings and historical advances in a hot research area! Find out more on how
ISSN 1664-8714
ISBN 978-2-88974-655-2
to host your own Frontiers Research Topic or contribute to one as an author by
DOI 10.3389/978-2-88974-655-2 contacting the Frontiers Editorial Office: frontiersin.org/about/contact
Frontiers in Psychology 1 March 2022 | Child Sexual Abuse: Empirical Research

CHILD SEXUAL ABUSE: EMPIRICAL
RESEARCH ON UNDERSTANDING AND
HELPING VICTIMS AND OFFENDERS
Topic Editors:
Noora Ellonen, University of Tampere, Finland
Robert Horselenberg, Maastricht University, Netherlands
Cristina Izura, Swansea University, United Kingdom
Taina Laajasalo, National Institute for Health and Welfare, Finland
Nadia Marie Wager, University of Huddersfield, United Kingdom
Citation: Ellonen, N., Horselenberg, R., Izura, C., Laajasalo, T., Wager, N. M., eds.
(2022). Child Sexual Abuse: Empirical Research on Understanding and Helping
Victims and Offenders. Lausanne: Frontiers Media SA.
doi: 10.3389/978-2-88974-655-2

Table of Contents
05 Editorial: Child Sexual Abuse: Empirical Research on Understanding and
Helping Victims and Offenders
Taina Laajasalo, Noora Ellonen, Robert Horselenberg, Cristina Izura and
Nadia Wager
08 Beliefs About Children’s Memory and Child Investigative Interviewing
Practices: A Survey in Dutch Child Protection Professionals from ‘Safe
Home’
Brenda Erens, Henry Otgaar, Lawrence Patihis and Corine de Ruiter
17 Risk Factors for Sexual Offending in Self-Referred Men With Pedophilic
Disorder: A Swedish Case-Control Study
Felix Wittström, Niklas Långström, Valdemar Landgren and Christoffer Rahm
27 “All of Me Is Completely Different”: Experiences and Consequences
Among Victims of Technology-Assisted Child Sexual Abuse
Malin Joleby, Carolina Lunde, Sara Landström and Linda S. Jonsson
42 Transgender Identity Is Associated With Bullying Involvement Among
Finnish Adolescents
Elias Heino, Noora Ellonen and Riittakerttu Kaltiala
53 Young People’s Experiences of Attending a Theater-in-Education
Program on Child Sexual Exploitation
Hannah May, Juliane A. Kloess, Kari Davies and
Catherine E. Hamilton-Giachritsis
68 Prioritizing Indecent Image Offenders: A Systematic Review and
Economic Approach to Understand the Benefits of Evidence-Based
Policing Strategies
Susan Giles and Laurence Alison
83 Characteristics and Behaviors of Anonymous Users of Dark Web Platforms
Suspected of Child Sexual Offenses
Jessica Woodhams, Juliane A. Kloess, Brendan Jose and
Catherine E. Hamilton-Giachritsis
94 Obstacles in the Process of Dealing With Child Sexual Abuse–Reports
From Survivors Interviewed by the Independent Inquiry Into Child Sexual
Abuse in Germany
Wiebke Schoon and Peer Briken
106 Child Sexual Abuse as Lifespan Trauma Within the Context of Intimate
Partner Violence: Experiences of Caribbean Women
Adele D. Jones
118 Swedish and Norwegian Police Interviewers’ Goals, Tactics, and Emotions
When Interviewing Suspects of Child Sexual Abuse
Mikaela Magnusson, Malin Joleby, Timothy J. Luke, Karl Ask and
Marthe Lefsaker Sakrisvold
131 Greater Knowledge Enhances Complainant Credibility and Increases Jury
Convictions for Child Sexual Assault
Jane Goodman-Delahunty, Natalie Martschuk, Eunro Lee and Annie Cossins

150 Non-intimate Relationships and Psychopathic Interpersonal and Affective
Deficits as Risk Factors for Criminal Career: A Comparison Between Sex
Offenders and Other Offenders
Fabio Ferretti, Andrea Pozza, Fulvio Carabellese, Adriano Schimmenti,
Gianluca Santoro, Gabriele Mandarelli, Giacomo Gualtieri,
Felice Carabellese, Roberto Catanesi and Anna Coluccia
162 Acceptance of Sexual Interest in Minors in Self-Referred Individuals Under
Treatment – An Exploratory Pilot Study
Ute Lampalzer, Safiye Tozdan, Fritjof von Franqué and Peer Briken

EDITORIAL
published: 16 February 2022
doi: 10.3389/fpsyg.2022.844639
Editorial: Child Sexual Abuse:

Empirical Research on
Understanding and Helping Victims
and Offenders
Taina Laajasalo 1,2*, Noora Ellonen 3 , Robert Horselenberg 4 , Cristina Izura 5 and
Nadia Wager 6
1
Special Services, National Institute for Health and Welfare, Helsinki, Finland, 2 Department of Psychology, University of
Helsinki, Helsinki, Finland, 3 Faculty of Social Science, University of Tampere, Tampere, Finland, 4 Department of Criminal Law
and Criminology, Faculty of Law, Maastricht University, Maastricht, Netherlands, 5 Department of Psychology, Swansea
University, Swansea, United Kingdom, 6 Department of Behavioural and Social Sciences, University of Huddersfield,
Huddersfield, United Kingdom
Keywords: child sexual abuse, editorial, prevention, victimization, offending, violence - prevention and control
Editorial on the Research Topic
Child Sexual Abuse: Empirical Research on Understanding and Helping Victims and Offenders
Child sexual abuse (CSA) is a global problem with multiple long-term adverse effects and public
health consequences. In a recent umbrella review, CSA was associated with elevated risks of various
negative long-term psychosocial, psychiatric, and physical health outcomes, and the population
attributable risk fractions for several common mental health problems were considerable (Hailes
et al., 2019). Early findings from the COVID-19 era are causing growing concerns about elevated
risks and rates of CSA (Fore, 2021). Therefore, the subject of this Research Topic is ever so
important. The aim was to create a multidisciplinary compilation of latest research findings on
issues directly or indirectly related to CSA. This aim was achieved: a total of 13 articles from several
Edited and reviewed by:
Colleen M. Berryessa, disciplines address CSA from multiple theoretical and methodological viewpoints. Prevention of
Rutgers University, United States CSA and its consequences demands taking into consideration the perspectives and needs of victims
*Correspondence:
and potential victims, perpetrators and potential perpetrators, criminal justice system, social and
Taina Laajasalo health care services as well as the context, culture, and society as a whole. The importance of this
taina.laajasalo@helsinki.fi comprehensive outlook is highlighted by the articles in this issue.
In this issue, three articles (Erens et al.; Goodman-Delahunty et al.; Magnusson et al.) bring
Specialty section: forward important topics related to criminal justice system. These are essential in getting CSA
This article was submitted to cases investigated, prosecuted and convicted, simultaneously securing the rights of the victims
Forensic and Legal Psychology, as well as the defendants. First, Erens et al. show how many professionals conducting CSA
a section of the journal assessments within the child protective services still endorse beliefs not in line with scientific
Frontiers in Psychology
evidence, and often do not use interviewing methods with empirical support. Magnusson et al.
Received: 28 December 2021 find associations between Nordic police interviewers’ self-reported goals, tactics, and emotions
Accepted: 13 January 2022 during CSA suspect interviews. For example, interviewers experiencing more negative emotions,
Published: 16 February 2022
such as anger, were more likely to employ confrontational tactics, such as focusing on obtaining
Citation: a confession or utilizing aggressive tactics like raising one’s voice. Further, Goodman-Delahunty
Laajasalo T, Ellonen N,
et al. show that educating jurors in the form of judicial direction and expert evidence from a
Horselenberg R, Izura C and Wager N
(2022) Editorial: Child Sexual Abuse:
psychologist significantly increases jurors’ CSA knowledge, which in turn enhances the credibility
Empirical Research on Understanding of the complainant and increases the conviction rate. As a whole, these articles underline the
and Helping Victims and Offenders. importance of training various officials working within the CPS and the justice system, and how
Front. Psychol. 13:844639. these efforts may aid in securing access to justice for all parties. Further, educating needs to
doi: 10.3389/fpsyg.2022.844639 extend to potential victims too: May et al. show how attending a theater-in-education program on
Frontiers in Psychology | www.frontiersin.org 5

1 February 2022 | Volume 13 | Article 844639
Laajasalo et al. Editorial: Child Sexual Abuse: Empirical Research
CSA helped participants gain new awareness and understanding of CSA can be identified. These include the prevalence of
of aspects related to child sexual exploitation and abuse. Results early sexualization of children and the social acceptance and
from a total of four focus groups tentatively indicate an attitudes contributing to sexual violence. Heino et al. show that
increase in young people’s awareness and knowledge around transgender identity and non-binary identity are associated with
CSA, including victims, perpetrators and abusive relationships, both being bullied and bullying others even when controlling for
as well as how to avoid harm and maintain safety. Future research a range of variables. As the authors note, it is known that a large
is needed to see whether training will translate into actual proportion of bullying among adolescents is of a sexual nature,
behavioral change, such as increases in reporting and disclosures and thus these findings are relevant to the prevention of CSA
of CSA. experiences. Schoon and Briken point out survivors’ obstacles
The focus of public policies and interventions on CSA in dealing with CSA, utilizing data from Independent Inquiry
perpetrators is still mostly on measures that take effect after the Into Child Sexual Abuse in Germany. Many of the experienced
offense has taken place. As an example, risk assessment measures obstacles were linked to the criminal justice system and mental
have been developed for correctional settings, limiting their health services, such as the duration of judicial proceedings or
utility with non-offending, at-risk individuals. An important lack of access to psychotherapy, providing guidance on how our
prevention strategy would be the utilization of validated risk service system can avoid re-traumatizing the victim and instead
assessment tools, addressing especially dynamic, potentially aid the victim in regaining control.
changeable risk factors among potential perpetrators. To this Digital development and internet communication
end, Wittström et al. address the risk of CSA among 55 have fundamentally changed the field of CSA. Battling
non-correctional patients with pedophilic disorder, which thus technological-assisted CSA requires better understanding
far is largely unknown. Measuring research-based dynamic of this evolving phenomenon. In their systematic review and
risk factors and self-rated risk of CSA among self-referred, economic approach, Giles and Alison underscore the large scale
help-seeking men, they found that these participants scored of the problem and its huge economic burden, which amounts
higher than controls (age-matched non-clinical control group) to billions in the UK alone. This translates to requirements
on all these domains. In addition, Lampalzer et al. assess how of research-informed prioritization by the police, where
the acceptance and integration of the sexual preference into the the offenders of highest risk downloading and sharing incident
individual self-concept is associated with pedophilia associated images of children should be identified and targeted first. Further,
urges and behaviors among help-seeking individuals. Findings Woodhams et al. posit how the use of the Dark Web for purposes
suggest that more acceptance of ones’ sexual preference might related to CSA is growing and analyze data from 53 anonymous
lead to both negative behaviors (the frequency of sexual activities suspected abusers. They described the suspects’ characteristics,
with minors) and positive behaviors (the frequency of use and their motivations for using the Dark Web, knowledge that
of legal imagery of children) in relation to treatment goals, may be of use when targeting offenders. Finally, Joleby et al.
highlighting the need for further research and individualized bring into attention the victims of technology-assisted CSA.
treatment plans. Finally, an article by Ferretti et al. compares Based on in-depth interviews of seven young women they show
sexual offenders (76% of whom were child molesters) to other how victimization has influenced their well-being profoundly,
offenders with heterogeneous criminal histories, but without sex both immediately and in the long term and how it has changed
offenses, utilizing validated risk assessment measures (HCR-20 how they perceive themselves and their relations to other people.
V3). They show the role of deficient non-intimate relationships as The road to evidence-based responses to CSA is not an
a risk factor for criminal careers of sex offenders. Sexual offenders easy one. The breadth of the Research Topics in this special
had a smaller likelihood of history of problems with violence and issue tell of the complexity surrounding CSA. The only way to
antisocial behavior than other offenders, but were more likely effectively prevent sexual violence is to adopt a comprehensive
to have problems in non-intimate relationships with family and strategy consisting of primary, secondary and tertiary preventive
friends. In addition, interpersonal and affective deficits measured measures. This is also required by international treaties and
on the Psychopathy Checklist-Revised (PCL-R) boosted this agreements such as the Council of Europe Convention on
association: when problems with non-intimate relationships were Protection of Children against Sexual Exploitation and Sexual
possibly or certainly present, these psychopathic traits increased Abuse. We hope that this special issue will contribute to the
the likelihood of being a sex offender as compared with other development and implementation of evidence-based prevention
offenders. Although questions of causality remain, these findings strategies and will encourage both researchers and practitioners
is important in terms of developing and tailoring treatment. to continue their efforts to eradicate CSA.
Regarding victims and potential victims, the heterogeneity
of individuals and their experiences must be acknowledged. AUTHOR CONTRIBUTIONS
Jones highlights the contextual specificities in child sexual abuse
regarding Caribbean women and their CSA experiences, and TL and NE wrote the editorial. TL, NE, RH,
also addresses the links between intimate partner violence and CI, and NW contributed to conception of the
violence against girls. Further, in vulnerable circumstances where editorial and approved the submission. All authors
patriarchal values, structural disadvantage and interpersonal contributed to the article and approved the
violence are commonplace, interconnected contextual risks submitted version.

Laajasalo et al. Editorial: Child Sexual Abuse: Empirical Research
REFERENCES Publisher’s Note: All claims expressed in this article are solely those of the authors
and do not necessarily represent those of their affiliated organizations, or those of
Fore, H. H. (2021). Violence against children in the time of COVID-19: the publisher, the editors and the reviewers. Any product that may be evaluated in
what we have learned, what remains unknown and the opportunities that this article, or claim that may be made by its manufacturer, is not guaranteed or
lie ahead. Child Abuse Neglect 116, 104776. doi: 10.1016/j.chiabu.2020.10
endorsed by the publisher.
4776
Hailes, H. P., Yu, R., Danese, A., and Fazel, S. (2019). Long-
term outcomes of childhood sexual abuse: an umbrella review. Copyright © 2022 Laajasalo, Ellonen, Horselenberg, Izura and Wager. This is an
Lancet Psychiatry 6, 830–839. doi: 10.1016/S2215-0366(19)30 open-access article distributed under the terms of the Creative Commons Attribution
286-X License (CC BY). The use, distribution or reproduction in other forums is permitted,
provided the original author(s) and the copyright owner(s) are credited and that the
Conflict of Interest: The authors declare that the research was conducted in the original publication in this journal is cited, in accordance with accepted academic
absence of any commercial or financial relationships that could be construed as a practice. No use, distribution or reproduction is permitted which does not comply
potential conflict of interest. with these terms.

ORIGINAL RESEARCH
published: 25 September 2020
doi: 10.3389/fpsyg.2020.546187
Beliefs About Children’s Memory and

Child Investigative Interviewing
Practices: A Survey in Dutch Child
Protection Professionals from ‘Safe
Home’
Brenda Erens 1* , Henry Otgaar 1,2 , Lawrence Patihis 3 and Corine de Ruiter 1
1
Clinical Psychological Science, Maastricht University, Maastricht, Netherlands, 2 Leuven Institute of Criminology, Catholic
University of Leuven, Leuven, Belgium, 3 University of Portsmouth, Portsmouth, United Kingdom
Knowledge of children’s memory and forensic interviewing skills are crucial in child abuse
investigations. Safe Home is the Dutch hotline where both professionals and citizens can
report concerns about child abuse or domestic violence. Professionals at Safe Home
often serve as first responders to determine the need for a child abuse investigation,
protective measures and/or further police investigation. In this study, child protection
professionals (N = 158) employed at Safe Home (i.e., behavioral scientists, medical
doctors, and social workers) completed an online survey on beliefs about memory
Edited by:
Taina Laajasalo,
functioning and forensic interviewing. In line with earlier studies, we expected to find a
National Institute for Health lack of knowledge about memory functioning among Safe Home workers. Furthermore,
and Welfare, Finland we expected limited use of forensic interviewing methods that have received empirical
Reviewed by: support. Indeed, we found many professionals endorsed beliefs not in line with current
Daniel Brooks Wright,
University of Nevada, Las Vegas, memory research, especially beliefs about repressed and recovered memories. Still,
United States high percentages of professionals also reported memory beliefs related to false memory
Mark Blades,
The University of Sheffield,
formation and suggestion that were in line with scientific evidence. Some professionals
United Kingdom reported using interviewing methods for which there is no empirical validation. Because
*Correspondence: child protection professionals are often the first to interview children about allegations
Brenda Erens of abuse, the current findings identify a need for training in child forensic interviewing,
brenda.erens@maastrichtuniversity.nl
including knowledge of human memory.
Specialty section:
Keywords: memory beliefs, child abuse investigation, child forensic interviewing, child protection, forensic
This article was submitted to interviewing
Forensic and Legal Psychology,
a section of the journal
Frontiers in Psychology INTRODUCTION
Received: 27 March 2020
Accepted: 28 August 2020 Child protection workers deal with complex cases of alleged child abuse on a daily basis. As part of
Published: 25 September 2020 their work, they interview children about alleged experiences of abuse. There are potential pitfalls
Citation: in interviewing children as previous research has documented children’s susceptibility to suggestive
Erens B, Otgaar H, Patihis L and interviewing techniques (e.g., Ceci et al., 2002). There are cases in which questionable interviewing
de Ruiter C (2020) Beliefs About techniques appear to have led to false memories of abuse in children (e.g., the McMartin case;
Children’s Memory and Child
Garven et al., 1998; see Otgaar et al., 2017, for a Dutch case). Use of evidence-informed child
Investigative Interviewing Practices: A
Survey in Dutch Child Protection
forensic interviewing methods, based on knowledge about children’s memory functioning (Cross
Professionals from ‘Safe Home’. and Hershkowitz, 2017) are therefore crucial. Professionals tasked with child abuse investigation
Front. Psychol. 11:546187. need to know which types of interview questions are best at minimizing inaccurate memories of
doi: 10.3389/fpsyg.2020.546187 events, as well as producing detailed recall (Lamb et al., 2007).

1 September 2020 | Volume 11 | Article 546187
Erens et al. Erroneous Beliefs in Child Protection Professionals
Within the Dutch child protection system, the use of in alleged child abuse cases and can start an investigation into
child interviewing protocols that are empirically validated is the child’s welfare. For a comparison between Safe Home and
limited (Kinderombudsman, 2013). Articles that criticize the child protective services in countries with similar economic and
decision making process in the child protection system regularly cultural characteristics (see Table 1).
appear in Dutch media (Huijer, 2014). For example, in their Because many child abuse allegations do not include physical
2014 annual report, the Dutch Youth Inspectorate expressed evidence (Euser et al., 2010), child investigative interviewing is
concern regarding the lack of objectivity and fact finding a crucial element in child abuse investigations. However, child
within the child protection system (Inspectie Jeugdzorg, 2014). forensic interviewing is challenging and suggestive questioning
Recently, the Dutch Minister for Legal Protection announced could result in inaccurate statements (Ceci and Bruck, 1993;
a plan of improvement regarding fact finding in the child Lamb et al., 2011; Otgaar et al., 2018). In addition, several
protection system (Dekker, 2018). The plan includes up to 20 factors can make children reluctant to disclose abuse, such as
strategies to improve child abuse investigations performed by feelings of shame and fear of the consequences after disclosure
different organizations. (McElvaney, 2015). When professionals interview children about
events they have experienced, they need to tap into episodic
Child Protection Work at Safe Home memory of the child, i.e., memory for events/episodes. In the
“Safe Home” (in Dutch: Veilig Thuis) is the Dutch hotline to be past few decades, a large body of research has uncovered ways in
called in case of concerns regarding a possible case of child abuse which interviewers can help or hinder episodic memory retrieval
or domestic violence. It was founded on January 1st, 2015. Safe (e.g., Lindsay, 1990; Melinder, 2002; Goodman and Melinder,
Home employs social workers, behavioral scientists, and medical 2007), and professionals’ beliefs about memory functioning play
doctors. Currently, there are 26 regional Safe Home organizations a role in this (Melinder et al., 2004).
in the Netherlands. These organizations provide advice on child
abuse to professionals who work with children (e.g., school Memory Beliefs of Professionals
teachers, sports coaches) and to citizens, and they investigate Holding inaccurate beliefs regarding the functioning of memory
allegations of child abuse and domestic violence. A citizen can file could impact behavior of a Safe Home professional. For
a report of possible abuse at Safe Home. During the first 6 months example, a professional who believes that certain psychological
of 2017, Safe Home received 42,090 reports of alleged child problems of a child (e.g., sleep problems) are linked to
abuse or domestic violence (Central Bureau of Statistics, 2017). repressed memories of child sexual abuse, could become
In a child abuse investigation, Safe Home gathers information confident that a child is abused, even if the interview
from a child interview and from collateral informants (mostly with the child uncovered limited factual evidence to support
professionals, such as medical doctors, police, and school staff). that conclusion. Previous research has examined memory
This information is used to conduct a so-called safety assessment beliefs among various professions and demonstrated a lack
to determine if follow-up action is needed to ensure the child’s of knowledge concerning memory functioning among mental
safety, both in the short and longer term. In case the investigation health clinicians (Gore-Felton et al., 2000), psychotherapists
reveals serious abuse that could lead to criminal prosecution, (Merckelbach and Wessel, 1998), and doctorate level clinical
Safe Home refers the case to the police as per government policy psychologists (Dammeyer et al., 1997). Recent studies revealed
(Kwakman, 2017). that many professionals hold beliefs that do not match
current findings from memory research. For example, Ost
Child Protection Work in Other Countries et al. (2017) surveyed memory beliefs in Chartered Clinical
It is important to realize that child protection services, such Psychologists, unchartered therapists (hypnotherapists), and
as Safe Home in the Netherlands, shares similarities with child undergraduate psychology students from the United Kingdom.
protection systems in other countries. In the Netherlands, Participants had to fill in the Memory Beliefs Questionnaire
Safe Home often collaborates with the police and with local (MBQ; Ost et al., 2017) containing 11 memory statements
municipalities when investigating a report of alleged child abuse that have broad scientific consensus [e.g., “It is possible for
or domestic violence. The police are responsible for the criminal an individual to develop false memories for a non-traumatic
investigation, while the municipalities provide the necessary help event” (true statement)]. Chartered Clinical Psychologists scored
or treatment to the families, provided by different mental health significantly more accurately on the memory statements than
organizations. Safe Home consults with its relevant partners in the hypnotherapists and the undergraduate psychology students.
the process of investigating and referring the case. This resembles, Among the hypnotherapists the authors found evidence for
for example, the child protection system in England (led by incorrect beliefs, such as the belief that memories from the
the Department of Education), where local authorities, clinical first year of life can be retrieved accurately in adulthood.
commission groups and the police have a joint responsibility in Some erroneous beliefs about memory appear to exist even
protecting child welfare (Child protection system in England, among well-trained and experienced clinical psychologists. For
2019). They work together with other relevant agencies to ensure example, 75% (n = 253) of the entire sample “strongly agreed”
the effectiveness of work to protect and promote child welfare, with the belief that “The mind is capable of unconsciously
such as making arrangements to identify and support children at blocking out memories of traumatic events.” The findings of
risk of harm. Although Safe Home collaborates with its partners, Ost et al. (2017) parallel the results of Patihis et al. (2013) who
it is the only organization in the Netherlands that provides advice found that a large percentage of alternative therapists (including

TABLE 1 | Comparison of child protection services in several countries.
Country Organization(s) Tasks Collaborations
Netherlands Safe Home (National hotline for child abuse and Give advice and investigate reports of alleged Council for Child Protection, police,
domestic violence) and the Council for Child Protection child abuse Public Prosecutor’s Office,
(can petition a supervision order at Court) municipalities
England Local authorities work together with clinical commission Investigating concerns and taking the National Society for the Prevention of
groups and the policea necessary steps to protect the child Cruelty to Children (NSPCC)
United States The National Children’s Alliance (NCA) is the national CACs provide a coordinated, evidence-based Law enforcement, prosecution, Child
association and accrediting body for a network of more response to children who have been abused in Protective Services, (mental) health
than 850 Child Advocacy Centre’s (CACs) represented all 50 states institutions
in different statesb
Scotland Child Protection Committees (CPCs) are responsible for Investigating concerns and taking the Within each local authority, CPCs work
multi-agency child protection policy, procedure, necessary steps to protect the child with local agencies, such as children’s
guidance and practicec social work, health services and the
police
Sweden Barnahus (children’s house)d Derived from the CACs in the USA: represents Law enforcement, prosecution, Child
a multi-professional approach to child victims of Protective Services, (mental) health
abuse with the double aim of facilitating the institutions
legal process and ensuring that the child
receives necessary support and treatment.
a Childprotection system in England. (2019, November). Retrieved from: https://learning.nspcc.org.uk/child-protection-system/england/ b Child protection system in
Scotland. (2019, July). Retrieved from: https://learning.nspcc.org.uk/child-protection-system/scotland/ c How the CAC model works. (2019). Retrieved from: https://www.
nationalchildrensalliance.org/cac-model/ d Johansson, S., Stefansen, K., Bakketeig, E., and Kaldal, A. (2017). Implementing the Nordic Barnahus model: Characteristics
and local adaptions. Collaborating Against Child Abuse, 1–31.
hypnotherapists) strongly agreed with false statements about sample consisted of social workers (n = 100), behavioral scientists
memory (for example, 82% of hypnotherapists (n = 50) agreed (n = 43) and medical doctors (n = 11).
that repressed and recovered memories can surface in therapy). Social workers who work at Safe Home have an educational
The above-mentioned studies indicate that different background in social welfare or a Bachelor-level psychology
professional groups demonstrate a lack of accurate knowledge degree. At Safe Home, their task is to investigate reports of
about memory in areas such as recovered memories. We alleged child abuse and domestic violence. Behavioral scientists
searched the PsychInfo database using combinations of the are Master-level psychologists. Behavioral scientists supervise
following keywords: child protection workers, (memory) beliefs, the social workers during their investigations. Also, behavioral
(memory) knowledge and child interviewing, which yielded scientists provide support during interviews with parents,
no results. To our knowledge, this is the first study of its kind informants, and victims. To be eligible to work as a behavioral
conducted with child protection workers. scientist at Safe Home, licensing by the youth quality register
(in Dutch: Stichting Kwaliteits Register Jeugd) is mandatory.
The Present Study Although social workers also need to be licensed, the criteria
The current study investigates memory beliefs in professionals are different and educational background is less important.
from 26 Safe Home organizations in The Netherlands: social Medical doctors perform medical examinations in cases of
workers, behavioral scientists, and medical doctors. We also child abuse reports, if deemed necessary. Medical doctors are
asked these professionals which child investigative interviewing required to have completed specialized training in investigating
method they are currently using and whether there is consensus child abuse cases.
within their organization about its application. Based on the
literature described earlier, several hypotheses were formulated. Materials
First, we expect professionals to show inaccurate knowledge In line with previous studies, we provided professionals with
regarding memory functioning. Second, we expected a lack of use memory statements, selected or slightly adapted (e.g., “Memory
of child interviewing methods that have empirical support across is not influenced by suggestion” was adapted to “Memories
all Safe Home organizations. Finally, we hypothesized that there cannot be influenced by suggestion”) from studies by Patihis
also would be little consensus on child interviewing methods et al. (2013) and Ost et al. (2017). We provided professionals
within each separate Safe Home organization. with nine statements about memory and added one statement
on the nature of questions during child investigative interviews.
We selected these statements because they cover important topics
MATERIALS AND METHODS for professionals working with child abuse cases (e.g., statements
on suggestion and memory) and having wrong beliefs could have
Participants implications for their (field) work (e.g., suggestive interviewing
Professionals (N = 154) from 26 Safe Home organizations in techniques). The selected statements have been used in several
The Netherlands completed an online survey via Qualtrics. Our earlier studies (e.g., Patihis et al., 2013; Ost et al., 2017; Houben

et al., 2019). The statements were translated into Dutch by workers, separately. The highest percentage of erroneous beliefs,
three independent individuals: one Master student in forensic that is beliefs that run counter to current memory research,
psychology, one Ph.D. student, and a professor of Forensic were observed for statements regarding repressed and recovered
Psychology chose the translation that fit best. memories. Seventy-four percent (n = 117) of the total sample
Participants had the option to either agree or disagree agreed with the statement “Repressed memories can be retrieved
with the 10 statements. All participants viewed the questions accurately by certain therapeutic techniques.” The statement
and statements in the same order. We did not provide an “Traumatic memories are often repressed because of their painful
intermediate option (I don’t know or Neither agree nor disagree), content” received 84% (n = 133) endorsement. Also, 64%
to prevent participants from answering neutrally. We asked (n = 101) of the total sample disagreed with the last statement
participants which method(s) their Safe Home organization used “During a child forensic interview, directive questions should
for conducting child investigative interviews. This was a multiple only be used when open questions have been exhausted.” High
choice question (four answer options) with an open comment percentages of answers in line with current memory research
area, in case the method the participant used was not listed were found for statement 6: “It is possible for an individual
among the four. Furthermore, we asked if consensus existed to develop false memories for non-traumatic events” (94.9%;
within their organization on applying a certain child interviewing n = 150) and statement 9: “Memories cannot be influenced by
method. We also asked whether they made use of additional suggestion” (95%; n = 150).
tools (“props”) during their child interviews, such as anatomically We also compared answers of the different professional groups
correct dolls or drawings. This was also a multiple choice question on the memory statements by means of Chi Square analyses
with an open comment section. Finally, we asked participants (Pearson χ2 -test). Table 3 provides a comparison of Group 1
whether they used a different interviewing method in cases of (behavioral scientists) and Group 2 (social workers). In The
alleged child sexual abuse compared to other types of child Netherlands, social workers do not have a Master’s degree, as
abuse (and if so, which method they employed). The statements opposed to behavioral scientists. Group 1 had more accurate
and response alternatives for memory knowledge are shown in knowledge on memory than Group 2. Specifically, there was a
Table 2, the answer choices consistent with current memory statistically significant difference (χ2 = 10.57, p = 0.001, Cramer’s
science are marked with an asterisk. V = 0.274) for the statement “Traumatic memories are often
repressed because of their painful content”.
Procedure Additionally, Table 5 shows a bivariate correlational analysis
A link to the survey in Qualtrics was distributed by e-mail. of the 10 memory statements, which was added to investigate
We e-mailed all 72 behavioral scientists from the 26 Safe whether the questionnaire items are correlated with each other.
Home organizations in The Netherlands and used the snowball As can be seen from the log odds ratios, not all items are
sampling method, asking them to distribute the survey among positively related to each other. This finding was not expected
the other professionals (social workers, medical doctors or (nine statements are supposedly related to each other, because
behavioral scientists) at their organization. A reminder was sent they cover the topic of memory beliefs; only item 10 covered
2 weeks after the first e-mail. For the medical doctors, we child interviewing).
made a request to the national forum (an online communication
platform) for physicians at Safe Home. This request contained
the same information as the e-mails sent to behavioral scientists Use of Interviewing Methods and
as well as the link to the Qualtrics survey. No reminder was Additional Tools
sent. The survey could be completed anonymously with the Our study shows that more than four out of 10 professionals
option of leaving an e-mail address to participate in a raffle (44.2%; n = 68) indicated they use the so-called Three Houses
for a 15 euro gift voucher. A completed questionnaire was method from “Signs of Safety” (Turnell and Edwards, 1997)
obtained from 158 professionals. These professionals consisted for interviewing children. Twenty-three percent (n = 35) of all
of 43 behavioral scientists (27% of the total sample), 100 social professionals indicated they do not use a standardized method for
workers (63%), and 11 medical doctors (7%) Approval for conducting child investigative interviews. Twenty-eight percent
this study was received from the Ethical Review Committee (n = 43) responded they used another method than the options
of the Faculty of Psychology and Neuroscience at Maastricht posed (Three Houses Method, self-developed questionnaire or
University (ERCPN- 181_01_07_2017). Data from the current no standardized method). In the open comment section, 21.5%
study are accessible via the Open Science Framework: https://osf. (n = 34) of these professionals indicated they used a variety of
io/rjnz7/. interviewing methods at their organization, dependent on the
child’s age and the nature and/or complexity of the reported
child abuse. Furthermore, 86.4% (n = 133) answered that there
RESULTS are differences between professionals in the use of interviewing
method(s) at their Safe Home organization, reflecting a lack of
Memory Beliefs consensus and/or standardization.
Table 3 shows the percentage of science-consistent answers to Furthermore, about one in five professionals reported they
the memory statements. This is shown for the total sample used human figure drawings (18.9%; n = 29) or Duplo
and for the behavioral scientists, medical doctors and social dolls (19.5%, n = 30) during child investigative interviews.

TABLE 2 | Memory statements – the response alternative believed to be the most correct according to current memory science is indicated by an asterisk.
Statements Response alternatives
1. Our memories are permanently stored in our brain, even though we can’t retrieve all of it. Agree – Disagree*
2. Early memories, from the first year of life, are accurately stored and retrievable. Agree – Disagree*
3. Repressed memories can be retrieved accurately by certain therapeutic techniques. Agree – Disagree*
4. The more emotion with which a memory is reported, the more likely it is to be accurate. Agree – Disagree*
5. Traumatic memories are often repressed because of their painful content. Agree – Disagree*
6. It is possible for an individual to develop false memories for non-traumatic events. Agree* – Disagree
7. Vivid memories are more likely to be accurate than vague memories. Agree – Disagree*
8. A negative memory for a childhood event is indicative of a traumatic childhood. Agree – Disagree*
9. Memories cannot be influenced by suggestion. Agree – Disagree*
10. During a child forensic interview, directive questions should only be used when open questions have been exhausted. Agree* – Disagree
Statements reproduced/adapted from Patihis et al. (2013) and Ost et al. (2017).
TABLE 3 | Percentage of behavioral scientists, medical doctors, social workers who responded correctly to10 statements about memory.
Statements Behavioral scientists Medical doctor Social workers Total (N = 154)

(n = 43) (n = 11) (n = 100)
1. Our memories are permanently stored in our 44 (n = 19) 73 (n = 8) 30 (n = 30) 37 (n = 57)

brain, even though we can’t retrieve all of it.
2. Early memories, from the first year of life, are 93 (n = 40) 100 (n = 11) 89 (n = 89) 91 (n = 140)
accurately stored and retrievable.
3. Repressed memories can be retrieved accurately 23 (n = 10) 46 (n = 5) 24 (n = 24) 25 (n = 39)
by certain therapeutic techniques.
4. The more emotion with which a memory is 98 (n = 42) 82 (n = 9) 90 (n = 90) 92 (n = 141)
reported, the more likely it is to be accurate.
5. Traumatic memories are often repressed 33 (n = 14) 9 (n = 1) 10 (n = 10) 16 (n = 25)
because of their painful content.
6. It is possible for an individual to develop false 95 (n = 41) 100 (n = 11) 92 (n = 92) 94 (n = 144)
memories for non-traumatic events.
7. Vivid memories are more likely to be accurate 65 (n = 28) 73 (n = 8) 57 (n = 57) 60 (n = 93)
than vague memories.
8. A negative memory for a childhood event is 88 (n = 38) 73 (n = 8) 85 (n = 85) 85 (n = 131)
indicative of a traumatic childhood.
9. Memories cannot be influenced by suggestion. 98 (n = 42) 100 (n = 11) 92 (n = 92) 94 (n = 145)
10. During a child forensic interview, directive 33 (n = 14) 64 (n = 7) 33 (n = 33) 35 (n = 54)
questions should only be used when open
questions have been exhausted.
Interviewing with Duplo dolls, which are similar to Playmobil Forty-five percent of the professionals indicated they use
dolls, was developed by Dutch therapist Marleen Diekmann a different interview method in cases of alleged child sexual
Schoenmaker, who used Duplo dolls during her work with child abuse compared to other types of child abuse. Methods
victims of war in foreign countries (Diekmann Schoenmaker mentioned were: psycho-sexual screening (17%; a diagnostic tool
and Van der Veer, 2003). Because she did not speak the to investigate the psychological and sexual development of the
children’s mother language, Diekmann used Duplo dolls to child) and a specific examination method for sexual abuse (9%).
facilitate communication with the children. Furthermore, 31.2% The latter method is intended to investigate “vague” signs of
(n = 48) of the professionals reported using additional tools possible statutory sexual offending. Training in the examination
other than those listed in the web survey. In the open method is offered by several private institutes in the Netherlands
comment section, over 28 different tools were mentioned, (e.g., Landelijk Opleidingscentrum Kindermishandeling; LOCK).
such as smiley lists (a set of smileys that depict different An example of a “vague” sign is a child who says “I have secrets
emotions children can use to indicate how they feel), socio with my Dad and I am not allowed to tell.” Examination methods
cards (cards with pictures of people or objects children can are designed to interview children in case there is not enough
use to indicate which people or objects play an important role evidence for the police to start an investigation. In the open
in their life), family compositions and the child’s own toys. comment section, several social workers indicated that interviews
Furthermore, 22.8% (n = 36) of the professionals indicated in cases of alleged sexual abuse are only conducted by behavioral
using multiple tools. scientists or medical doctors at their Safe Home organization.

TABLE 4 | Chi square analysis of behavioral scientists vs. social workers who responded inaccurately (%) on 10 statements about memory.
Statement Behavioral Social χ2 p-value

scientists (n = 43) workers(n = 100) (two-tailed)
1. Our memories are permanently stored in our brain, even 24 (56) 70 (70) 2.69 0.101
though we can’t retrieve all of it.
2. Early memories, from the first year of life, are accurately 3 (7) 11 (11) 0.551 0.458
stored and retrievable.
3. Repressed memories can be retrieved accurately by 33 (77) 75 (75) 0.02 0.899
certain therapeutic techniques.
4. The more emotion with which a memory is reported, the 1 (2) 9 (9) 2.10 0.148
more likely it is to be accurate.
5. Traumatic memories are often repressed because of their 29 (67) 88 (88) 10.57 0.001
painful content.
6. It is possible for an individual to develop false memories 2 (5) 6 (6) 0.12 0.728
for non-traumatic events.
7. Vivid memories are more likely to be accurate than vague 15 (35) 41 (41) 0.603 0.437
memories.
8. A negative memory for a childhood event is indicative of 5 (12) 13 (13) 0.072 0.789
a traumatic childhood.
9. Memories cannot be influenced by suggestion. 1 (2) 6 (6) 0.913 0.339
10. During a child forensic interview, directive questions 29 (67) 65 (65) 0.017 0.897
should only be used when open questions have been
exhausted.
The bold value means the p value was considered significant.
TABLE 5 | Log odds ratios for the bivariate comparisons of the 10 statements about memory.
Item 1 Item 2 Item 3 Item 4 Item 5 Item 6 Item 7 Item 8 Item 9
Item 2 4.58
Item 3 0.72 4.01
Item 4 −0.26 −2.74 −0.54
Item 5 0.73 −0.42 0.15 0.11
Item 6 0.57 0.47 −0.60 0.47 0.23
Item 7 0.84 0.14 −0.02 0.46 0.39 0.44
Item 8 0.44 1.09 0.10 0.65 0.13 −2.60 1.04
Item 9 −0.03 0.47 0.84 2.62 −1.24 −1.54 −0.68 −2.60
Item 10 −0.10 −0.95 0.09 −1.28 −0.20 0.51 0.15 −0.34 −1.14
DISCUSSION include more detailed analysis of underlying traits to explain

these correlations.
Consistent with our expectations, many professionals endorsed We found that for questions on how memory works (e.g.,
beliefs that are not in line with current scientific research “Our memories are permanently stored in our brain, even though
on memory, especially controversial beliefs on the existence we can’t retrieve all of it”), behavioral scientists responded more
of repressed and recovered memories (see also Otgaar in line with memory science than social workers. Statistically
et al., 2019). Other memory beliefs of the professionals significant differences were found for one of the 10 statements.
were largely in keeping with scientific consensus. Most This finding is in line with previous research that compared
professionals held accurate beliefs regarding the formation different professional groups on memory beliefs. For example,
of false memories and the influence of suggestion on memory. Patihis et al. (2013) found that psychology researchers were more
However, the majority of professionals disagreed with the accurate about the non-existence of repressed memories and
fact that directive questions should only be used in a child agreed more often that memory can be unreliable compared to
investigative interview when open prompt questions have been some psychology practitioner groups. Furthermore, in a recent
exhausted. This runs counter to current scientific literature, study, Akhtar et al. (2018) found that police officers and the
which emphasizes the importance of open prompts (also general public endorsed more erroneous memory beliefs, as
called “invitations”) to encourage children to provide a opposed to memory experts who endorsed more scientifically
narrative of what they have experienced (Lamb et al., 2007). supported memory beliefs.
Since bivariate comparisons of the 10 memory statements Our results suggest that for some, but not all memory beliefs,
showed that although expected, not all items were positively there is a gap between beliefs held by professionals working at
correlated to each other. Future research should therefore Safe Home in the Netherlands and the current scientific literature

on memory. In line with earlier studies on memory beliefs (e.g., have supportive evidence that it helps optimize the retrieval of
Patihis et al., 2013; Melinder and Magnussen, 2015; Ost et al., accurate and detailed memories. On the contrary, at face value,
2017) many professionals endorsed beliefs regarding repressed the instruction to draw a “dream house” could obviously elicit
and recovered memories. The fact that the majority of our sample fantasies instead of memories.
agreed that repressed memories can be retrieved accurately by Furthermore, 21% of the professionals reported that they
certain therapeutic techniques raises concern. Empirical support did not use a standardized method for conducting child
for repression in adults who were abused as children is lacking investigative interviews, 22% reported using multiple methods,
(Corelli et al., 1997; Piper et al., 2008). One potential concern and 87% indicated different professionals apply different
is that beliefs in repressed memories of child (sexual) abuse methods at their Safe Home organization. This lack of a
could bias professionals at Safe Home during the investigative standardized child forensic interviewing method concurs with
process, affecting their interpretation of children’s behaviors recent research among child protection professionals in the
or statements, and affecting their choice of questions during USA (Rivard and Compo, 2017). Rivard and Compo (2017)
interviews with children. created an online survey which was distributed among Child
As for child investigative interviewing, the majority of Advocacy Centres and other child sexual abuse investigative
professionals disagreed with the statement “During a child agencies throughout the United States. This study also showed
forensic interview, directive questions should only be used when a diversity of interviewing protocols across agencies and
open questions have been exhausted.” Valuable and accurate professionals were trained in different methods. In most
information can be obtained by using open prompts (e.g., “Tell cases, Safe Home professionals interview children before
me what happened;” Lamb et al., 2018). Directive questions the police become involved in a case, and the threshold
are focused prompts which, according to several empirical for criminal investigation is not reached in every reported
studies, actually elicit less relevant information from alleged child case. However, if such initial interviews are done incorrectly
abuse victims than open prompts (Cyr and Lamb, 2009; Lamb this could contaminate children’s memory, with obvious
et al., 2009). The fact that only a minority of the responding repercussions if the case is subsequently referred for further
professionals were aware of this, raises concern that they will police investigation. Because child interviewing methods which
use directive questions too early in their child investigative have received empirical support are lacking, current practices
interviews. Such early directive questions could in some cases raise doubts regarding the quality of statements collected from
become suggestive or leading. children by Safe Home.
As hypothesized, we found that interviewing methods that Moreover, professionals indicated that additional tools, such
received empirical support are not being used in the majority as human figure drawings, anatomically correct dolls and
of interviews with alleged victims of child abuse at Safe Home. Duplo dolls, are used during child investigative interviews. The
The method reported most frequently (44.3%; n = 70) was the use of such tools is highly controversial. To our knowledge,
Three Houses method, which to our knowledge, lacks scientific the use of Duplo dolls for child abuse investigative purposes
support as a child investigative interviewing method. The Three has not been empirically investigated. Besides, several studies
Houses method is part of “Signs of Safety” (Turnell and Edwards, have demonstrated the harmful effects of anatomically correct
1997). Signs of Safety was developed in Australia as a method for dolls on children’s memory (e.g., Bruck et al., 2000) and
developing a constructive relationship between child protection the suggestive nature of these dolls can lead to false reports
workers and family members of (substantiated) abused children about sexual abuse. For example, Bruck et al. (2000) found
that re under child protection authority. Its purpose is to bring that under certain conditions, the use of dolls increased
the child’s voice into the supervision and monitoring by child incorrect reports of inappropriate behavior, such as touching
protective services. The child is instructed to draw three houses body parts that were not actually touched. Also, studies
according to different themes: the house of worries (danger), the have found that use of human figure drawings decreased the
house of good things (safety) and the house of dreams (change). accuracy of children’s reports of events (Brown et al., 2007;
How the houses are drawn depends on the child’s characteristics Otgaar et al., 2012).
(e.g., age, cognitive abilities, creativity). There are no predefined
rules as to how every house should be drawn. Typically, the Limitations
interviewer asks the child to draw and tell everything that is Before we draw our final conclusions, a number of limitations
well and fine at home (safety), things that are not well at of the present study should be acknowledged. The current
home (danger) and which things the child wold like to be data were gathered by means of a snowball method: behavioral
improved (change). Previous research on the Three Houses scientists working at Safe Home were asked to distribute
method has focused primarily on the reported experiences of the online survey among colleagues at their organization.
professionals, parents, and children with the method (for an This snowball method may have resulted in selection bias,
overview, see Wheeler and Hogg, 2012). Research has shown in that especially professionals interested in memory and
that the Three Houses method stimulates open communication child investigative interviewing filled in the questionnaire.
with families and it provides abusive parents more insight As a consequence of the snowball method we also do not
into the experiences of their children (Westbrook, 2006). The know the exact response rate, because it is unclear how
Three Houses method is, however, not designed to assist in many colleagues of the behavioral scientists received the
fact-finding in cases of child abuse allegations, nor does it survey. Furthermore, the sample size of the medical doctors

(n = 11) was quite small compared to the behavioral scientists and questions (e.g., “Where did he touch you?”- when the child has
social workers. This could be due to the fact that for technical already disclosed being touched) or option-posing questions,
reasons no reminder e-mails could be sent to the medical where the child can choose between two different answers
doctors. In addition, our survey data on child investigative (e.g., “Did he touch you above or beneath your clothes?”). The
interviewing at Safe Home represent professionals’ self-report protocol consists of different phases, including rapport building,
about their child interviewing practice. We do not know to explaining ground rules (for example, what it means to tell the
what extent these data reflect actual practice. More specifically, truth), training in episodic memory, substantive phase (asking
these findings do not inform us about the frequency with which about the alleged events), and closing with a neutral topic.
certain interview methods and/or additional tools are used. Interviewers trained in the NICHD protocol tend to use more
However, the finding that most surveyed professionals reported open prompts and fewer suggestive questions than they did
using non-scientifically supported child interviewing methods, is before training (e.g., Yi et al., 2016).
highly relevant. Our findings are worrisome because workers at Safe Home
are the first line of professionals who interview children
about an alleged experience of abuse. Incorrect memory beliefs
CONCLUSION and deficient child interviewing methods could result in false
positives (concluding a child was abused, while actually the
Taken together, our study suggests a need for training of
child was not) or false negatives (concluding the child was not
Safe Home professionals in both memory functioning and the
abused, when the child in fact was abused). Hence, it is vital that
practice of evidence-based child investigative interviewing. In
professionals at Safe Home adopt empirically-based methods for
line with previous research (Patihis et al., 2013; Ost et al.,
interviewing children, and that they are educated on the science
2017; Patihis and Pendergrast, 2019), incorrect knowledge, on
of memory relevant to applied settings.
issues such as repressed memories for trauma, are common
among professionals working in child protection. Furthermore,
professionals do not use a child interviewing method which has
received empirical support and they use additional tools, such as
DATA AVAILABILITY STATEMENT
human figure drawings, that increase the risk of false memory The datasets generated for this study can be found in online
reports (Otgaar et al., 2012). Because Safe Home is the Dutch repositories. The names of the repository/repositories and
“first responder” organization that deals with reports of alleged accession number(s) can be found at: https://osf.io/rjnz7/.
child abuse by both citizens and professionals, its professionals
should be appropriately trained in the relevant knowledge and
skills. Most of the required knowledge and skills are not taught ETHICS STATEMENT
in these professionals’ primary educational programs (e.g., social
work, child psychology). The studies involving human participants were reviewed and
We recommend implementing evidence-based child approved by the Ethical Review Committee Psychology and
interviewing methods, for example, the National Institute of Neurocience. The patients/participants provided their written
Child Health and Development (NICHD) interview protocol, informed consent to participate in this study.
developed by Lamb et al. (2007). This interview-protocol was
developed to obtain reliable and detailed accounts from children
by the use of open prompts (“Tell me what happened”) that AUTHOR CONTRIBUTIONS
facilitate free recall from episodic memory (Olafson, 2012; La
Rooy et al., 2015). Only when these open prompts do not elicit All authors have contributed to the manuscript and provided
relevant material any longer, the interviewer can use directive approval for the publication.
REFERENCES Ceci, S. J., Gilstrap, L., and Fitneva, S. (2002). Children’s Testimony.
Child and Adolescent Psychiatry. Oxford: Blackwell Science,
Akhtar, S., Justice, L. V., Knott, L., Kibowski, F., and Conway, M. A. (2018). The 117–127.
‘common sense’ memory belief system and its implications. Int. J. Evid. Proof Central Bureau of Statistics (2017). Veilig Thuis Eerste Halfjaar 2017: Nog
22, 289–304. doi: 10.1177/1365712718784045 Steeds Geen Vergelijkbare Cijfers. [Safe Home, First Sixth Months of
Brown, D. A., Pipe, M., Lewis, C., Lamb, M. E., and Orbach, Y. (2007). Supportive 2017: Still No Comparable Numbers.]. Thapathali: Central Bureau of
or suggestive: do human figure drawings help 5- to 7-year-old children to report Statistics.
touch? J. Consult. Clin. Psychol. 75, 33–42. doi: 10.1037/0022-006X.75.1.33 Child protection system in England (2019). Available online at: https://learning.
Bruck, M., Ceci, S. J., and Francoeur, E. (2000). Children’s use of anatomically nspcc.org.uk/child-protection-system/england/
detailed dolls to report genital touching in a medical examination: development Corelli, T. B., Hoag, M. J., and Howell, R. J. (1997). Memory, repression, and sexual
and gender comparisons. J. Exp. Psychol. Appl. 6, 74–83. doi: 10.1037//0278- abuse: forensic implications for the mental health professional. J. Am. Acad.
7393.6.1.74 Psychiatry Law 25, 31–47.
Ceci, S. J., and Bruck, M. (1993). Suggestibility of the child witness: a historical Cross, T. P., and Hershkowitz, I. (2017). Psychology and child protection:
review and synthesis. Psychol. Bull. 113, 403–439. doi: 10.1037/0033-2909.113. promoting widespread improvement in practice. Psychol. Public Policy Law 23,
3.403 503–518. doi: 10.1037/law0000141

Cyr, M., and Lamb, E. M. (2009). Assessing the effectiveness of the NICHD Meeting on Cognitive Psychology and Neuropsychology, eds M. S. Korsnes, A.
investigative interview protocol when interviewing French-speaking alleged Raftopoulos, and A. Demetriou (Nicosia: Cyprus University Press), 25–31.
victims of child sexual abuse in Quebec. Child Abuse Negl. 33, 257–268. doi: Melinder, A., Goodman, G. S., Eilertsen, D. E., and Magnussen, S. (2004). Beliefs
10.1016/j.chiabu.2008.04.002 about child witnesses: a survey of professionals. Psychol. Crime Law 10, 347–
Dammeyer, M. D., Nightingale, N. N., and McCoy, M. L. (1997). Repressed 365. doi: 10.1080/10683160310001618717
memory and other controversial origins of sexual abuse allegations: beliefs Melinder, A., and Magnussen, S. (2015). Psychologists and psychiatrists serving
among psychologists and clinical social workers. Child Maltreat. 2, 252–263. as expert witnesses in court: what do they know about eyewitness memory?
doi: 10.1177/1077559597002003007 Psychol. Crime Law 21, 53–61. doi: 10.1080/1068316X.2014.915324
Dekker, S. (2018). Verbetering Feitenonderzoek in de Jeugdbeschermingsketen. Merckelbach, H., and Wessel, I. (1998). Assumptions of students
[Improvement of Evidence-Based Investigation in the Youth Care System.]. and psychotherapists about memory. Psychol. Rep. 82, 763–770.
Available online at: https://www.tweedekamer.nl/kamerstukken/brieven_ doi: 10.2466/pr0.1998.82.3.763
regering/detail?id=2018Z 0662anddid=2018D32523 Olafson, E. (2012). A call for field-relevant research about child forensic
Diekmann Schoenmaker, M., and Van der Veer, G. (2003). An extra language in interviewing for child protection. J. Child Sex. Abuse 21, 109–129. doi: 10.1080/
counseling and training. Intervention 1, 36–39. 10538712.2012.642469
Euser, E. M., van IJzendoorn, M. H., Prinzie, P., and Bakermans-Kranenburg, M. J. Ost, J., Easton, S., Hope, L., French, C. C., and Wright, D. B. (2017). Latent variables
(2010). Prevalence of child maltreatment in the Netherlands. Child Maltreat. 15, underlying memory beliefs of chartered clinical psychologists, hypnotherapists
5–17. doi: 10.1177/1077559509345904 and undergraduate students. Memory 25, 57–68. doi: 10.1080/09658211.2015.
Garven, S., Wood, J. M., Malpass, R. S., and Shaw, J. S. III (1998). More than 1125927
suggestion: the effect of interviewing techniques from the McMartin Preschool Otgaar, H., de Ruiter, C., Howe, M. L., Hoetmer, L., and van Reekum, P.
case. J. Appl. Psychol. 83, 347–359. (2017). A case concerning children’s false memories of abuse: recommendations
Goodman, G. S., and Melinder, A. (2007). Child witness research and forensic regarding expert witness work. Psychiatry Psychol. Law 24, 365–378. doi: 10.
interviews of young children: a review. Legal Criminol. Psychol. 12, 1–19. doi: 1080/13218719.2016.1230924
10.1348/135532506X156620 Otgaar, H., Horselenberg, R., van Kampen, R., and Lalleman, K. (2012). Clothed
Gore-Felton, C., Koopman, C., Thoresen, C., Arnow, B., Bridges, E., and Spiegel, and unclothed human figure drawings lead to more correct and incorrect
D. (2000). Psychologists’ beliefs and clinical characteristics: judging the veracity reports of touch in children. Psychol. Crime Law 18, 641–653. doi: 10.1080/
of childhood sexual abuse memories. Prof. Psychol. Res. Pract. 31, 372–377. 1068316X.2010.532129
doi: 10.1037/0735-7028.31.4.372 Otgaar, H., Howe, M. L., Merckelbach, H., and Muris, P. (2018). Who is the better
Houben, S. T. L., Otgaar, H., Roelofs, J., Wessel, I., Patihis, L., and Merckelbach, H. eyewitness? Sometimes children but at other times adults. Curr. Dir. Psychol.
(2019). Eye movement desensitization and reprocessing (EMDR) practitioners’ Sci. 27, 378–385. doi: 10.1177/0963721418770998
beliefs about memory. Psychol. Conscious. Theory Res. Pract. doi: 10.1037/ Otgaar, H., Howe, M. L., Patihis, L., Merckelbach, H., Lynn, S. J., Lilienfeld, S. O.,
cns0000211 [Epub ahead of print]. et al. (2019). The return of the repressed: the persistent and problematic claims
Huijer, J. (2014). Fact finding in the youth care system. A legal perspective. of long-forgotten trauma. Perspect. Psychol. Sci. 14, 1072–1095. doi: 10.1177/
Nederlands Juristenblad 13, 834–837. 1745691619862306
Inspectie Jeugdzorg (2014). Annual Report. Available online at: Patihis, L., Ho, L. Y., Tingen, I. W., Lilienfeld, S. O., and Loftus, E. F. (2013). Are
www.inspectiejeugdzorg.nl the ‘Memory Wars’ over? A scientist-practitioner gap in beliefs about repressed
Kinderombudsman (2013). Is de Zorg Gegrond? Analyse van het Feitenonderzoek memories. Psychol. Sci. 25, 519–530. doi: 10.1177/0956797613510718
aan de Basis van Ingrijpende Jeugdzorgbeslissingen.[Is the Concern Well- Patihis, L., and Pendergrast, M. H. (2019). Reports of recovered memories of
Founded? Analysis of Evidence-Based Research in Far-Reaching Youth Care abuse in therapy in a large age-representative U.S national sample: therapy
Decisions]. Available online at: http://www.dekinderombudsman.nl/ul/cms/ type and decade comparisons. Clin. Psychol. Sci. 7, 3–21. doi: 10.1177/
fckuploaded/2013.KOM008Isdezorggegr nd.pdf 2167702618773315
Kwakman, E. (2017). Handreiking Samenwerken bij Strafbare Kindermishandeling. Piper, A., Lillevik, L., and Kritzer, R. (2008). What’s wrong in believing in
[Guide to Collaboration in Criminal Child Abuse Cases]. Available online at: repression? A review for legal professionals. Psychol. Public Policy Law 3,
https://www.rijksoverheid.nl/documenten/publicaties/2017/11/09/handreiking 223–242. doi: 10.1037/a0014090
samenwerken-bij-strafbare-kindermishandeling Rivard, J. R., and Compo, N. S. (2017). Self-reported current practices in child
La Rooy, D., Brubacher, S. P., Aromaki-Stratos, A., Cyr, M., and Hershkowitz, I. forensic interviewing: training, tools and pre-interview preparation. Behav. Sci.
(2015). The NICHD protocol: a review of an internationally-used evidence- Law 35, 253–268. doi: 10.1002/bsl.2290
based tool for training child forensic interviewers. J. Criminol. Res. Policy Pract. Turnell, A., and Edwards, S. (1997). Aspiring to partnership, the signs of safety
1, 76–89. doi: 10.1108/JCRPP-01-2015-0001 approach to child protection. Child Abuse Rev. 6, 179–190.
Lamb, M. E., Brown, D. A., Hershkowitz, I., Orbach, Y., and Esplin, P. W. (2018). Westbrook, S. (2006). Utilizing the Signs of Safety Framework to Create Effective
Tell Me What Happened: Questioning Children About Abuse. Chichester: Wiley. Relationships with Child Protection Service Recipients. St. Paul, MN: MSW
Lamb, M. E., La Rooy, D. J., Malloy, L. C., and Katz, C. (2011). Children’s Testimony: Clinical Research.
A Handbook of Psychological Research and Forensic Practice. Chichester: Wiley. Wheeler, J., and Hogg, V. (2012). “Signs of safety and the child protection
Lamb, M. E., Orbach, Y., Herkowitz, I., Esplin, P. W., and Horowitz, D. movement,” in Solution-Focused Brief Therapy. A Handbook of Evidence-Based
(2007). A structured forensic interview protocol improves the quality and Practice, eds C. Franklin, T. S. Trepper, W. Gingerich, and E. E. McCollum
informativeness of investigative interviews with children: a review of research (London, UK: Oxford University Press), 203–215.
using the NICHD investigative interview protocol. Child Abuse Neglect 31, Yi, M., Jo, E., and Lamb, M. E. (2016). Effects of the NICHD protocol training on
1201–1231. doi: 10.1016/j.chiabu.2007.03.021 child investigative interview quality in Korean police officers. J. Police Crim.
Lamb, M. E., Orbach, Y., Sternberg, K. J., Aldridge, J., Pearson, S., Steward, H. L., Psychol. 31, 155–163.
et al. (2009). Use of a structured investigative protocol enhances the quality of
investigative interviews with alleged victims of child sexual abuse in Britain. Conflict of Interest: The authors declare that the research was conducted in the
Appl. Cogn. Psychol. 23, 449–467. doi: 10.1002/acp.1489 absence of any commercial or financial relationships that could be construed as a
Lindsay, D. S. (1990). Misleading suggestions can impair eyewitnesses’ ability potential conflict of interest.
to remember event details. J. Exp. Psychol. Learn. Mem. Cogn. 16,
1077–1083. Copyright © 2020 Erens, Otgaar, Patihis and de Ruiter. This is an open-access article
McElvaney, R. (2015). Disclosure of child sexual abuse: delays, non-disclosure and distributed under the terms of the Creative Commons Attribution License (CC BY).
partial disclosure. What the research tells us and implications for practice. Child The use, distribution or reproduction in other forums is permitted, provided the
Abuse Rev. 24, 159–169. doi: 10.1002/car.2280 original author(s) and the copyright owner(s) are credited and that the original
Melinder, A. (2002). “Children’s memory of a mildly stressful event – what help publication in this journal is cited, in accordance with accepted academic practice. No
is helpful?,” in Studies of the Mind: Proceedings of the First Norwegian Cypriot use, distribution or reproduction is permitted which does not comply with these terms.

ORIGINAL RESEARCH
published: 26 November 2020
doi: 10.3389/fpsyg.2020.571775
Risk Factors for Sexual Offending in

Self-Referred Men With Pedophilic
Disorder: A Swedish Case-Control
Study
Felix Wittström 1 , Niklas Långström 2 , Valdemar Landgren 3* and Christoffer Rahm 4
1
Center for Sexual Medicine (ANOVA), Karolinska University Hospital, Stockholm, Sweden, 2 Department of Medical
Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden, 3 Gillberg Neuropsychiatry Centre, Institute
of Neuroscience and Physiology, University of Gothenburg, Stockholm, Sweden, 4 Department of Clinical Neuroscience,
Karolinska Institutet, Stockholm, Sweden
Background: The risk of child sexual abuse among non-forensic, non-correctional

patients with Pedophilic Disorder (PD) is largely unknown.
Methods: We recruited a consecutive sample of 55 help-seeking, non-correctional
Edited by: adult men diagnosed with DSM-5 PD at a university-affiliated sexual medicine outpatient
Taina Laajasalo, unit in Sweden. PD participants were compared with 57 age-matched, non-clinical
National Institute for Health
and Welfare, Finland control men on four literature-based dynamic risk domains and self-rated child
Reviewed by: sexual abuse risk.
Alexander F. Schmidt,
Johannes Gutenberg University
Results: PD participants scored higher than controls on all tested domains (0–3 points);
Mainz, Germany expectedly so for pedophilic attraction (2.5 vs. 0.0, Cohen’s d = 2.40, 95% confidence
Katarina Alanko,
interval (CI): [1.91–2.89]), but also for sexual preoccupation (1.6 vs. 1.0, d = 1.11, 95%
Åbo Akademi University, Finland
CI: [0.71–1.51]), impaired self-regulation (1.4 vs. 1.0, d = 0.44, 95% CI: [0.06 to 0.81]),
*Correspondence:
Valdemar Landgren impaired cognitive empathy and antisocial traits (0.9 vs. 0.1, d = 1.18, 95% CI: [0.78–
valdemar.landgren@gu.se 1.59]), and self-rated child sexual abuse risk (1.0 vs. 0.0, d = 1.56, 95% CI: [1.13–1.98]).
When summarizing all five domains into a pre-specified composite score (0–15 points),
Specialty section:
This article was submitted to PD subjects scored substantially higher than matched control men (7.5 vs. 2.1, d = 2.12,
Forensic and Legal Psychology, 95% CI: [1.65–2.59]). Five (9%) PD participants self-reported any previous conviction
for a contact child sexual offense and eight (15%) for possession of child sexual abuse
Received: 11 June 2020
material or non-contact sexual offending (adult or child victim). Eighteen subjects (34%)
Accepted: 12 October 2020 acknowledged past week, child-related sexual behaviors.
Published: 26 November 2020
Citation:
Conclusion: Self-referred, help-seeking men with PD scored higher (small to very large
Wittström F, Långström N, effect sizes) than non-clinical control men on psychiatric measures of dynamic risk of
Landgren V and Rahm C (2020) Risk child sexual abuse suggested in prior research with correctional samples diagnosed with
Factors for Sexual Offending
in Self-Referred Men With Pedophilic PD. Our findings, including the composite risk measure, might inform clinical practice,
Disorder: A Swedish Case-Control but needs validation against actual sexual offending behavior.
Study. Front. Psychol. 11:571775.
doi: 10.3389/fpsyg.2020.571775 Keywords: pedophilic disorder, dynamic risk, child sexual abuse, case-control study, risk assessment

1 November 2020 | Volume 11 | Article 571775
Wittström et al. Pedophilic Disorder and Sexual Offending Risk
INTRODUCTION seek treatment for PD without prior identified child sexual

offending (Schaefer et al., 2010; Dombert et al., 2016).
Pedophilic Disorder (PD) is defined as intense and persistent We assessed the prevalence of dynamic risk factors for sexual
sexual attraction to prepubescent children associated with offending in self-referred, non-correctional adult men diagnosed
negative consequences for the individual or others (American with PD and age-matched control men. We also asked for self-
Psychiatric Association, 2013). Although many individuals with reported sexual offending behaviors against minors. Finally, to
PD do not sexually abuse children (Martijn et al., 2020), PD synthesize the empirical and theoretical literature, we constructed
remains an important risk factor and treatment target in child a pre-specified composite risk score. We hypothesized that PD
sexual abuse prevention (Jordan et al., 2011; Association for the participants would score substantially higher than controls on all
Treatment of Sexual Abusers [ATSA], 2014; Baur et al., 2016) and five risk domains and the composite score.
an estimated 50% of convicted child sexual offenders may meet
PD diagnostic criteria (Seto, 2018).
A methodological problem with prior PD research is that METHODS
most studies concern individuals involved with the criminal
justice system. This entails substantial risks that findings Setting
secondary to selection biases following subject status as an This case-control study was part of the research project
identified and convicted child sexual offender diagnosed with Pedophilia at Risk—Investigations of Treatment and Biomarkers
PD may be over-interpreted as causal for the pedophilic (PRIOTAB), conducted from March 2016 to April 2019 at the
attraction as such. Karolinska University Hospital in Stockholm, Sweden. PRIOTAB
Individuals are increasingly seeking help for PD in sexual included clinical interviews, psychological testing and self-
medicine contexts and general psychiatry settings, such as the reports, neuroimaging, blood sampling and enrollment in a
German prevention project Dunkelfeld (Schaefer et al., 2010; double-blind, randomized controlled efficacy trial (RCT) of the
Knack et al., 2019). Treatment decisions are partly based on testosterone-suppressing medication degarelix (Landgren et al.,
clinically perceived risk of child sexual abuse, and current 2020). Individuals calling PrevenTell1 , a national telephone
guidelines suggest SSRI medication or psychotherapy for low risk helpline addressing unwanted sexuality were screened for PD
and antiandrogen medication for high risk individuals (Thibaut and invited to participate. PrevenTell is a low-threshold service
et al., 2010). However, validated measures of reoffending risk providing counseling and treatment for sexual risk behaviors.
such as the Static-99R and STABLE-2007 (Hanson et al., 2016; The target group is people with self-experienced risky behavior,
Brankley et al., 2019) were developed in correctional settings compulsive preoccupation with sexual thoughts or actions, sexual
and have limited utility with non-offender, at-risk individuals. interest in children or impulses to force someone into sex.
Hence, risk assessment tools, addressing particularly dynamic, PrevenTell has been providing services since 2011, financed
potentially changeable risk factors, are needed also for general through the tax-funded health care system and targeted support
psychiatry and sexual medicine. Improved knowledge about from the Swedish Government. PrevenTell has received public
sexual offending risk among self-referred individuals with PD visibility through repeated media coverage and increasing
could advance treatment tailoring and the prevention of child exposure in online search engines, and receives an average of
sexual offending (Laws, 2000; Långström et al., 2013; Khan et al., four calls per day.
2015; Knack et al., 2019).
Assessing risk of sexual offending is important but complex Participants
and ethically challenging (Craig et al., 2005; Khan et al., Men aged 18–66 years with DSM-5 PD ascertained by
2015). Dynamic and static recidivism risk factors, although telephone interview prior to intake and confirmed on site
not necessarily causal, have indeed been identified among by a board-certified general psychiatrist were eligible for
convicted sexual offenders. A classic systematic review (Hanson study inclusion. Psychiatric exclusion criteria included severe
and Morton-Bourgon, 2005) found deviant sexual interest psychosis, severe and acutely increased substance misuse, or
(e.g., pedophilia), sexual preoccupation, impaired self-regulation suicide risk, respectively. Medical exclusion criteria included
and antisocial traits among the most prominent dynamic or contraindications to magnetic resonance imaging (MRI) and
potentially changeable risk factors (Cohen’s d = 0.2–0.4) for drug trial inclusion; previously known or newly ascertained
sexual offense recidivism in known sexual offenders. Further, severe osteoporosis, electrocardiogram-verified prolonged
cognitive empathy has been linked theoretically to antisocial QT/QT interval (> 450 ms), kidney or liver insufficiency,
behavior and a recent meta-analysis suggested that it is severe asthma, hypersensitivity to study drug or participation in
moderately lower (Hedges’ g = −0.58) among child sexual another drug study during the previous 3 months. Karolinska
offenders compared to the general population, but not compared Trial Alliance assists Karolinska University Hospital in recruiting
to sexual offenders of adults (Morrow, 2019). The motivation- participants for clinical trials and, along with independent study
facilitation model for sexual offending (Seto, 2019) suggests that monitoring, helped the PRIOTAB project by disseminating
pedophilic interest and hypersexuality are motivational factors, information about the opportunity to participate as a non-
antisocial behavior trait factors, and impaired self-control a patient control in the study. Healthy controls were recruited
state factor. However, sexual offending risk and actual offending
behavior is mostly unknown among individuals who voluntarily 1
www.preventell.se

primarily through adverts on Karolinska Institutet’s homepage, prespecified composite measure (SChiMRA, see Supplementary
and through Karolinska Trial Alliance. Exclusion criteria were Material) developed by us to quantify self-rated risk of sexual
the same for PD participants and controls. Finally, 55 PD offending against children and past week sexually abusive
participants and 57 age-matched controls were included in data behavior toward children.
analyses (Figure 1). The five putative dynamic risk domains for child sexual
abuse; pedophilic attraction (DSM-5 diagnostic criteria), sexual
Procedure preoccupation (SDI, HBI and ongoing offensive behavior
PD men and age-matched control men were all examined by according to SChiMRA-B), impaired self-regulation (CCPT-II),
a psychiatrist, completed self-rating questionnaires, provided impaired cognitive empathy and antisocial traits (RAADS-14,
blood samples, and underwent neuropsychological testing by RMET and ASPD symptoms according to MINI 6.0), and
a clinical psychologist. Self-report measures were filled out in self-rated child sexual abuse risk (SChiMRA-A), were assessed
privacy on site and participants were instructed to ask study staff and compared between PD participants and control men. We
for assistance if they had questions or difficulty answering the predefined the composite score as described in Supplementary
questionnaires. All participants underwent assessments during Table 1 by assigning 0–3 points to each of the five domains
the same time of the day starting in the morning. Approximately and summing them with equal weights to a total score ranging
6 h were required for the assessments, the first two provided data from 0 to 15. Hence, the composite risk score is based on
presented in this case-control study. self-, expert-completed and behavioral measures and potentially
treatment sensitive since all five included domains could capture
Ethical Considerations change over time. The five-domain score was used as the primary
A designated research nurse oversaw procedures where patient outcome measure in the PRIOTAB RCT (Landgren et al., 2020).
identification was necessary for clinical safety and insurance Here we report both the five-domain score and a four-domain
reasons; research subjects were otherwise only known to score (range 0–12) excluding pedophilic attraction; part of the
researchers by their initials. Participants were initially informed inclusion criteria for PRIOTAB but also the grouping criterion
about health professionals’ obligation according to the Swedish for the present case-control study. A detailed description of the
Social Services Act to immediately notify the social services when instruments is provided in Supplementary Material.
a named child is at imminent risk of abuse or maltreatment.
Subjects were asked at every visit to inform about such children Statistical Analyses
in their vicinity. According to Swedish law, professionals are also We computed Cohen’s ds with 95% confidence intervals to
allowed, but not obliged, to supersede confidentiality and tell the express effect sizes of potential differences between cases and
police any potential crime against children admitted by a patient. controls. Using the freely available Practical Meta-Analysis Effect
If a study participant reported any such actions, the choice to Size Calculator (Effect Size Calculator, 2020) provided by the
report or not to the police was discussed with an external ethico- Campbell Collaboration and based on Lipsey and Wilson (2001),
legal advisory board linked to the Karolinska University Hospital. Cohen’s ds were calculated from 2 by 2 (occasionally 3 by 2) tables
The study was approved by the Swedish Central Ethical Review for binary (frequency) data, from standardized mean differences
Board (no: Ö 26-2014). Participants were offered reimbursement for parametric test results and based on p-values and sample
for transports to study visits and a financial compensation of sizes for non-parametric test results. Following Cohen (Cohen,
SEK 1000 (the equivalent of 94 Euros before taxation) upon 1988), Cohen’s ds were interpreted as small (0.20–0.49), medium
study completion. All subjects provided oral and signed informed (0.50–0.79) or large (0.80+) effects.
consent and were offered treatment as usual after the study. Some data are missing since study participants occasionally
did not complete self-rating questionnaires and no imputation
Instruments procedures were employed. The number of individuals
We compared PD and control participants on three self- with complete data for each variable is reported in table
reports, the Sexual Desire Inventory (SDI) (Spector et al., 1996) footnotes. Data were analyzed using SPSS Version 24
measuring sexual interest; the Hypersexual Behavior Inventory for Windows.
(HBI) (Reid et al., 2011) assessing hypersexuality; and the
Ritvo Autism and Asperger Diagnostic Scale—Screening Tool
(RAADS-14) (Eriksson et al., 2013) mentalizing subscale. We RESULTS
complemented this with three expert-administered measures;
Conners’ Continuous Performance Test—2nd edition (CCPT- Table 1 suggests that PD participants and control men were
II) (Conners, 2000) tapping impulsivity; the Reading the similar regarding education level, non-sexual offenses and being
Mind in the Eyes Test (RMET) (Baron-Cohen et al., 2001) a parent, whereas PD participants were less often employed
addressing the understanding of others’ emotions; and Antisocial (Cohen’s d = −0.90), more often single (d = 0.58), had less
Personality Disorder (ASPD) symptoms according to the MINI frequently lived with an intimate partner for 2+ years (d = −0.63),
Neuropsychiatric Interview 6.0 (Sheehan et al., 1998). had lower IQ (d = −0.79)—although on average within the
Additionally, participants completed measures addressing normal range (95–105), and more often self-reported convictions
pedophilic attraction and related behavior and self-rated risk. We for any sexual offense (d = 1.52) or non-contact sexual offenses
evaluated PD symptoms based on DSM-5 criteria and used a (d = 1.25). Further, PD participants scored higher compared to

FIGURE 1 | Flow chart for study recruitment of matched population control men and help-seeking, non-correctional adult men with Pedophilic Disorder in Sweden.
controls on all dynamic sexual offending risk domains and the domain scores with and without adjustment for IQ. Naturally,
predefined composite risk score (Table 2). Effect sizes for risk group status was moderately to strongly significantly correlated
domains were significant and small (impaired self-regulation, with risk scores before adjustment (impaired self-regulation,
d = 0.44), large (sexual preoccupation d = 1.11, impaired cognitive Spearman’s r = −0.22, p < 0.05; low empathy and antisocial traits,
empathy and antisocial traits, d = 1.18, self-rated risk, d = 1.56) r = −0.55, p < 0.01; four-domain risk score, r = −0.71, p < 0.01).
and, following inclusion criteria for PD subjects, extremely large Adjusting for IQ in partial correlation analyses decreased
(pedophilic attraction, d = 2.40). Individual items pertaining associations somewhat (impaired self-regulation, r = −0.13,
to pedophilic attraction and self-rated risk (such as past week p = 0.19; low empathy and antisocial traits, r = −0.48, p < 0.01;
watching, socializing with, or sexually interacting with children) four-domain risk score, r = −0.66, p < 0.01). The second post hoc
were not directly compared across groups, only as part of sensitivity analysis examined the robustness of the composite
domain and composite scores. Most notably, however, 12 (22%) score difference by excluding the 12 PD participants with a
PD participants reported prior convictions for sexuality-related history of sexuality-related offending. The latter had a composite
crimes; five (9%) stated a contact sex offense against a child score Mdn = 8 (n = 12, IQR 6.0–9.5) compared to Mdn = 8
(<15 years of age) and eight (15%) any non-contact sexual in those without (n = 41, IQR 6.0–9.0). Exclusion reduced the
offense; mostly illegal possession of child sexual abuse material composite score difference compared to controls from d = 2.12
(CSAM), legally referred to as “child pornography”. Eighteen PD (95% CI: 1.65 to 2.59) to d = 2.11 (1.61 to 2.61).
subjects (34%) acknowledged past week, child-related sexual at-
risk behavior, one of them reported actual child sexual interaction
during the past week. 66% estimated having a 40% or higher risk DISCUSSION
of future child-related sexual at-risk behavior; if there existed an
easy way to escape detection. Finally, we conducted two post hoc We compared non-correctional, self-referred adult men
sensitivity analyses. First, as IQ is empirically linked to measures diagnosed with DSM-5 Pedophilic Disorder (PD) to age-matched
of self-regulation and empathy (Vellante et al., 2013; Chien et al., non-clinical control men to elucidate dynamic sexual offending
2017), we correlated group status (PD or control participant) to risk factors in a non-forensic clinical context. There were four

TABLE 1 | Socio-demographic, criminological and sexuality baseline characteristics among help-seeking, non-correctional adult men with Pedophilic Disorder and
matched non-clinical control men in Sweden.
Characteristic Pedophilic Disorder men Non-clinical control men Cohen’s d (95% CI)
(n = 55) (n = 57)
Age, range, M (SD) 18–66, 36 (12) 18–64, 36 (12) −0.05 (−0.42 to 0.32)
Full scale IQ (WAIS-IV), Mdn (IQR) 101 (23) 115 (15)a −0.79 (−1.17 to−0.40)
Education, highest level, n (%)
Primary school ≤ 9 years 6 (11) 2 (4)
Secondary school 1–3 years 26 (47) 23 (40) −0.36 (−0.73 to 0.02)
Postsecondary education 23 (42) 32 (56)
Employed, n (%) 32 (58) 50 (88) −0.90 (−1.43 to−0.38)
Parent/guardian, n (%) 20 (36) 26 (46) −0.21 (−0.63 to 0.21)
Currently lives as single, n (%) 35 (64) 21 (37) 0.58 (0.16 to 0.99)
Ever lived with partner for more than 2 years, n (%) 26 (47) 42 (74) −0.63 (−1.06 to−0.05)
Pedophilic sexual attraction
Age of discovery of pedophilic sexual attraction, range, M (SD), Mdn (IQR)b 6–39, 18 (7), 16 (13–23) – NA
Attraction primarily to boys, n (%) 8 (15) – NA
Attraction primarily to girls, n (%) 43 (78) – NA
Attraction to boys and girls, n (%) 4 (7) – NA
Exclusive attraction to prepubescent children, n (%) 12 (22) – NA
Self-reported convictions, n (%)c
Any sex offense 12 (22) 1 (2) 1.52 (0.37 to 2.66)
Contact sex offense (child)c 5 (9) 0 (0) 0.96 (−0.24 to 2.16)
Non-contact sex offense (child or adult) or CSAM offensec,d 8 (15) 0 (0) 1.25 (0.09 to 2.42)
Any non-sexual offense 8 (15) 13 (23) −0.30 (−0.84 to 0.23)
Bolded figures are significant at p < 0.05.

NA, not applicable.
For reasons of completeness regarding Cohen’s d estimates, we exchanged frequencies of “0” to “1” in effect size calculations.
a Missing data for 1 control.
b Excluded 4 participants reporting that they “had always known” about their pedophilic sexual attraction and did not provide a specified age.
c One PD patient reported both contact and non-contact sexual offenses.
d CSAM offense = possession of child sexual abuse material.
main findings. First, HBI-assessed hypersexual behavior, lower more cognitively high-functioning individuals volunteering
RAADS-14 mentalizing-based cognitive empathy, and DSM-5 for study participation. In addition, our sensitivity analysis
Antisocial Personality Disorder symptoms were much more suggested that IQ is unlikely to be a major confounder of
pronounced among PD cases than controls; effect sizes were the link between group status (case or control) and the
large. Second, and contrarily, SDI hyposexuality, ADHD-like composite risk score. Although PD participants had less lifetime
self-regulation measured with the CCPT-II test and RMET-based experience of living with a partner and were more often
theory of mind impairments did not differ meaningfully between single vs. controls, the prevalence of being a parent did
PD and control men. Third, a non-trivial proportion of PD not differ.
participants reported prior convictions for sexuality-related
offending, current child-related sexual at-risk behavior and at Pedophilic Disorder
least moderate risk of future child-related sexual at-risk behavior. Since PD was a core inclusion criterion, the lowest possible
Fourth, when dynamic risk domains were combined into a score in this domain was 2 for PD participants; hence, the
predefined composite measure, the resulting summary score effect size for the difference vs. controls on this domain risk
was very much larger among PD men vs. controls. This large score was extremely large. Although reluctance to disclose
difference remained after excluding the pedophilic attraction symptoms cannot be ruled out, null findings regarding PD
domain and each risk domain score was weakly to very much symptoms among age-matched control men agrees with the
higher in cases compared to controls. notion that persistent pedophilic attraction is infrequent in the
general population, perhaps 1% at most (Santtila et al., 2015;
Socio-Demographic Characteristics Seto, 2018).
Although education levels were similar, unemployment was
moderately more prevalent among PD participants than Sexual Preoccupation
controls. PD men also had lower IQ, albeit not clinically SDI baseline scores did not differ meaningfully between PD
significantly so. However, both employment rates and higher and control participants, indicating that hypoactive sexual
IQ among controls may partly reflect selection bias toward motivation is not over-represented in help-seeking, non-forensic

TABLE 2 | Comparisons of five domains of dynamic risk factors for child sexual abuse among help-seeking, non-correctional adult men with Pedophilic Disorder and
age-matched, non-clinical control men in Sweden.
Risk factor/domain Pedophilic Disorder Non-clinical control Cohen’s d (95% CI)

men (n = 55) men (n = 57)
Pedophilic attraction
Domain risk score (range 0–3), M (SD) 2.5 (0.6) 0.0 (0.0) 2.40 (1.91 to 2.89)
Sexual preoccupation
Sexual Desire Inventory scorea (range 12–109), Mdn (IQR) 70 (19) 65 (14) 0.21 (−0.16 to 0.59)
Hypersexual Behavior Inventory scoreb (range 19–95), Mdn (IQR) 56 (22) 24 (10) 1.70 (1.26 to 2.14)
Past week watched child sexual abuse material or observed children for sexual
arousalc , n (%)
Not at all 37 (70) 57 (100) –
A few days 11 (21) 0 (0) –
More than half the days 1 (2) 0 (0) –
Nearly every day 4 (8) 0 (0) –
Past week socializing with children for sexual arousal, n (%)
Not at all 50 (94) 57 (100) –
A few days 3 (6) 0 (0) –
More than half the days/nearly every day 0 (0) 0 (0) –
Past week direct sexual interaction with children
Not at all 52 (98) 57 (100) –
A few days 1 (2) 0 (0) –
More than half the days/nearly every day 0 (0) 0 (0) –
Domain risk score (range 0–3)a , M (SD) 1.6 (0.7) 1.0 (0.2) 1.11 (0.71 to 1.51)
Impaired self-regulation
Conners’ Continuous Performance Test aspects most like ADHD normsb (range 2.3 (1.8) 1.8 (1.5) 0.31 (−0.06 to 0.69)
0–12), M (SD)
Domain risk score (range 0–3), M (SD) 1.4 (1.0) 1.0 (1.0) 0.44 (0.06 to 0.81)
Impaired cognitive empathy and antisocial traits
Ritvo Autism and Asperger Diagnostic Scale – Screening Tool mentalizing 10 (12) 2 (3) 0.81 (0.43 to 1.20)
subscale scorec (range 0–21), Mdn (IQR)
Reading the Mind in the Eyes Test scorec (range 0–36), Mdn (IQR) 27 (7) 29 (5) −0.33 (−0.71 to 0.04)
No. of DSM-5 Antisocial Personality Disorder symptoms (range 0–12), M (SD) 3 (3) 1 (2) 1.23 (0.83 to 1.64)
Domain risk score (range 0–3)c , M (SD) 0.9 (0.8) 0.1 (0.4) 1.18 (0.78 to 1.59)
Self-rated substantial risk of child sexual abuseb , n (%)
Watching child sexual abuse material or observing children for sexual arousal 35 (66) 0 (0) –
Socializing with children for sexual arousal 9 (17) 0 (0) –
Direct sexual interaction with children 11 (21) 0 (0) –
Domain risk score (range 0–3)b , M (SD) 1.0 (1.0) 0.0 (0.0) 1.56 (1.13 to 1.98)
Five-domain composite risk score (range 0–15)a,d , M (SD) 7.5 (1.7) 2.1 (1.1) 2.12 (1.65 to 2.59)
Four-domain composite risk score (excluding pedophilic attraction, range 4.9 (1.9) 2.1 (1.1) 1.62 (1.18 to 2.05)
0–12), M (SD)a
Bolded figures are significant at p < 0.05.

a n = 53 and n = 55 subjects for PD and control groups, respectively.
b n = 53 and 57.
c n = 54 and 57.
d For a description of the composite risk score see Supplementary Table 1.
PD patients. In contrast, we found a large difference between fact, paraphilic and hypersexuality comorbidity might fuel
PD and control men regarding HBI hypersexuality scores, help-seeking behaviors (Kafka, 1997; Walton et al., 2017).
which also contributed substantially to the large difference The median HBI score for our PD subjects is above the
between PD and control men in the sexual preoccupation suggested cut off-score of 53 when screening for hypersexual
domain risk score. The HBI was designed to capture repetitive disorder (Reid et al., 2011), suggesting such comorbidity and
sexual behaviors in response to dysphoric mood states or highlighting their motivation to seek help. Simultaneously,
stressful life events and the degree to which they are the median HBI score in control men (24.0) was lower
perceived as uncontrollable (Reid et al., 2011). Hypersexual than that of a large Spanish non-clinical sample (30.5)
behavior appears to co-occur with paraphilic interest; in (Ballester-Arnal et al., 2019).

Impaired Self-Regulation participants was that those with PD and prior child sexual contact
We found no significant differences between PD and control self-reported lower mentalizing ability (d = 0.38, p < 0.05),
men in binary self-regulation measures derived from the CCPT- and a trend toward impaired emotion attribution (d = 0.31,
II, but a moderate, significant difference when integrated into p = 0.05) (Schuler et al., 2019). This is partly consistent with
the risk domain score. The CCPT-II is used supportively the poorer self-rated mentalizing ability and trend toward poorer
in both assessment and treatment evaluation of ADHD, but emotion attribution found among our non-forensic but help-
effect sizes of the mean difference between ADHD patients seeking PD men.
and controls has only been moderate (Hervey et al., 2004). The large difference in Antisocial Personality Disorder
Due to the insufficient discriminative power for ADHD, symptoms should be vetted toward mixed results in prior
CCPT-II results are not considered diagnostic. Yet, test- research. Cohen et al. (2018) studied individuals convicted of
measured executive dysfunction is not a prominent characteristic child sexual offending (n = 50), non-convicted, non-clinical
of our non-forensic PD patients. Importantly, this agrees individuals with pedophilic attraction (n = 195) and non-PD
with studies suggesting that executive dysfunction is more controls (n = 60). Their findings suggested more antisocial traits
strongly associated with child sexual offending than with among convicted offenders, but not among non-convicted PD
PD as such (Joyal et al., 2007; Suchy et al., 2009; Eastvold individuals, compared to controls. Another, likely underpowered
et al., 2011; Schiffer and Vonlaufen, 2011). Massau et al. study of CSAM offenders (n = 23), non-contact offenders
(2017) compared men with pedophilic attraction with or (n = 15) and contact child molestation offenders (n = 49) (Jung
without prior contact child sexual offending to men with no et al., 2013) found self-reported antisocial traits to fall within
pedophilic attraction but with or without prior contact child clinically significant ranges across groups. However, more ASPD
sexual offending. One measure of impulsivity was minimally symptoms in PD participants could partly be inflated by the
higher in offending compared to non-offending participants question specifically asking about “Repeated illegal acts as an
(η2 = 0.032, p = 0.03) but did not differ according to adult” with previous CSAM offending being somewhat common
pedophilic attraction. in our patient population.
Child-Related Sexually Motivated

Impaired Cognitive Empathy and
Behavior: Self-Report and Self-Assessed
Antisocial Traits
The RAADS-14 mentalizing subscale, RMET and Antisocial
Risk
Personality Disorder symptoms all comprised the impaired The proportion of our PD participants self-reporting any
cognitive empathy and antisocial traits domain but contributed previous conviction of a sexuality-related offense supports
differently to the large difference in this domain risk score the notion that a non-negligible minority of help-seeking
between PD and control participants. PD patients may have prior sexual offense convictions.
The RAADS-14 mentalizing subscale attempts to Additionally, we observed a large proportion of PD participants
capture Autism Spectrum Disorder (ASD)-related atypical acknowledging past week, child-related sexual behaviors such
communication and social interaction (Baghdadli et al., 2017) as viewing CSAM or observing children for sexual arousal
and our PD participants scored considerably higher than and endorsing a clinically significant risk that they would, if
controls. Eriksson and colleagues (Eriksson et al., 2013) found not discovered, watch CSAM or children directly off-line for
a median RAADS-14 mentalizing score of 15 among adult men sexual purposes. First, this supports the notion that watching
in Sweden with ASD and 4 among men with other psychiatric CSAM may be a diagnostic marker for PD (Seto et al.,
disorders than ASD or ADHD. Thus, PD participants’ median 2006). Second, it implies that help-seeking individuals with
fell between those of male ASD patients and men with other PD may be at risk of committing child sexual abuse and
psychiatric morbidity, whereas that of our controls agree with could benefit from risk-reducing treatment. Our findings are
those of the non-psychiatric controls (Eriksson et al., 2013). consistent with prior research suggesting that most self-identified
However, importantly, the full diagnostic criteria for ASD were individuals with PD or respondents in anonymous online
not used by us, and a prior ASD diagnosis did not exclude surveys endorse past or present CSAM viewing (Riegel, 2004;
study participation. Neutze et al., 2011; Klein et al., 2015). The present results
The similarity in RMET scores between cases and controls is reflect the past week situation when PD patients seek help,
somewhat surprising; theoretically, RMET should measure the and complement the lifetime prevalence numbers reported in
same construct as the RAADS-14 mentalizing domain. Some the seminal German study of 137 self-identified PD individuals
research (Oakley et al., 2016) suggests RMET is primarily a where 69% reported previous contact child sexual offending
measure of emotion recognition. Thus, PD participants’ self-rated (Neutze et al., 2011).
mentalizing impairments might capture more complex real-
life social situations than the RMET. Recently, Schuler et al. Static and Dynamic Risk of Child Sexual
(2019) compared men with pedophilic attraction with (n = 85) Abuse in PD
and without (n = 72) prior contact child sexual offenses to Static and dynamic risk factors co-occur and interact over time
male non-offending controls (n = 128) on several aspects of (Craig et al., 2005). To further investigate the overlap of static and
empathy. The only significant difference between PD and control dynamic risk among PD patients, future studies might measure

static risk factors with tools such as the recently validated Child CONCLUSION
Pornography Offender Risk Tool (CPORT) (Eke et al., 2019).
The CPORT predicts general sexual and child pornography The help-seeking and self-identified PD individuals in this
recidivism (AUCs = 0.72–0.74) among individuals adjudicated study exhibited offending risk-increasing characteristics. A non-
for CSAM offending but without known contact sexual offenses. trivial minority also report prior CSAM and direct child sexual
Dynamic measures for sexual offending risk among convicted offending or current child-related at-risk sexual behavior. This
offenders have been developed (Hanson and Morton-Bourgon, may need consideration in clinical services for help-seeking
2009). Two such established measures are the STABLE-2007 and individuals with PD and supports the idea of PD as a potential
ACUTE-2007, encompassing for example sexual preoccupation, treatment target in child sexual abuse prevention. Low-threshold
self-regulation problems and antisocial behavior. They both specialist services such as the PrevenTell helpline, the recruitment
appear to add incremental predictive power beyond static risk base for the current study, might be important resources in
alone to risk assessments (Brankley et al., 2019). However, these such prevention efforts. Except from the sexual interest in
measures were designed for individuals currently in contact minors per se, risk-related characteristics may include concurrent
with the legal system for sexual offending, whereas medical hypersexuality or sexual preoccupation, mentalizing or cognitive
professionals generally lack risk measures for non-forensic empathy deficits, and antisocial traits. In contrast, non-forensic
clinical use. We propose that our dynamic composite risk score PD patients may not necessarily differ from healthy general
might aid in further development of such clinically useful risk population controls in terms of hypoactive sexual desire and
assessment and management methods. continuous performance tests of inattention, impulsivity, and
vigilance. We present a preliminary, dynamic composite risk
Strengths and Limitations score that might be useful in developing risk assessment measures
PD patients were recruited through a national helpline in for use in child sexual abuse prevention efforts in sexual medicine
Sweden, resulting in a consecutive nationwide sample of and general psychiatry.
self-referred, non-forensic participants. Consequently, included
PD patients were likely similar to those clinicians would
see in sexual medicine and general psychiatry settings, and
representative of the target group often addressed in selective DATA AVAILABILITY STATEMENT
prevention of child sexual abuse. Further, the truthfulness of
The datasets presented in this article are not readily available
the self-report provided by participants was strengthened by a
because the dataset contains sensitive and intimate details about
pseudonymization procedure. Finally, most employed measures
the participants. Requests to access the data will be evaluated
had been properly validated in other clinical settings, enabling
on a case by case basis. Such requests should be directed to CR,
comparisons of this patient population to those diagnosed with
christoffer.rahm@ki.se.
other psychiatric disorders. Regarding limitations, we were not
able to prospectively investigate associations between empirically
and theoretically suggested dynamic risk measures and actual
future CSAM or contact child sexual offending. Further, poor ETHICS STATEMENT
concurrent validity has been suggested across various measures
of the cognitive empathy construct (Chen et al., 2017), indicating The studies involving human participants were reviewed and
that it is indeed difficult to measure. Inevitably, any study that approved by the Swedish Central Ethical Review Board. The
requires opting in or help-seeking efforts will be vulnerable to patients/participants provided their written informed consent to
selection biases toward subjects who are indeed motivated to participate in this study.
participate. Here by taking part in a project involving a novel
treatment addressing impairing sexual preoccupation, impaired
cognitive empathy, antisocial traits and self-assessed offending
risk. Last, the validity of self-ratings of sexual offending risk and AUTHOR CONTRIBUTIONS
sexually motivated at-risk or abusive behavior against children
CR designed the study and collected the data. FW drafted the
has not been investigated. However, by safeguarding participant
initial manuscript and CR, VL, and NL reviewed the manuscript
anonymity, these self-reports may be as truthful as possible in a
for important intellectual content. All authors contributed
clinical PD study with face-to-face interactions.
to data analyses.
Future studies should investigate prospectively the ability of
the measures tested here to predict sexual offending, preferably
by looking at risk behaviors and situations, self-reports of child
sexual offending, and suspicions/arrests for suspected sexual FUNDING
offenses including CSAM crime (Långström et al., 2013). Also,
the malleability of these dynamic risk factors or potential Funding was provided by the Swedish Society for Medical
treatment effect mediators should be investigated, as attempted in Research; Gothenburg Society of Medicine; Swedish Society of
the recently published results of the PRIOTAB RCT of degarelix Medicine; the Söderström Königska Foundation; the Fredrik and
(Landgren et al., 2020). Ingrid Thuring Foundation; the Center for Psychiatry Research,

Department of Clinical Neuroscience, Karolinska Institutet, SUPPLEMENTARY MATERIAL

Stockholm; the Research unit of Skaraborg Hospital, Skövde;
and through the regional ALF agreement between Karolinska The Supplementary Material for this article can be found
Institutet and Region Stockholm. The funders had no influence online at: https://www.frontiersin.org/articles/10.3389/fpsyg.
on study design, analyses, conclusions, or the decision to publish. 2020.571775/full#supplementary-material
REFERENCES Eriksson, J. M., Andersen, L. M., and Bejerot, S. (2013). RAADS-14 Screen:
validity of a screening tool for autism spectrum disorder in an adult psychiatric
American Psychiatric Association (2013). Diagnostic and Statistical Manual population. Mol. Autism 4:49. doi: 10.1186/2040-2392-4-49
of Mental Disorders, 5th Edn. Washington, D.C: American Psychiatric Hanson, R. K., and Morton-Bourgon, K. E. (2005). The characteristics of persistent
Association. sexual offenders: a meta-analysis of recidivism studies. J. Consult. Clin. Psychol.
Association for the Treatment of Sexual Abusers [ATSA] (2014). Practice 73, 1154–1163. doi: 10.1037/0022-006x.73.6.1154
Guidelines for the Assessment, Treatment, and Management of Male Adult Hanson, R. K., and Morton-Bourgon, K. E. (2009). The accuracy of recidivism
Sexual Abusers. Beaverton, OR: ATSA. Association for the Treatment of Sexual risk assessments for sexual offenders: a meta-analysis of 118 prediction studies.
Abusers. Psychol. Assess. 21, 1–21. doi: 10.1037/a0014421
Baghdadli, A., Russet, F., and Mottron, L. (2017). Measurement properties of Hanson, R. K., Thornton, D., Helmus, L.-M., and Babchishin, K. M. (2016). What
screening and diagnostic tools for autism spectrum adults of mean normal sexual recidivism rates are associated with Static-99R and Static-2002R scores?
intelligence: a systematic review. Eur. Psychiatry 44, 104–124. doi: 10.1016/j. Sex. Abuse 28, 218–252. doi: 10.1177/1079063215574710
eurpsy.2017.04.009 Hervey, A. S., Epstein, J. N., and Curry, J. F. (2004). Neuropsychology of
Ballester-Arnal, R., Castro-Calvo, J., Gil-Julia, B., Giménez-García, C., and Gil- adults with Attention-Deficit/Hyperactivity disorder: a meta-analytic review.
Llario, M. D. (2019). A validation study of the spanish version of the hypersexual Neuropsychology 18, 485–503. doi: 10.1037/0894-4105.18.3.485
behavior inventory (HBI): paper-and-pencil versus online administration. Jordan, K., Fromberger, P., Stolpmann, G., and Müller, J. L. (2011). The role of
J. Sex. Marital. Ther. 45, 283–302. doi: 10.1080/0092623x.2018.1518886 testosterone in sexuality and paraphilia–a neurobiological approach. Part II:
Baron-Cohen, S., Wheelwright, S., Hill, J., Raste, Y., and Plumb, I. (2001). The testosterone and paraphilia. J. Sex. Med. 8, 3008–3029. doi: 10.1111/j.1743-
“Reading the Mind in the Eyes” test revised version: a study with normal adults, 6109.2011.02393.x
and adults with asperger syndrome or high-functioning autism. J. Child Psychol. Joyal, C. C., Black, D. N., and Dassylva, B. (2007). The neuropsychology and
Psychiatry 42, 241–251. doi: 10.1111/1469-7610.00715 neurology of sexual deviance: a review and pilot study. Sex. Abuse 19, 155–173.
Baur, E., Forsman, M., Santtila, P., Johansson, A., Sandnabba, K., and Långström, doi: 10.1177/107906320701900206
N. (2016). Paraphilic sexual interests and sexually coercive behavior: a Jung, S., Ennis, L., Stein, S., Choy, A. L., and Hook, T. (2013). Child pornography
population-based twin study. Arch. Sex. Behav. 45, 1163–1172. doi: 10.1007/ possessors: comparisons and contrasts with contact- and non-contact sex
s10508-015-0674-2 offenders. J. Sex. Aggress. 19, 295–310. doi: 10.1080/13552600.2012.741267
Brankley, A. E., Babchishin, K. M., and Hanson, R. K. (2019). STABLE- Kafka, M. P. (1997). Hypersexual desire in males: an operational definition and
2007 demonstrates predictive and incremental validity in assessing risk- clinical implications for males with paraphilias and paraphilia-related disorders.
relevant propensities for sexual offending: a meta-analysis. Sex. Abuse Arch. Sex. Behav. 26, 505–526.
13:1079063219871572. doi: 10.1177/1079063219871572 Khan, O., Ferriter, M., Huband, N., Powney, M. J., Dennis, J. A., and Duggan, C.
Chen, K.-W., Lee, S.-C., Chiang, H.-Y., Syu, Y.-C., Yu, X.-X., and Hsieh, C.-L. (2015). Pharmacological interventions for those who have sexually offended or
(2017). Psychometric properties of three measures assessing advanced theory of are at risk of offending. Cochrane Database Syst. Rev. 2015:CD007989.
mind: evidence from people with schizophrenia. Psychiatry Res. 257, 490–496. Klein, V., Schmidt, A. F., Turner, D., and Briken, P. (2015). Are sex drive and
doi: 10.1016/j.psychres.2017.08.026 hypersexuality associated with pedophilic interest and cild sexual abuse in a
Chien, Y.-L., Chou, M.-C., Chiu, Y.-N., Chou, W.-J., Wu, Y.-Y., Tsai, W.-C., male community sample? PLoS One 10:e0129730. doi: 10.1371/journal.pone.
et al. (2017). ADHD-related symptoms and attention profiles in the unaffected 0129730
siblings of probands with autism spectrum disorder: focus on the subtypes of Knack, N., Winder, B., Murphy, L., and Fedoroff, J. P. (2019). Primary
autism and Asperger’s disorder. Mol. Autism 8, 37. and secondary prevention of child sexual abuse. Int. Rev. Psychiatry 17,
Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences, 2nd Edn. 181–194.
Hillsdale, N.J: L. Erlbaum Associates. Landgren, V., Malki, K., Bottai, M., Arver, S., and Rahm, C. (2020). Effect of
Cohen, L. J., Ndukwe, N., Siegfried, R., Kopeykina, I., Yaseen, Z. S., and Galynker, gonadotropin-releasing hormone antagonist on risk of committing child sexual
I. (2018). Attraction versus action in pedophilic desire: the role of personality abuse in men with pedophilic disorder: a randomized clinical trial. JAMA
traits and childhood experience. J. Psychiatr. Pract. 24, 374–387. doi: 10.1097/ Psychiatry 77, 897–905. doi: 10.1001/jamapsychiatry.2020.0440
pra.0000000000000348 Långström, N., Enebrink, P., Laurén, E.-M., Lindblom, J., Werkö, S., and Hanson,
Conners, C. K. (2000). Conners’ Continuous Performance Test II (CPT II V.5). R. K. (2013). Preventing sexual abusers of children from reoffending: systematic
North Tonawanda, NY: Multi-Health Systems, Inc. review of medical and psychological interventions. BMJ 9:347.
Craig, L. A., Browne, K. D., Stringer, I., and Beech, A. (2005). Sexual recidivism: Laws, D. R. (2000). Sexual offending as a public health problem: a North American
a review of static, dynamic and actuarial predictors. J. Sex. Aggress. 11, 65–84. perspective. J. Sex. Aggress. 5, 30–44. doi: 10.1080/13552600008413294
doi: 10.1080/13552600410001667733 Lipsey, M. W., and Wilson, D. B. (2001). Practical Meta-Analysis. Thousand Oaks,
Dombert, B., Schmidt, A. F., Banse, R., Briken, P., Hoyer, J., Neutze, J., et al. (2016). CA: SAGE Publications.
How common is men’s self-reported sexual interest in prepubescent children? Martijn, F. M., Babchishin, K. M., Pullman, L. E., and Seto, M. C. (2020). Sexual
J. Sex. Res. 53, 214–223. doi: 10.1080/00224499.2015.1020108 attraction and falling in love in persons with pedohebephilia. Arch. Sex. Behav.
Eastvold, A., Suchy, Y., and Strassberg, D. (2011). Executive function profiles of 49, 1305–1318. doi: 10.1007/s10508-019-01579-9
pedophilic and nonpedophilic child molesters. J. Int. Neuropsychol. Soc. 17, Massau, C., Tenbergen, G., Kärgel, C., Weiß, S., Gerwinn, H., Pohl, A., et al.
295–307. doi: 10.1017/s1355617710001669 (2017). Executive functioning in pedophilia and child sexual offending. J. Int.
Effect Size Calculator (2020). Effect Size Calculator. Available from: Neuropsychol. Soc. 23, 460–470.
https://campbellcollaboration.org/escalc/html/EffectSizeCalculator-SMD5.php Morrow, E. P. (2019). Cognitive, affective, and general empathy in individuals
(accessed May 18, 2020). convicted of a sexual offense: a meta-analysis. Sex. Abuse 32, 883–906. doi:
Eke, A. W., Helmus, L. M., and Seto, M. C. (2019). A validation study of the 10.1177/1079063219858062
child pornography offender risk tool (CPORT). Sex. Abuse 31, 456–476. doi: Neutze, J., Seto, M. C., Schaefer, G. A., Mundt, I. A., and Beier, K. M. (2011).
10.1177/1079063218762434 Predictors of child pornography offenses and child sexual abuse in a community

sample of pedophiles and hebephiles. Sex. Abuse 23, 212–242. doi: 10.1177/ the development and validation of a structured diagnostic psychiatric
1079063210382043 interview for DSM-IV and ICD-10. J. Clin. Psychiatry 59(Suppl. 20),
Oakley, B. F. M., Brewer, R., Bird, G., and Catmur, C. (2016). Theory of mind is not 22–33.
theory of emotion: a cautionary note on the reading the mind in the eyes test. Spector, I. P., Carey, M. P., and Steinberg, L. (1996). The sexual desire inventory:
J. Abnorm. Psychol. 125, 818–823. doi: 10.1037/abn0000182 development, factor structure, and evidence of reliability. J. Sex. Marital Ther.
Reid, R. C., Garos, S., and Carpenter, B. N. (2011). Reliability, validity, and 22, 175–190. doi: 10.1080/00926239608414655
psychometric development of the hypersexual behavior inventory in an Suchy, Y., Whittaker, J. W., Strassberg, D. S., and Eastvold, A. (2009).
outpatient sample of men. Sex. Addict. Compulsivity 18, 30–51. doi: 10.1080/ Neurocognitive differences between pedophilic and nonpedophilic
10720162.2011.555709 child molesters. J. Int. Neuropsychol. Soc. 15, 248–257. doi: 10.1017/
Riegel, D. L. (2004). Effects on boy-attracted pedosexual males of viewing boy s1355617709090353
erotica. Arch. Sex. Behav. 33, 321–323. doi: 10.1023/b:aseb.0000029071.89455. Thibaut, F., Barra, F. D. L., Gordon, H., Cosyns, P., Bradford, J. M., and Wfsbp
53 Task force on sexual disorders. (2010). The world federation of societies
Santtila, P., Antfolk, J., Räfså, A., Hartwig, M., Sariola, H., Sandnabba, N. K., et al. of biological psychiatry (WFSBP) guidelines for the biological treatment of
(2015). Men’s sexual interest in children: one-year incidence and correlates in a paraphilias. World J. Biol. Psychiatry 11, 604–655. doi: 10.3109/15622971003
population-based sample of Finnish male twins. J. Child Sex. Abuse 24, 115–134. 671628
doi: 10.1080/10538712.2015.997410 Vellante, M., Baron-Cohen, S., Melis, M., Marrone, M., Petretto, D. R., Masala,
Schaefer, G. A., Mundt, I. A., Feelgood, S., Hupp, E., Neutze, J., Ahlers, C. J., C., et al. (2013). The “Reading the Mind in the Eyes” test: systematic review of
et al. (2010). Potential and Dunkelfeld offenders: two neglected target groups psychometric properties and a validation study in Italy. Cogn. Neuropsychiatry
for prevention of child sexual abuse. Int. J. Law Psychiatry 33, 154–163. doi: 18, 326–354. doi: 10.1080/13546805.2012.721728
10.1016/j.ijlp.2010.03.005 Walton, M. T., Cantor, J. M., Bhullar, N., and Lykins, A. D. (2017).
Schiffer, B., and Vonlaufen, C. (2011). Executive dysfunctions in pedophilic and Hypersexuality: a critical review and introduction to the “Sexhavior
nonpedophilic child molesters. J. Sex. Med. 8, 1975–1984. doi: 10.1111/j.1743- Cycle”. Arch. Sex. Behav. 46, 2231–2251. doi: 10.1007/s10508-017-
6109.2010.02140.x 0991-8
Schuler, M., Mohnke, S., Amelung, T., Dziobek, I., Lemme, B., Borchardt, V.,
et al. (2019). Empathy in pedophilia and sexual offending against children: a
multifaceted approach. J. Abnorm. Psychol. 128, 453–464. Conflict of Interest: The authors declare that the research was conducted in the
Seto, M. C. (2018). Pedophilia and Sexual Offending Against Children: absence of any commercial or financial relationships that could be construed as a
Theory, Assessment, and Intervention, 2nd Edn. Washington, DC: American potential conflict of interest.
Psychological Association.
Seto, M. C. (2019). The motivation-facilitation model of sexual offending. Sex. Copyright © 2020 Wittström, Långström, Landgren and Rahm. This is an open-
Abuse 31, 3–24. doi: 10.1177/1079063217720919 access article distributed under the terms of the Creative Commons Attribution
Seto, M. C., Cantor, J. M., and Blanchard, R. (2006). Child pornography offenses License (CC BY). The use, distribution or reproduction in other forums is permitted,
are a valid diagnostic indicator of pedophilia. J. Abnorm. Psychol. 115, 610–615. provided the original author(s) and the copyright owner(s) are credited and that the
doi: 10.1037/0021-843x.115.3.610 original publication in this journal is cited, in accordance with accepted academic
Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, practice. No use, distribution or reproduction is permitted which does not comply
E., et al. (1998). The mini-international neuropsychiatric interview (M.I.N.I.): with these terms.

ORIGINAL RESEARCH
published: 07 December 2020
doi: 10.3389/fpsyg.2020.606218
“All of Me Is Completely Different”:

Experiences and Consequences
Among Victims of
Technology-Assisted Child Sexual
Abuse
Malin Joleby 1* , Carolina Lunde 1 , Sara Landström 1 and Linda S. Jonsson 2
1
Department of Psychology, University of Gothenburg, Göteborg, Sweden, 2 Department of Social Sciences, Ersta Sköndal
Bräcke University College, Stockholm, Sweden
The aim of the present study was to gain a first-person perspective on the experiences
of technology-assisted child sexual abuse (TA-CSA), and a deeper understanding of the
way it may affect its victims. Seven young women (aged 17–24) with experience of TA-
CSA before the age of 18 participated in individual in-depth interviews. The interviews
were teller-focused with the aim of capturing the interviewee’s own story about how they
made sense of their experiences over time, and what impact the victimization had on
them in the short and long terms. Thematic analysis of the interviews revealed a broad
Edited by: range of abusive experiences that had profoundly impacted the individuals’ lives, health
Robert Horselenberg,
Maastricht University, Netherlands
and self-concepts. Three dominant themes emerged from the analysis – From thrilling
Reviewed by:
to abusive, Negative effect on health and wellbeing, and A new self after the abuse.
Anita Santos, From thrilling to abusive captures the wide range of experiences described, starting
Instituto Universitário da Maia (ISMAI), from the child’s own sexual curiosity to descriptions of having been manipulated or
Portugal
Matt DeLisi, threatened into engaging in sexual activity, as well as the sometimes long and complex
Iowa State University, United States process of understanding the severity of one’s experiences. Negative effect on health
*Correspondence: and wellbeing describes the victimization’s comprehensive impact on the life and health
Malin Joleby
malin.joleby@psy.gu.se
of the participants, how they blamed themselves for what had happened, and the
struggle of having to live with the constant fear of pictures from the abuse resurfacing.
Specialty section: A new self after the abuse depicts how the victimization impacted the way participants
This article was submitted to
viewed and thought about themselves in relation to others, and distorted their views of
a section of the journal their bodies. The findings are discussed in relation to previous research on both offline
CSA and TA-CSA, as well as theoretical and practical implications.
Received: 14 September 2020
Accepted: 13 November 2020 Keywords: technology-assisted child sexual abuse, victim, internet, online abuse, thematic analysis, experiences,
Published: 07 December 2020 CSA, consequences
Citation:
Joleby M, Lunde C, Landström S
and Jonsson LS (2020) “All of Me Is
INTRODUCTION
Completely Different”: Experiences
and Consequences Among Victims
In recent years, child sexual abuse that is conducted through the means of internet has
of Technology-Assisted Child Sexual received increasing attention in the media following several large-scale cases involving
Abuse. Front. Psychol. 11:606218. many victims, and numerous countries worldwide have reported an increase in cases
doi: 10.3389/fpsyg.2020.606218 (Bentley et al., 2019; Swedish National Council for Crime Prevention, 2019). In a

1 December 2020 | Volume 11 | Article 606218
Joleby et al. Experiences and Consequences of TA-CSA
ground breaking sentencing in 2018, Sweden became the first cases of online sexual abuse, as not all offenders aim for such a
country in the world to sentence a man for the rape of a child, meeting (e.g., Briggs et al., 2011; Joleby et al., 2020b).
for crimes that had been conducted solely online (B 11734-17). Online sexual victimization can take many forms, such as
The man, while located in Sweden, had coerced children in the online dissemination of abuse pictures (Martin, 2015), sexual
United Kingdom, the United States, and Canada into performing solicitation (Mitchell et al., 2001), online grooming (Whittle et al.,
sexual acts on themselves. Evidently, this type of technology- 2013), and sextortion (Wolak and Finkelhor, 2016). The focus of
assisted child sexual abuse (TA-CSA) is of global concern and this article is on TA-CSA in which children have been incited to
an issue that requires attention. Following the increase in TA- engage in online sexual activity. Such activity can include sexual
CSA victims worldwide, there is a significant desire to learn more chat, generating sexual photos and/or videos, performing sexual
about the phenomenon of TA-CSA. So far, research has largely acts live via webcam, or engaging in sexually humiliating activities
focused on offenders and offending behavior (e.g., Briggs et al., online (Whittle et al., 2013; Kloess et al., 2019; Joleby et al., 2020a).
2011; Webster et al., 2012; Kloess et al., 2014; Black et al., 2015; With regard to the relationship between offender and victim, TA-
De Santisteban et al., 2018), with attention more recently turning CSA includes victims who may have perceived the offender as
to its potential consequences. Although the existing knowledge being a romantic partner, as well as victims who have experienced
remains sparse, initial results indicate that the consequences of pressure and threats (Whittle et al., 2013; Wolak and Finkelhor,
TA-CSA can be as severe and harmful as offline CSA (Whittle 2016; Joleby et al., 2020b), and no offline meeting is required for
et al., 2013; Hamilton-Giachritsis et al., 2017; Jonsson et al., 2019; the situation to be considered TA-CSA.
Joleby et al., 2020a). Adolescents with experience of TA-CSA (and This article aims to provide a deeper understanding of the
no offline CSA) reported more trauma symptoms than a reference experiences of TA-CSA and its potential consequences. More
group, at least at the same level as adolescents with experiences specifically, the aim is to understand how victims of TA-CSA
of penetrative offline CSA (and no TA-CSA) (Jonsson et al., have made sense of their experiences over time, and what impact
2019). A wide range of psychological consequences similar to the victimization had on them in the short and long terms.
those reported among victims of offline CSA (e.g., psychological As this study was conducted in Sweden, a brief introduction
suffering, self-harming or suicidal behavior, sleeping problems, to the Swedish legislation is necessary. In Sweden, the age of
trust issues, impaired relationships, and difficulties at school) sexual consent is 15 years (Swedish Penal Code, paragraph 6).
were identified in court documents regarding victims of TA- Consequently, all sexual acts involving children under the age of
CSA (Joleby et al., 2020a). It has been suggested that the 15 are illegal. In accordance with the United Nations’ Convention
permanence of pictures of the abuse, which often exist in TA- on the Rights of the Child (UNICEF, 1989), individuals under
CSA, potentially complicate the impact of abuse even further and the age of 18 are defined as children. Thus, some sexual acts
can lead to increased feelings of self-blame (Hamilton-Giachritsis (if they harm the child’s health and development or if the child
et al., 2017). This argument resonates with a study showing that is under the care of the offender) between adults and children
knowledge of abuse pictures simply existing, or knowledge of between the age of 15 and 17 are also illegal (Swedish Penal
them having been distributed, was related to higher levels of Code, paragraph 6).
post-traumatic stress symptoms compared to being exposed to
undocumented CSA (Jonsson and Svedin, 2017).
The existence of a relationship between TA-CSA victimization MATERIALS AND METHODS
and psychological suffering is thus initially supported. In order
to gain a deeper understanding of what this relationship looks Recruitment
like and which processes are behind it, this study uses in- Recruitment to the present study took place between spring
depth interviews with victims of TA-CSA. The aim is to provide 2018 and spring 2020. Initially, the recruitment criteria were (i)
a rich first-person perspective on the experiences of TA-CSA having been subjected before the age of 18 to TA-CSA (ii) which
and the way it may affect its victims. This understanding is had been reported to the police. However, due to recruitment
vital in order to provide victims with sufficient support and difficulties, the second criterion was excluded after a few months.
help for coping with their experiences. Worryingly, there is a Recruitment efforts were all aimed at a Swedish audience and
common assumption that TA-CSA is a less severe form of sexual included (a) inviting lawyers (n ∼ 50) to send recruitment flyers
abuse. Research has shown that professionals demonstrate a about the project to their clients in recent legal cases of TA-
limited understanding of TA-CSA, may view it as less serious, CSA, (b) distributing recruitment flyers in waiting rooms at
and fail to prioritize its victims by minimizing their abusive youth clinics, youth centers, child and youth psychiatric centers,
experiences (Hamilton-Giachritsis et al., 2017, in press). Legally, psychologists’ receptions, and support organizations in several
these crimes are also considered less violating and result in lower municipalities, and (c) distributing digital recruitment flyers on
penalties for the offenders. Despite the worldwide agreement that social media via support organizations, spokespersons engaged
children should be protected from all forms of CSA (UNICEF, in issues of sexual abuse, and celebrities who are known to debate
1989; Council of Europe, 2007), 162 countries have failed issues of sexual abuse. The flyer included the recruitment criteria,
to criminalize sexual grooming of children online unless the brief information about the focus of the interview (“hearing
offender also attempts to organize an offline meeting with the about your experiences and how your health has been before,
child (International Centre for Missing and Exploited Children during, and after the abuse”), and information about the right
[ICMEC], 2017). This leaves children legally unprotected in many to anonymity. Flyers were aimed at both males and females. The

recruitment efforts resulted in nine people (all female) getting Linköping Youth Life Event Scale (Nilsson et al., 2010), and a
in touch to receive more information about the study. One of questionnaire specifying the type of online abuse they had been
them asked for information to forward to a friend, and one was exposed to and its context. They were given the choice to either
under the age of 15 and did not want to ask her parents for fill out the surveys on the spot, or to take them away and post
consent (which is a requirement according to research ethics their answers afterward (pre-stamped envelopes were provided).
guidelines in Sweden), and was thus not allowed to participate All participants except one chose to fill out the survey on the
in the study. The remaining seven were booked for interviews. spot. On leaving, participants received a pamphlet with contact
Six of the interviewees found out about the study through the information for support organizations working with victims
same celebrity (a female Swedish artist, author, and social media of sexual abuse.
personality), and the seventh participant did not remember
where she had found the information. Analysis
All interviews were conducted during summer 2019. The The aim of the study was to gain a first-person perspective
interviews took place at the interviewees’ choice of location, at on the experiences of TA-CSA, and a deeper understanding
either a library or a university in the chosen city, according to of the way it may affect its victims. Thus, qualitative methods
the participants’ requests. The first author – the only researcher with an explorative and descriptive approach fit well. The study
who knew the identities of the participants – had no prior employed a qualitative design by using thematic analysis, which
relationship with the participants and only met them on the one is “a method for identifying, analyzing and reporting patterns
occasion of the interview. The participants did not receive any within data” (Braun and Clarke, 2006, p. 79). The data was
compensation for their participation (but were offered water and analyzed inductively, which involves open-minded (theory-free)
fruit during the interview). exploration of data (Braun and Clarke, 2013), and generates
themes that are closely linked to the data without any theoretical
Interviews constraint. The analysis focused on the surface meaning of the
The interviews were based on the teller-focused interview method data, identifying themes on a semantic level.
(Hydén, 2014) due to its suitability for interviews about sensitive The procedures recommended by Braun and Clarke (2006)
topics. This method endorses a dialectical way of thinking about to ensure a rigorous thematic analysis were undertaken. All
the relationship between the interviewer and the interviewee, in interviews were conducted, audio-recorded, and transcribed
that they are two partners with different tasks and responsibilities verbatim by the first author, who thus had a high degree of
during the interview. The aim is to reduce the power imbalance familiarity with the data prior to the commencement of coding.
between the two, and to provide a relationally safe space in In order not to disembody the participants, but instead create
which the informant feels safe to share his or her story. The a fuller contextual understanding of their stories, the thematic
method is oriented toward narration and resembles unstructured analysis used a case-based approach in which each transcript was
interviewing (Brinkmann, 2014) in that it uses open questions. systematically coded separately.
Before starting the interview, the first author informed the Throughout the analysis, an iterative approach was employed
participants that the aim of the interview was to let them tell their by revisiting and adjusting existing codes, revising themes, and
story. The interview used open invitations like “Tell me about checking the fit with the original data (Braun and Clarke, 2013).
yourself and who you are” and “Tell me about what happened,” In order to understand how the participants made sense of
and open questions like “What was your life like before the their experiences and how they were impacted by the TA-CSA,
abuse?,” which were followed up with cued prompts (“You said X, the interview covered the lives and psychological health of the
please tell me more about that”), facilitating utterances (nodding, participant before, during, and after the abuse. Therefore, it was
“Umm,” etc.), clarifying questions (“Do you mean Y?”), and crucial to view each initial code in the context of the whole story
relevant follow-up questions. The aim of the interview was to of the participant. A table was created for each participant to
allow the participant’s story to develop naturally, but still make give an overview of all initial codes that were identified within
sure that all preplanned topics were covered (friends, school, that participant’s story before, during, and after the abuse. After
mental health, and family, etc.). Most questions were broad or the initial coding, each initial code was revisited and compared
open (see full interview guide in Supplementary Materials), to other initial codes within that participant’s story, in order to
and covered the life of the participants before, during, and look for relationships between them. To be transparent (Yardley,
after the abuse. 2008), we clarify with an example. One participant had an initial
The interviews lasted for 36 to 90 min, with an average code of “had a lot of friends” before the abuse and “almost no
interview time of 68 min (this is the time of the audio-recorded friends” after the abuse. After revisiting the original quotations, it
interview and thus exclude the pre-interview information and became apparent that this change in social network was due to the
the completion of questionnaires). Before the interviews, the participant no longer daring to trust anyone. This more general
participants received information about their participation (that part of the participant’s story was assigned the initial theme “Trust
it is voluntary, that they will remain anonymous, that they have issues.” Another participant had the initial codes “Easy to be
the right to withdraw their participation at any time, how the retraumatized”, “People are naïve”, and “The world is evil”, which
data would be stored, and reported), and filled out a consent were merged into the initial theme “Can’t trust anyone.” After
form. After the interviews, the participants were asked to fill creating one table for each participant, the initial themes were
out the Rosenberg Self-Esteem Scale (Rosenberg, 1965), the compared across participants. The initial themes “Trust issues”

and “Can’t trust anyone” (together with similar initial themes girls. The boy turned out to be much older than he had claimed,
from other participants) were merged into the overarching final and Beatrice found herself suddenly involved in a legal process.
theme “Difficult to trust people.” Clara stated that she was happy growing up, and that her
The coding and naming of themes for the first three transcripts childhood was nothing out of the ordinary. At the age of 13, she
were carried out by the first and last authors, who coded each fell in love with a boy in her school class. The boy took advantage
transcript separately and then created initial themes through of Clara’s feelings and managed to persuade her to send him a
discussion. The remaining four transcripts were initially coded picture of her in her underwear and to engage in sexual activity
and analyzed by the first author, and the last author verified that while chatting with him online. Six months after this event, he
all coding accurately represented the raw data through a process threatened to disseminate the picture of her unless she sent him a
of reading and re-reading all transcripts. The structure and nude picture within a few hours. Clara was scared, and did as he
finalized themes were set by the first and last author together, and demanded. Despite this, the boy disseminated the nude picture of
approved by the second and third author. The iterative process her to his friends at school.
of repeated discussions between two authors during the analytic Denise portrayed herself as a super-extrovert child who loved
process ensured that the findings were credible and dependable. being the center of attention and had lots of friends that she
Extracts from the interviews were used as illustrative examples to enjoyed hanging out with. While growing up, she suffered from
support the analytical claims (Braun and Clarke, 2013). anxiety and was sexually abused offline by her mother’s new
boyfriend. At the age of 11, Denise had online contact with a boy
Ethics she thought was her age. One night, he opened up his webcam
This research project was granted full ethical approval by and Denise was shocked to see a grown man masturbating. This
the Regional Ethical Review Board in (Linköping, Sweden). event really scared her. When Denise was 14, she met a girl online
According to Swedish law on ethical review of research (Swedish whom she fell in love with. Due to the previous event, she asked
law, 2003:460, Paragraph 18), participants above the age of 15 for pictures of the girl to make sure the girl existed. Denise and
have the right to consent to participate in research. Thus, no the girl chatted online daily for many months. Eventually, Denise
consent from the guardian or legal next of kin is needed. found out that the girl was in fact a man, who used the naked
pictures she had sent to blackmail her into sending more pictures.
Participants Emma explained that she grew up in a loving and caring
This study involves interviews with seven young women who family, and described herself as a really “good girl.” When Emma
self-identified as victims of TA-CSA. They were aged between 7 was young, she had good self-confidence, but grew more insecure
and 13 at the first occasion of the online abuse, and between 17 as she approached adolescence. When she was around 13 years
and 24 at the time of the interviews. All participants had been old, men started contacting her online. Although Emma felt
subjected to TA-CSA, and many of them had also been sexually a little disgusted, she also enjoyed the attention and the men
abused offline either before or after the TA-CSA. See Table 1 for managed to persuade her to engage in online sexual activity by
a summary of the participants’ experiences and survey answers. manipulating her with compliments and flattery. One of the men
To provide a context for the data (Levitt et al., 2018), a short communicated with Emma daily and convinced her that “love
description of each participant, based on how they described had no age limit.” Emma eventually met with him offline and was
themselves and their experiences during the interviews, follows. raped by him. Four years after this event, Emma was contacted by
All participants’ names have been changed to pseudonyms and the police who had found pictures and videos of her on the man’s
all identifying features removed to protect their identities. All computer during an investigation involving 70 other children.
quotes in the manuscript have been slightly edited to facilitate Frida reported that she grew up with supportive parents, but
understanding, and translated to English. that she was a shy and somewhat unsocial child who was teased
Anna was subjected to psychological abuse by her parents at school. Frida described her young self as worrying and having
while growing up. She explained that school was the only place a lot of anxiety, but online she could be anyone she wanted and
where everything was fine. She was involved in many after- therefore used the internet to find people to talk to. At 7 years of
school activities, and only had a few friends. At age 13, when age, men started taking the initiative for webcam sex, and Frida –
Anna’s grandfather had recently died, she turned to the internet who was initially curious and appreciated the affirmation it gave
to find someone to talk to. An older man initially listened to and her – was persuaded. During a period of about 6 years, Frida was
comforted her, but eventually demanded that she take off her repeatedly incited to engage in sexual activities with men online.
clothes and engage in masturbation. This was the first abusive Gabriella described herself as being an energetic and
experience that Anna encountered online. As Anna grew older, exuberant person who loves to be seen and heard, and has a great
she started engaging in online sex as a self-harming behavior, need for attention. At around 11 years of age, Gabriella spent a lot
which eventually resulted in her also being paid for sexual of time on the computer playing games, and found it exciting to
activities offline. talk to new people. One day she received a message saying “I miss
Beatrice described her childhood as being great, with many seeing you in underwear.” Although confident that she had never
friends to hang out with. When she was 12 years old, she engaged sent anyone pictures of herself, she was worried that the man had
in online sexual activity with a boy she thought was her age. Two hacked her webcam and she started chatting with him online.
years later, the police contacted her after finding videos of her The man, who claimed to be 19 years old, showered Gabriella
on the boy’s computer during an investigation involving 30 other with compliments and questions. The man managed to persuade

TABLE 1 | Characteristics and survey-scores across participants.
Namea Age at onset Length of Age at Number of Online sexual Experience of LYLES score Rosenberg
of abuse contact with interview online abuseb offline sexual nIPE/IPE/ACCd self-esteem
offender/s offenders abusec scale score
Anna 13 5 years 24 Numerous Penetration, Yes, i.a. by 2/6/4 Low (4/30)

sex in exhange online offenders
for money (sex in
exchange for
money)
Beatrice 12 2–4 weeks 19 One Penetration No info 5/1/1 Low (15/30)
Clara 13 3 years 17 One Penetration No info 3/1/3 Low (4/30)
Denise 11 1 month and 19 Two Penetration Yes 9/8/7 Low (12/30)
6–8 months
Emma 13 2 years 20 Numerous Penetration Yes, by one of 4/No info/No info Low (11/30)
the online
offenders
Frida 7 7 years 22 Numerous Penetration, Yes 8/2/6 Low (10/30)
sex in
exchange for
money
Gabriella 11 2 years 19 One Sexual Yes 9/3/6 Low (14/30)
conversations
a All
names are fictitious. b The most serious type of sexual act that the child was incited to perform of sexual conversation, sexual posing, masturbation, and penetration,
as well as whether they had received money in exchange for sexual activity. c The question regarding experiences of offline sexual abuse was not limited to before the
age of 18, thus the offline sexual abuse may have occurred any time during the participant’s life. d LYLES = Linköping Youth Life Experience Scale, 41-item questionnaire
measuring potentially traumatic life events. nIPE, non-interpersonal event; IPE, interpersonal event; ACC, adverse childhood circumstances. The figures indicate the
number of experiences reported for each category. For comparison, a community sample of 172 non-bullied Swedish females (age 15–20) reported on average 6.39
nIPE, 1.87 IPE, and 1.09 ACC (Nilsson et al., 2012).
her to engage in increasingly sexual conversations during their presented in the two themes Falling into the hands of the offender
2 years of contact, but she refused to send nude pictures even and Realizing the severity.
when he tried to force her. About a year after the contact had
ended, Gabriella was shocked to learn that 68 other children – Falling Into the Hands of the Offender
some of whom had been forced to engage in extreme and violent All participants had different stories about how they had come
sexual behaviors – had reported the man to the police. into contact with their offender(s), ranging from having their
curiosity exploited to being exposed to outright threats. Most
participants, however, described how they had been duped by
RESULTS the offender’s skills in manipulating them in different ways.
Some participants described the excitement of having someone
The thematic analysis resulted in three main themes – From showing an interest in them and making them feel seen and
thrilling to abusive, Negative effect on health and wellbeing, and heard: ‘It was a very sensitive age. And coming into contact
A new self after the abuse – with two, three, and three subthemes, with someone was really big, almost thrilling in a way. I was
respectively – see Table 2. Each theme is presented below, very excited that “Oh, someone wants to talk to me”.’ Several
supported by illustrative quotations and relevant contextual participants also explained their curiosity and excitement at
information. Please note that participants sometimes used engaging in something previously unexplored. Beatrice explains:
diminishing and mitigating phraseology when talking about their
abusive experiences. Nonetheless, when an adult engages in “It was super exciting! It was. . . well, about the time where you
sexual activity with a child, it is always the adult’s responsibility. start to explore yourself and stuff. So it was really cool, but also. . .
top secret. I knew that mom and dad absolutely could not know
The participants self-identified as victims of TA-CSA, and
about it. So it was really exciting, but also extremely important
according to Swedish legislation their experiences would be
that no-one would find out.”
deemed to constitute sexual abuse.
A common theme in the interviews was that the contact
From Thrilling to Abusive initially gave the participants a self-esteem boost and the
The analysis of how the participants made sense of the abuse attention they longed for, which made them overstep their own
revealed that their view and understanding of their experiences boundaries. Emma explained how the flattery prevailed over
had changed significantly over time. The participants had been her doubts: “Since this person had showed an interest in me
enticed or lured in different ways into engaging in online sexual and believed I was cute and hot and all of that, I kind of
activities, and most of them saw their abuse in a different light in wanted to do it, to get this affirmation, all the time.” When
retrospect than they did at the time of the abuse. The results are she expressed regret or doubts about the sexual activities, the

TABLE 2 | Result from the thematic analysis (Braun and Clarke, 2006) of interviews with seven victims of technology-assisted child sexual abuse (TA-CSA) about their
experiences and how the victimization affected them. A brief description of each sub-theme is provided.
Themes Brief descriptions
From thrilling to abusive

Falling into the hands of the offender All participants had in different ways been allured, manipulated, or forced into engaging in online sexual activities, although
some participants expressed that they at the time believed they voluntarily engaged in the sexual activities.
Realizing the severity Most participants had a different understanding of their abuse looking back at it, as they first in retrospect understood the full
severity of what they had been exposed to.
Negative effect on health and wellbeing
Everything collapsed Despite having different backgrounds, different experiences, and having their abuse revealed in different ways, all participants
described having been negatively affected by the TA-CSA. Many of them suffered consequences that largely had affected,
and still did affect, their lives and wellbeing, even though several years had passed since the abuse ended.
Self-blaming The self-blame that many participants described had several sources. From shame about having been sexually curious or for
taking an active role in the sexual activities, to shame for not having been able to shield themselves from the online abuse.
Fear of pictures resurfacing The reality of, or the fear for, pictures being disseminated caused additional stress and led the after-effects of the online
abuse to continue long after the abuse ended. This affected the psychological well-being of the participants for a long time,
and for some it also affected the decisions they took regarding their future careers.
A new self after the abuse
Trying to make sense of who I am Some participants struggled with understanding who they were after the abuse. They believed that the victimization had
fundamentally altered them as persons, and thus created a discrepancy between who they really were, and what themselves
and others thought of them.
Difficult to trust people For some of the participants, the TA-CSA had created a distrust in people and a skepticism of the good in others. This could
result in participants limiting themselves or the people they surrounded themselves with.
Distorted view of my body The abuse affected most participant’s relationship to and feelings toward their bodies in several ways. From leading to
difficulties to view or enjoy their bodies or sexualities, to a failure to listen to their own boundaries or bodily needs.
offender showered her with compliments and manipulated her ‘I remember I was crying, but for each thing I did, he said “Now
into thinking that their relationship was genuine and that she I have a picture of that.” [. . .] I didn’t know how my parents would
could say no if she wanted to. Gabriella also described how she react. Would they be able to talk to me? Would they be able to look
was initially duped by her offender, who became increasingly at me if they found out what I had done? So I just did as he said,
aggressive throughout their contact. and he masturbated on cam.’
“I thought it was love, that he cared about me. But that was not Gabriella also received threats about having pictures of her
really the case, he wanted something completely different. But in uploaded to the internet. When Gabriella started to realize that
my world, I guess I was naïve and believed he wanted to be my there was something devious about the “boy” she had been
friend, and that he cared about me. [. . .] And it’s easy for an 11- chatting with online and tried blocking him, he became angry.
year-old to be naïve.”
Although she was aware that the man did not have any naked
pictures of her, she nevertheless felt threatened:
Several participants were exposed to explicit threats. Anna,
who was extremely upset after her grandfather had died, went
“Okay, he does not have any pictures, but what happens if he
online to find someone to talk to. She started chatting with a 60-
writes something, or manipulates a photo. [. . .] I thought that
year-old man on a chat roulette site, whom she initially thought people I don’t know very well on Facebook might see it and not
was just being kind to her. understand that it is photoshopped, that he had done something
to the picture. People would think it was me. So I was pretty
“To begin with, he was very. . . considerate? He asked why I was scared.”
crying and what had happened. [. . .] I felt pretty safe with him,
and he let me cry. And then he asked me to remove my sweater,
and I said “No, I don’t want to,” and he said that if I didn’t do it he Instead of having the threats directed only toward herself,
would stop talking to me. [. . .] I think I felt, I really needed that, Emma described how one man threatened to commit suicide
to be listened to. So I did as he said. And he continued to talk to if she did not comply with his demands. Emma explained her
me as if nothing had happened, as if it wasn’t weird. So I kind of reaction: “I panicked. Like shit, what if, what have I done then?
forgot about it.” If someone finds this [the chat logs], they might think I have
kind of murdered someone.” Evidently, the threats made by the
The man, however, took advantage of Anna’s trust and took offenders were extremely effective. Anna, however, explained that
a screenshot of her. He then threatened to search out her the threats would not have worked on her today: “He said he
IP address, find her parents, and send the picture to them knew where I lived. Of course he didn’t, I mean now I understand
unless she engaged in different sexual activities. The man had a that he couldn’t know. But I didn’t really understand that at
hold on Anna: the time.”

Realizing the Severity had negatively affected their health and wellbeing in both the
Between four and 15 years had passed between the (first occasion short and long terms. This is presented in the three themes
of) TA-CSA and the interview. For most participants, the passing Everything collapsed, Self-blaming and Fear of pictures resurfacing.
of time had changed their perspective, and when they looked back
at their experiences now, they had come to the realization that the Everything Collapsed
situation was not what they initially thought it was. For instance, All participants had different psychological baggage before the
sexual activities that at the time felt relatively positive and online abuse occurred, and they all described their upbringing
exciting could in retrospect be viewed differently. This changed and mental health differently. Beatrice, Clara, Emma, and
view could be due to participants growing older and gaining Gabriella all said that their childhood prior to the online abuse
perspective. Emma said: “You do not have the ability to judge was good and safe, and none of them mentioned any previous
the consequences of your actions when you’re thirteen, fourteen, psychological problems. By contrast, Anna, Denise, and Frida
fifteen. You don’t see the consequences, that what you are doing described suffering prior to the online abuse. Nevertheless, it was
is wrong, until maybe afterward.” Frida initially believed that the clear that all the participants were negatively affected by the abuse
sexual contact with men online offered her experiences she would specifically, regardless of their previous mental health. Anna, who
otherwise miss out on, and she appreciated the gifts and money grew up in a dysfunctional family, said:
she received. She described communicating with other young “Well I believe I was feeling. . . yeah, yeah I was already feeling bad
girls who were engaged in similar behaviors, and that she was (laughs) before. Or I was. . . lonely. A bit sad. Thought everything
caught up in it. This made it difficult to have a balanced view of was a bit hopeless. But afterward. . . it feels like everything became
what was going on, and it was only afterward that she realized the much more precarious. It wasn’t that I was completely naïve
men had taken advantage of her: “The thing is, at the time you about people being mean, and such. But it went from me being
don’t really consider what is going on [. . .], but afterward you comfortable with the feeling of sadness (laughs), to me being
realize how sick it was.” completely turned off, and I started dissociating quite a lot.”
For other participants, the changed view of the abuse occurred
Similarly, Denise mentioned that she had suffered from
when they found out that the person they had engaged in sexual
anxiety as a child, but that she still lived a functional life before
activities with had lied about who they were. This could turn
the abuse changed everything:
an activity that they had experienced as seemingly mutually
consenting, with someone they cared about and who they thought ‘I talked less to people, I was making less contact with people, I
cared about them, into an experience of betrayal and abuse. didn’t use social media. I turned off most things. I had extremely
Gabriella, who had been chatting online for about 2 years with high absence from school, and the school was like “You will lose
what she thought was a boy who was a few years older, was all your grades”. Everything just fell apart for me. I had almost no
shocked to find out about his true identity. friends. I ate my own emotions and felt like crap. I was scared and
tired of trying.’
‘It evoked quite a lot of feelings, because I felt, I think it was a slap
to the face, like “Oh, this is what has happened,” kind of. To just For Denise, Anna, and Frida, who were already struggling
get all this served at one time was really tough, [. . .] to find out for other reasons, the online abuse made everything collapse
everything, his identity, all the other people that had been abused, and they became suicidal. The other participants, who seemingly
and stuff like that.’ lived relatively untroubled lives before the abuse, also described
different forms of psychological suffering. Beatrice described how
Some participants were abused by numerous different people she was feeling very well and had a great upbringing, but that this
over the course of several years. While they experienced many of suddenly changed when she developed depression following the
the situations as abusive and coercive, it was only in retrospect realization of her abuse 5 years ago:
that they understood the full extent of the consequences it would
lead to. What at the time might have appeared to be their own “I haven’t been genuinely happy for a longer period since before
choice, had in fact been them being manipulated and used by the whole police interrogation thing. Of course, I have happy
people online. Emma explained how she realized in hindsight moments and have had them for several years, but it has been
that “this person was not feeling well” when she looked back at several years since I was feeling good and happy in general.”
herself. “That’s when it started to catch up with me, that I started Aside from general depressive symptoms, sleeping problems,
to understand what I had been exposed to, and everything that and anxiety, feelings of shame could also negatively affect the
had happened. Yeah, everything caught up with me.” Several participants’ wellbeing. Emma mentioned the profound effect
participants described how they felt repulsed when they thought shame had on her:
back to what they had been incited to engage in. Anna explained:
‘I felt very disgusted by myself and by people in general. Both “It was all the shame. It has like, kind of stopped me. This, all this
what I had been asked to do, and that I had been able to do it to shame. I was so extremely ashamed. And that is kind of, shame
myself.’ is maybe the feeling that is most difficult to bear, because it really
paralyzes you. It affected how I wanted to live.”
Negative Effect on Health and Wellbeing Some described that the initial abuse taught them to ignore
Although some participants took a while to realize the full extent their own will, which resulted in them developing self-harming
of their experiences, all participants described how the TA-CSA behavior which allowed other people to take power over their

bodies. Anna explained: “I just did what people told me to. my long walks and feel safe in that. So it varies a lot. But. . . it feels
And afterward, I couldn’t really understand what I was doing. I like it’s getting better.”
remember that. That I had anxiety about it, like why do I do this?
Why can’t I just quit?” Frida also described the negative spiral of It is important to highlight that one interview stood out from
destructiveness that followed the online abuse: the rest. Despite describing that she had indeed been negatively
affected by the abuse, Gabriella, who had been encouraged to
“Between the ages of 12 and 14, I really noticed that I, kind of, it engage in sexualized conversations, also stated that it had not
sounds so sick, but I felt that I kind of wanted to sell my body, affected her too much: “Of course, it might have shaped me, but
thoughts like that. And I don’t know why I felt like that, it was not very much, not so that it has left deep traces on me, but
more that I felt like a failure, I would never find someone, so I perhaps to some extent.”
might as well just do this, and then I could kill myself. [. . .] The
contact with men online made me feel so bad, you kind of lose Self-Blaming
your self-esteem, and sense of reality.”
A common theme was to put the blame on themselves for what
All participants experienced negative consequences due to the they had been subjected to. For some participants, this was due to
abuse, but the trauma appeared at different times and in different them believing that their sexual curiosity was something taboo,
situations, as described in a previous subtheme (Realizing the and that the activities they engaged in were inappropriate. Emma
severity). In contrast to Frida and Anna, who described the TA- explained that she kept her online contact a secret, “so in some
CSA as having been disgusting or traumatic, some talked about way I kind of knew it was wrong,” and Frida expressed that her
the online sexual activities in more neutral or positive terms, behavior was “abnormal.” For others, the self-blame stemmed
and the real trauma seemed to have come afterward when they from having been fooled. Due to having been misled online
realized that they had been tricked or deceived, or when they before, Denise asked for photos and an address to make sure that
realized that other people had knowledge of the abuse. Beatrice the girl she was chatting with online was who she claimed to be.
explained that the sexual activity she engaged in with what she When she realized that she had been deceived by a man despite
thought was a boy her age was “nothing I really thought about. these precautions, she berated herself:
It happened and then I moved on, kind of. It was nothing I “I felt that I was so very stupid for having done this, like how can
gave much attention to, it was nothing I kept thinking about. you trust someone you don’t actually know whether you know?
Until this legal process.” Being thrown into a legal process and [. . .] I felt so very mad at myself for doing it, and often felt that
becoming aware that the sexual conversations, photos, and videos if something happens now, it’s still my fault to some extent. I’ve
she had shared were now being viewed by others was traumatic. contributed to it, erm, so I guess I only have myself to blame.”
After the first police interrogation, Beatrice described “three to
four years of pure shame about myself.” It seemed as if a large Emma said that she did not dare to tell her friends about the
part of the trauma could stem from the social shame of having abuse she was subjected to, as it would reveal her own stupidity:
a secret revealed, and not only the realization that one has been “I remember thinking that this would never have happened
subjected to abuse. Similarly, Emma described her reaction to the to them. It is so typically me. Because they are so good and
first phone call from the police, in which they explained that they perfect and all that. They would 100% understand that this was
had found pictures and videos of her on a suspect’s computer: a pedophile you were talking to, but I’m stupid, and I didn’t
understand that.”
‘And I was like “Okay what kind of pictures?” And they were like “I
think you know what kind of pictures,” and I was like “Well okay Another factor for self-blame was that the sexual abuse
yes I know.” And then the world fell apart, I was extremely sad, I was conducted online and that there was no physical meeting
panicked.’ between the participants and their offenders. Some participants
expressed that they should therefore have been able to shield
For almost all participants, several years had passed since the
themselves from the abuse, and that if they had just turned off
online abuse ended, but nevertheless, many of the participants
their computers, “none of this would have happened.” Emma
described thinking about it a lot, struggling with ongoing
explained:
depression, undergoing psychiatric evaluations, only being able
to go to university part-time, or being in the midst of recovery. “Since I was behind a computer screen all the time, I kind of,
However, despite still struggling, most participants described a afterward I thought I had myself to blame. [. . .] It has led me to
positive development and most saw some kind of brightening on think that I have not been subjected to anything, but it is like I
the horizon. Frida said that she had not been thinking about it a have subjected myself to it.”
lot during recent years, and Beatrice explained that she had been
able to let go of the feelings of shame and instead felt anger: “I am Similarly, the participants had to take an active role in their
mad. I am extremely mad. [. . .] and not only at this man, but. . . abuse, for instance by following directions or demands from their
at society as a whole.” Denise described her psychological health offenders, or by playing along with their sexualized conversations.
as going up and down: Frida was involved in a legal process, during which she had to
listen to her conversations with the offender being read aloud in
“Some days are horrible, and I don’t want to get out of bed. Other court. She described how she felt guilty when she realized that she
days I can go outside for a run. And just be out in nature, and take had also been taking the initiative.

‘I knew it was wrong somehow [the sexual activities], but I listened A few participants also expressed concerns that they had done
to what he said. [. . .] I’ve repressed this so much, that I’ve been a something illegal when producing the pictures. When Beatrice
driving force in this, but when I read, when I sat in the courtroom was summoned to the police to be questioned, she was assigned
and had it read to me, I got like a shock, like “shit,” “what the hell”.’ a lawyer. This convinced her that it was her who had done
something wrong. Anna said that she was concerned someone
Most participants expressed initial self-blame, but as more
would report her to the police:
time passed after the abuse, many came to realize that the guilt
was not theirs but the offender’s. Despite this realization, some “I had read somewhere that it counted as producing child
struggled to rid themselves of the feelings of shame. Gabriella pornography if you uploaded pictures or videos of yourself if you
said that she is aware in hindsight that it was not her fault, but were under 18, and I remember being scared about that as well,
nevertheless said: that someone would find the pictures and report me.”
‘I still feel an underlying sense of shame, even if I don’t want it to Another traumatizing aspect of having the abuse documented
be like that. It’s still not something that I would like to talk about. was described by Beatrice and Emma, who had gone through legal
[. . .] It might be that this shame is still in me today.’
processes in which their videos and chat logs were scrutinized.
Since the videos were part of the evidence, Emma had to sit
Fear of Pictures Resurfacing outside the courtroom knowing that everyone in court was
One of the major causes of anxiety was the existence of pictures watching videos of her touching herself.
or videos of the abuse. Anna was confronted by a colleague who
recognized her from videos online, and this caused her such ‘It’s so surreal, and it’s so very humiliating having to walk back in
extreme stress that she did not return to that job. It is important to and sit down in the chair, look someone in the eyes and confirm
emphasize that it was not only when pictures were disseminated “Yes, that is me in the video.” So it has been extremely strenuous.
that anxiety or distress occurred. Just the fear that someone could Partly due to knowing that there are people out there that you
see the pictures or videos was enough to cause concern, and don’t know, you don’t know who has these pictures or where
numerous participants described worrying about who had seen they end up, but also, I don’t know, having to live with the
facial expressions, the looks you received walking back into the
them. Several participants described the existence of pictures as
courtroom knowing what they have just watched.’
something of a ticking bomb; “I always have to be prepared” and
“this will eventually come back to me.” Clara explained how she
deleted her whole Facebook profile because she was so worried A New Self After the Abuse
that the picture would be posted there: “I was so scared that In addition to the psychological consequences caused by the
my parents would see, and I didn’t want proof that I had done abuse, the interviews revealed that the victimization also affected
this.” Denise described how she was “just sitting at home, always some participants’ self-concept and how they acted. This is
indoors, waiting for something to happen. [. . .] I was scared presented in the three themes Trying to make sense of who I am,
to death.” The fear was also present out in public, as Emma Difficult to trust people, and Distorted view of my body.
exemplified:
Trying to Make Sense of who I Am
‘Every time an older man, or a man, gives me more than a glance
During the interviews, participants received the open invitation
when I walk past, I automatically think “This person has seen these
“Tell me about yourself and who you are.” The answers revealed
videos, this is a pedophile.” It’s kind of sick to think like that.
I don’t do it as frequently anymore, but when I was younger I that some participants struggled with their views of themselves
thought like that right away, and I kind of panicked.’ and how this contrasted with the views others had of them, or
how they themselves wanted to be seen. More specifically, Anna,
For Anna, this worry caused her to develop a fear of men, Denise, and Emma all described themselves as being viewed as
resulting in her not being able to buy groceries for 2 years “good girls” and that their online victimization threatened that
if the cashier was a man. Furthermore, the knowledge that façade. Denise explains: “I was always portrayed as such a good
compromising pictures or videos might exist could have far- child, because school went really well, and I always did well. And
reaching consequences in some of the participants’ life decisions. then this happens. That is not who I am. And all of a sudden I
Emma used to think that her future was ruined: “[. . .] because, if am bad.” They described the online victimization as something
I was to succeed with anything, and in any way become a public wrong or immoral that they themselves were responsible for. Not
person, these videos would resurface. And then the whole world living up to the high expectations of them created an internal
would see.” Similarly, Anna described that she had always wanted incongruity that made them question themselves. Anna explained
to be an author. At one time she was nominated for a literature how it was very difficult to cope with the discrepancy caused by
award and had a real chance of becoming an author, but she the fact that “one side of me was subjected to abuse, and one side
realized: “No, that will not happen. I cannot have my picture was functioning as usual.” Anna was very active in student union
and my name linked to something that I am not sure what will projects at school, had a job outside school, and had excellent
happen to, that is public.” She also described feeling that nothing grades. She explained: “Everyone viewed me as proper and high-
online ever really ceased to exist, “so of course I am worried, but functioning [. . .] and that was the image of myself that I liked. I
(sighs). . . I try not to think about it too much or else I would not was scared that someone would find the pictures or videos and
be able to have a life (laughs).” come to realize who I really was.”

Clara, who was abused by a boy at her school, described portrayed how her worry had also changed her in more ways than
how the boy also manipulated her into treating her friends and minimizing her network.
brother badly. This had affected Clara’s self-esteem, since she now
“I used to be a very outgoing person. I was. . . I was always the
believed that she could not “know what is right and wrong in a
center of attention. [. . .] I loved to be seen. I loved to be heard,
relationship.” She described how the victimization became a sign and I would put on talent shows and sing and dance and show
of her character: “I have a very hard time accepting myself, like off. [. . .] Now I don’t want to be out around people. [. . .] I find it
my personality above all. Because all I do is just wrong and mean very hard to be seen. I don’t want to be the center of attention. I’m
and evil.” Likewise, Anna also questioned her own character. terrified of attention.”
When she was younger, her self-hatred revolved around her
looks, her geekiness, or things she was teased for at school. Now, Anna, who had been sexually abused by many men over a
her self-hatred evolved into questions like: “Am I a good person? period of several years, also struggled with trusting people. She
Am I dangerous? Am I ruined? Am I broken? Things like that.” said that she had lost her faith in society and that it was so easy to
She continued: re-traumatize herself by just going online and confirming that the
world was an evil place. This made it difficult for her to have any
‘I look at myself like an “after” now. Before, I used to view myself friends, and she described the isolation as the most challenging
as a whole person, that was just what I was, but now I feel like I’m part:
living in some kind of closing credits to something, and so I have
to try to make something good out of it. As if all the important “It’s difficult to explain to people why I don’t trust others. It feels
things have already happened.’ like people in general have a very. . . naïve view of the world
(laughs). And they think that people are kind and good, and that
Some participants described how being lured into engaging there are specific indicators of what a kind and good person is,
in TA-CSA affected their self-esteem and self-confidence, and and that they could never know anyone who would be able to. . .
Denise described how she struggled with coming to terms with (sighs) hurt a child for instance. And that makes it difficult. . . erm,
the fact that the abuse had made her a different person: “It’s tough to relate to other people. I feel very abnormal. . . and I think that
to realize that. . . that it has affected you so much. More than you that is the hardest part, because I have had to come to terms with
the fact that I might have a future without many people around
think. Because all of me is completely different.”
me, and figure out how I can make that feel valuable.”
Difficult to Trust People Distorted View of My Body

During the interviews, it became apparent that the online abuse All the participants except Gabriella reported that their
had led some of the participants to lose trust in other people. victimization had affected how they viewed their body or how
This manifested itself in different ways, and affected their lives to they thought about their sexuality. Beatrice described how she
different extents. For Gabriella, Beatrice, and Denise, it resulted in used to think of her body as a tool before the abuse:
them being very careful online with private accounts and cautious
about what kind of information they shared about themselves ‘Before it happened, I didn’t reflect on my appearance, it was my
online. Gabriella said: “I don’t trust many people on the internet. body and I needed it to go to school and to do stuff. [. . .] What
It takes quite a lot for me to feel like, okay, this is not a fake person I’m working on most these days is what I call “clothing anxiety.”
but someone who is decent and does not want to hurt me.” She Picking an outfit in the morning takes at least an hour, and half of
the time it results in me not being able to leave.’
described how she used to be scared of being abused again, but
that it had also taught her to be less naïve and easy to fool, which Denise talked about how she did not want to see herself
she viewed as some kind of positive. Denise also expressed that naked and could not look at herself in the mirror without feeling
the victimization had changed how she operated online: ashamed. As a result, she used to cover up as much as possible:
‘I get so paranoid always. As soon as I receive a request to follow “I wore XXXL in everything, because then you can’t tell what I
me online [. . .] I ask my friends “Do you know this person? Do look like. Eh, and I’ve always thought that if people don’t see what
you know who it is?” [. . .] I keep everything private. I don’t upload I look like, then I can’t be accused of contributing to whatever
anything that I’m not 100 percent comfortable with. I use the might happen.”
block button frequently (laughs).’
At the time of the interview, several years had passed since
Both Beatrice’s and Denise’s distrust in people online also the online abuse ended. Denise described how things were slowly
spilled over into their life outside the internet and affected which getting better, and that she was learning to dress more the way she
people they surround themselves with. Beatrice explained that she wanted to:
has chosen friends that “she can really trust.” Similarly, Denise
described having a hard time letting people into her life, and “It feels fantastic to finally not only look at these things that you
want to wear, but also to wear them and show them off. But also
“didn’t dare to trust anyone else” after the abuse. Before the
that I can undress in front of my partner [. . .] and we can have
abuse, she described herself as being “the one everyone wanted sex. These steps have been tough.”
to hang out with. And I wanted to be with everyone (laughs). I
loved hanging out with people.” By contrast, she described her These extracts demonstrate how the abuse could lead to a
current social network as going “from having had everyone and distorted view of, and guilt about, one’s appearance and sexual
anyone, I now have three real friends, and my family.” Denise desire. Denise explained that she used to be open about her

sexuality, but that the abuse led her to feel that she did not believe that they were communicating with a peer. Thus,
want to and should not enjoy sex after what she had been for some participants, the realization of abuse was not an
through. With other participants, having been exposed to other inner insight, but occurred when police contacted them and
people’s sexualization at a young age seemed to have blurred the revealed the true identity of the offender, which could be a
boundaries around their own sexuality and the right to their shocking experience. TA-CSA offenders are, however, not always
bodies. For Clara, Emma, and Frida, the abuse caused them deceptive (Wolak and Finkelhor, 2013). In these cases, the
to overstep their own limits and lose track of their self-respect manipulation and psychological grooming (Craven et al., 2006)
for their bodies. Clara described how the abuse made her think that the victims were subjected to led them to gradually gain
that “any guy can decide over my body in some way.” Emma an understanding of the true nature of their experiences as
explained how she did not feel like she owned her own sexuality they grew older and gained perspective. For the participants,
because she had always been told what to do. This led her to the boundaries for when an experience was considered abuse
agree to things that she did not really want to take part in online, could evidently be blurred. This can have implications for
and she also allowed guys at school to touch her body because professionals meeting this victimized group. Therapists and
“the attention I received from guys kind of meant more to me support workers should consider what impact the involvement
than how I was feeling.” Frida said that the online abuse was of technology might have, and should be aware of the possibility
what caused her distorted view of sex, which led her to have a that young people may have difficulties understanding their
dysfunctional relationship with men and to develop feelings of experiences as abuse. For the same reason, law enforcement
not being worth anything. should be cautious when approaching children whom they
Anna, who developed self-harming behavior in which adults suspect have been victims of TA-CSA, in order not to cause
paid her for sexual activities, described how all her abusive the victim any additional trauma in connection with the
experiences had led her to shut off her body: disclosure, as the child themselves may not be aware that they
have been exploited.
“I don’t think about it. At all. Most of the time, if it does not give
me discomfort, umm. . . I forget that it exists, which means that I
Participants often expressed that they had initially been
do not reflect on the fact. . . that it’s cold outside. I feel very little excited and sometimes part of inciting the sexual activity.
physical discomfort.” Considering that sexual curiosity is a significant aspect of
development (e.g., Kastbom et al., 2012), and that digital advances
have led many young people to explore their sexuality online
DISCUSSION (Valkenburg and Peter, 2011; Anastassiou, 2017; Madigan et al.,
2018), this initial excitement is not a deviating trait. Offenders
This study’s findings, from in-depth interviews with seven taking advantage of children’s natural curiosity, however, seems
young females who self-identified as victims of TA-CSA, to impact on the self-blame experienced by the victims, and
build upon previous research showing that TA-CSA may have constitutes an additional obstacle to realizing they had been
severe consequences (Hamilton-Giachritsis et al., 2017; Jonsson exploited. The fact that the abuse took place online (“should
et al., 2019; Joleby et al., 2020a) and provide an insight have been able to turn off the computer”) and required the
into the underlying processes between TA-CSA victimization participants to be active in the acts (“felt like I subjected
and psychological suffering. Despite the participants’ different myself to the abuse”) led to further self-blame. Professionals
experiences (from being abused by a single offender, to several have also noted that victims of TA-CSA are more often blamed
years of repeated abuse by numerous offenders), they all by others and seen as participating in the abuse than victims
provided detailed accounts of how their victimization had a of offline CSA (Hamilton-Giachritsis et al., 2017). Higher
negative effect on their health and wellbeing, not seldom of levels of self-blame among CSA victims have, in turn, been
extremely serious proportions. The abuse impacted on several associated with increased psychological distress (Coffey et al.,
aspects of their lives, such as their relationships with others, 1996). This illustrates the importance of professionals dealing
their self-respect, and their ability to cope with everyday life. with abused children working to counter feelings of blame
Research has shown CSA to be predictive of internalizing (Hamilton-Giachritsis et al., 2017).
outcomes (Muniz et al., 2019), and many of the consequences Another consequence of the abuse being conducted via digital
described in the interviews match those that several decades technology was the constant fear of pictures resurfacing. This
of research on offline CSA have reported, namely general fear was one of the major causes of anxiety and impacted the
depressive symptoms, re-victimization, sexual problems, anxiety, lives of the participants in several ways, contributing to the long-
poor self-esteem, and interpersonal problems (Paolucci et al., term effects of the abuse. Always having to be prepared to be
2001; Maniglio, 2009). confronted and fearing one might be recognized out in public
The impact of the abuse could be both direct and delayed, contributed to the never-ending aspect of the abuse, which has
depending on the participant’s understanding of the abusive previously been reported among victims of online abuse images
situation and the time taken to realize its severity. Understanding (Leonard, 2010). In sum, these findings lend support to the idea
that one’s experiences can be labeled as sexual abuse can that some aspects of TA-CSA complicate the impact of the abuse
be a long and complex process (Hjelen Stige et al., 2020), (Hamilton-Giachritsis et al., 2017).
and the online element may add extra complexity as it Consequences following CSA are known to vary widely among
allow offenders to hide their identity, leading the victims to victims (Maniglio, 2009), which was also evident in this study.

In line with research showing the cumulative negative effect Another way of explaining both the short- and long-term
of numerous traumatic experiences (Felitti et al., 2019), the outcomes following CSA is through the potential damage to self-
participants who reported psychological suffering and trauma concept that may occur in response to victimization. Broadly,
previous to the TA-CSA seemingly experienced the most severe self-concept refers to the way individuals think about, evaluate,
impact following the abuse (e.g., suicidality). The participant and perceive themselves (Baumeister, 1999), and one of its
with the least extensive abusive experience (being enticed to core aspects is self-esteem. Self-concept is proposed to have
engage in sexual conversations but managing to refuse to send a mediating role between CSA and its negative outcomes on
nude pictures, while the other participants had been incited psychological health, and the relationship between CSA and
to penetrate themselves) stood out from the group in that she negative self-concept and self-esteem has been theoretically and
only reported limited suffering. Whether this is due to the empirically explored (Kendall-Tackett et al., 1993; Stern et al.,
type of abuse she had been subjected to cannot be determined. 1995; Turner et al., 2010; Cantón-Cortés et al., 2012; Lamoureux
It may instead be that some individuals are not severely et al., 2012; Halvorsen et al., 2020). This study, in line with studies
affected by TA-CSA, as research on offline CSA has shown on offline CSA, showed that the abuse influenced the ways in
(Kendall-Tackett et al., 1993). which participants thought about and understood themselves, as
This study identified both immediate and long-lasting well as how they thought others would perceive them. The time
negative impacts on the psychological health of the participants period during which the participants were subjected to TA-CSA
following the TA-CSA. One way of explaining how and why is an important time in regards to sexual development (Diamond
CSA can result in both short- and long-term consequences and Savin-Williams, 2009), fitting in and being accepted by
is by using the four traumagenic dynamics model (Finkelhor peers concerns (Steinberg, 2011), and the development of a
and Browne, 1985). The model suggests that the experience of positive self-concept (Berger, 2018). Therefore, negative sexual
CSA changes the child’s cognitive and emotional orientation experiences during this sensitive period may have particularly
to the world, by distorting the child’s self-concept, world crucial implications. With regard to self-esteem, the Rosenberg
view, and affective capacities. Organizing the results from this Self-Esteem Scale (Rosenberg, 1965) is the most commonly
study under the framework of this model (Finkelhor and used measuring tool, and individuals with a history of offline
Browne, 1985) can thus aid in understanding the findings. The CSA report a higher frequency of low and medium self-esteem,
model proposes four different dynamics (traumatic sexualization, relative to individuals without such a history (e.g., García et al.,
betrayal, powerlessness, and stigmatization) that mediate the 2019). In line with this well-established finding, all participants
psychological outcome of CSA (Finkelhor and Browne, 1985). in this study reported low levels of self-esteem, expressed
Each dynamic can be expressed in many different ways, and both during the interviews and in the Rosenberg Self-Esteem
when analyzing the findings from this study through the lens Scale screening (Rosenberg, 1965) reported in Table 1. Again,
of this model, all four dynamics can be identified. Firstly, the processes between TA-CSA and subsequent psychological
traumatic sexualization refers to having one’s sexuality shaped suffering do not seem to be unique, but follow the same patterns
in developmentally inappropriate ways. This can, for instance, as those of offline CSA.
lead to sexual problems, sexual re-victimization, and negative
attitudes toward one’s sexuality and body (Finkelhor and Browne,
1985), all of which were expressed in the interviews. Secondly, Limitations
betrayal refers to discovering that one has been manipulated by First, this study’s results are based on a relatively small sample
a trusted person, and can result in distrusting people and lead (seven young female victims of TA-CSA). The small number of
to isolation (Finkelhor and Browne, 1985). This was captured participants is due to the sensitive nature of the topic and thus
in the theme Difficult to trust people. Thirdly, powerlessness the extreme difficulty in accessing this population. The sample
refers to the process in which one’s will, desire, and sense of size was, however, appropriate for the method chosen.
efficacy are continually contravened (Finkelhor and Browne, Second, all participants were self-selected and self-identified as
1985). Several participants expressed how they ignored their victims of TA-CSA. It is possible that this self-selection resulted
own will and blindly followed the demands of the offenders. in a biased sample, as victims who have experienced a more
According to the model, powerlessness can result in fear, anxiety, negative impact of abuse may have been more inclined to share
and re-victimization, which were also identified in the interviews. their experiences by taking part in the study. The majority of
Fourthly, stigmatization refers to the negative connotations the participants reported experiences of offline sexual abuse in
surrounding sexual victimization, such as shame, guilt, and addition to the TA-CSA (occurring either before or after the
badness, or that the activity was seen as taboo or deviant. TA-CSA). The focus of the interviews was, however, the TA-
This view can become incorporated into one’s self-image and, CSA, and any information that was brought up regarding other
for instance, lead to feelings of guilt and shame, as well as traumatizing experiences or the impact of these has been omitted
self-destructive behavior and suicide attempts (Finkelhor and from the analysis.
Browne, 1985), which were all described in the interviews. While Third, several years had passed between the (first) occasion of
this study in no way provided a full evaluation of the applicability TA-CSA and the interview, which means that the participants’
of the four traumagenic dynamics model on TA-CSA, it is evident stories are based on their retrospective recollections of
that the experiences and consequences of TA-CSA follow the their experiences. While this might have affected what
same pattern as those for offline CSA. they remembered and reported, it also provided a valuable

insight into how the perception and impact of the abuse had first-person perspective on the experiences of TA-CSA and
developed over time. the potential processes between victimization and subsequent
psychological suffering.
CONCLUSION
This study advanced our knowledge of how victims of TA-
CSA make sense of their experiences, and provided an in- No datasets can be shared due to the sensitive nature of the data.
depth understanding of the ways in which TA-CSA can lead
to psychological distress. The stories from seven victims of TA-
CSA illustrated the ways in which they experienced that their ETHICS STATEMENT
victimization had profoundly impacted their lives and health, in
both the short and long terms. Besides serious negative impacts This study involving human participants were reviewed and
on their mental health and relationships with other people, approved by the Regional Ethical Review Board in Linköping,
victims described how the victimization impacted their self- Sweden. Written informed consent from the participants’ legal
concept by altering the ways in which they viewed themselves in guardian/next of kin was not required to participate in this
relation to others. Adding this all together, it is clear that many of study in accordance with the national legislation and the
the consequences and the processes between victimization and institutional requirements.
psychological suffering are similar to those of offline CSA. In
addition, two factors appeared to add complexity to TA-CSA
victimization, that is increased levels of self-blame due to the AUTHOR CONTRIBUTIONS
victims’ own participation in the abuse and the fact that it was
conducted online, and the long-lasting fear of pictures or videos MJ, CL, SL, and LJ designed the study. MJ conducted the
of the abuse resurfacing. interviews. MJ and LJ performed the thematic analysis. MJ
In sum, viewing these results in light of other initial research created a draft of the manuscript, which was revised by CL, SL,
demonstrate the potential severity of TA-CSA, and indicates the and LJ. All authors contributed to the article and approved the
need to challenge the assumption that it is a less severe form submitted version.
of sexual abuse.
FUNDING
MANUSCRIPT CONTRIBUTIONS
This work was supported by a grant from the Swedish Crime
Given the recent increase in police reports regarding TA-CSA, Victim Compensation and Support Authority (08983/2016).
there is a significant desire to learn more about the phenomenon.
So far, research has mainly focused on offenders and offending
behavior, with attention more recently turning to its potential ACKNOWLEDGMENTS
consequences. While there seems to be a common assumption
that TA-CSA is a less severe form of sexual abuse, initial research The authors would like to express their sincere gratitude to the
indicate that the consequences can be as severe and harmful as participants who bravely and generously shared their stories.
for offline CSA. The aim of this study was to provide a deeper
understanding of how victims of TA-CSA make sense of their
experiences and its potential consequences. Interviews with seven SUPPLEMENTARY MATERIAL
victims of TA-CSA revealed the victimizations comprehensive
impact on the life and health of the participants, and the The Supplementary Material for this article can be found
sometimes complex process of understanding the severity of online at: https://www.frontiersin.org/articles/10.3389/fpsyg.
their experiences. In sum, the present findings provided a 2020.606218/full#supplementary-material
REFERENCES Berger, K. S. (2018). The Developing Person Through Childhood and Adoles-Cence.
New York, NY: Macmillan.
Anastassiou, A. (2017). Sexting and young people: a review of the qualitative Black, P. J., Wollis, M., Woodworth, M., and Hancock, J. T. (2015). A linguistic
literature. Qual. Rep. 22, 2231–2239. analysis of grooming strategies of online child sex offenders: implications for
Baumeister, R. F. (1999). “Self-concept, self-esteem, and identity,” in Personality: our understanding of predatory sexual behavior in an increasingly computer-
Contemporary Theory and Research, 2nd Edn, eds V. J. Derlega, B. A. Winstead, mediated world. Child Abuse Negl. 44, 140–149. doi: 10.1016/j.chiabu.2014.12.
and W. H. Jones (Chicago: Nelson-Hall), 339–375. 004
Bentley, H., Burrows, A., Hafizi, M., Kumari, P., Mussen, N., O’Hagan, O., Braun, V., and Clarke, V. (2006). Using thematic analysis in psychology. Qual. Res.
et al. (2019). How Safe are our Children? An Overview of Data on Child Psychol. 3, 77–101. doi: 10.1191/1478088706qp063oa
Abuse Online. Available online at: https://learning.nspcc.org.uk/media/1747/ Braun, V., and Clarke, V. (2013). Successful Qualitative Research: A Practical Guide
how-safe-are-our-children-2019.pdf for Beginners. Washington, DC: Sage.

Briggs, P., Simon, W. T., and Simonsen, S. (2011). An exploratory study of Internet- Jonsson, L., and Svedin, C. G. (2017). Barn Utsatta för Sexuella Övergrepp på Nätet
initiated sexual offenses and the chat room sex offender: has the Internet [Children Exposed to Sexual Abuse Online]. Linköping: Linköping University
enabled a new typology of sex offender? Sex. Abuse 23, 72–91. doi: 10.1177/ Electronic Press/Stiftelsen Allmänna Barnhuset.
1079063210384275 Jonsson, L. S., Fredlund, C., Priebe, G., Wadsby, M., and Svedin, C. G. (2019).
Brinkmann, S. (2014). “Unstructured and semi-structured interviewing,” in The Online sexual abuse of adolescents by a perpetrator met online: a cross-sectional
Oxford Handbook of Qualitative Research, ed. P. Leavy (Oxford: Oxford study. Child Adolesc. Psychiatry and Ment. Health 13:32.
University Press), 277–299. Kastbom, ÅA., Larsson, I., and Svedin, C. G. (2012). Parents’ observations and
Cantón-Cortés, D., Cortés, M. R., and Cantón, J. (2012). The role of traumagenic reports on the sexual behaviour of 7 to 13 years old children. Reprod. Syst. Sex.
dynamics on the psychological adjustment of survivors of child sexual abuse. Disord. 1, 2–7.
Eur. J. Dev. Psychol. 9, 665–680. doi: 10.1080/17405629.2012.660789 Kendall-Tackett, K. A., Williams, L. M., and Finkelhor, D. J. P. B. (1993). Impact
Coffey, P., Leitenberg, H., Henning, K., Turner, T., and Bennett, R. T. (1996). of sexual abuse on children: a review and synthesis of recent empirical studies.
Mediators of the long-term impact of child sexual abuse: perceived stigma, Psychol. Bull. 113:164. doi: 10.1037/0033-2909.113.1.164
betrayal, powerlessness, and self-blame. Child Abuse Negl. 20, 447–455. doi: Kloess, J. A., Beech, A. R., and Harkins, L. (2014). Online child sexual exploitation:
10.1016/0145-2134(96)00019-1 prevalence, process, and offender characteristics. Trauma Violence Abuse 15,
Council of Europe (2007). Council of Europe Convention on the Protection of 126–139. doi: 10.1177/1524838013511543
Children Against Sexual Exploitation and Sexual Abuse: Lanzarote, 25. X. 2007. Kloess, J. A., Hamilton-Giachritsis, C. E., and Beech, A. R. (2019). Offense processes
Lanzarote: Council of Europe. of online sexual grooming and abuse of children via Internet communication
Craven, S., Brown, S., and Gilchrist, E. (2006). Sexual grooming of children: review platforms. Sex. Abuse 31, 73–96. doi: 10.1177/1079063217720927
of literature and theoretical considerations. J. Sex. Aggress. 12, 287–299. doi: Lamoureux, B. E., Palmieri, P. A., Jackson, A. P., and Hobfoll, S. E. J. P. T. T.
10.1080/13552600601069414 (2012). Child sexual abuse and adulthood-interpersonal outcomes: examining
De Santisteban, P., Del Hoyo, J., Alcázar-Córcoles, M. Á, and Gámez-Guadix, pathways for intervention. Psychol. Trauma 4, 605–613. doi: 10.1037/a0026079
M. (2018). Progression, maintenance, and feedback of online child sexual Leonard, M. M. (2010). “I did what i was directed to do but he didn’t touch me”:
grooming: a qualitative analysis of online predators. Child Abuse Negl. 80, the impact of being a victim of internet offending. J. Sex. Aggress. 16, 249–256.
203–215. doi: 10.1016/j.chiabu.2018.03.026 doi: 10.1080/13552601003690526
Diamond, L. M., and Savin-Williams, R. C. (2009). “Adolescent sexuality,” in Levitt, H. M., Bamberg, M., Creswell, J. W., Frost, D. M., Josselson, R., and Suárez-
Handbook of Adolescent Psychology, Vol. 1, eds R. M. Lerner and L. Steinberg Orozco, C. (2018). Journal article reporting standards for qualitative primary,
(Hoboken, NJ: Wiley & Sons), 479–523. qualitative meta-analytic, and mixed methods research in psychology: the APA
Felitti, V. J., Anda, R. F., Nordenberg, D., Williamson, D. F., Spitz, A. M., Edwards, publications and communications board task force report. Am. Psychol. 73:26.
V., et al. (2019). Relationship of childhood abuse and household dysfunction to doi: 10.1037/amp0000151
many of the leading causes of death in adults: the adverse childhood experiences Madigan, S., Ly, A., Rash, C. L., Van Ouytsel, J., and Temple, J. R. (2018). Prevalence
(ACE) study. Am. J. Prev. Med. 56, 774–786. doi: 10.1016/j.amepre.2019.04.001 of multiple forms of sexting behavior among youth: a systematic review and
Finkelhor, D., and Browne, A. (1985). The traumatic impact of child sexual abuse: meta-analysis. JAMA Pediatr. 172, 327–335. doi: 10.1001/jamapediatrics.2017.
a conceptualization. Am. J. Orthopsychiatry 55, 530–541. doi: 10.1111/j.1939- 5314
0025.1985.tb02703.x Maniglio, R. (2009). The impact of child sexual abuse on health: a systematic review
García, J. A., Checay Olmos, F., Lucas Matheu, M., and Tesifón Carreño, P. (2019). of reviews. Clin. Psychol. Rev. 29, 647–657. doi: 10.1016/j.cpr.2009.08.003
Self esteem levels vs global scores on the rosenberg self-esteem scale. Heliyon Martin, J. (2015). Conceptualizing the harms done to children made the subjects
5:e01378. doi: 10.1016/j.heliyon.2019.e01378 of sexual abuse images online. Child Youth Serv. 36, 267–287. doi: 10.1080/
Halvorsen, J. E., Solberg, E. T., and Stige, S. H. (2020). “To say it out loud is to 0145935x.2015.1092832
kill your own childhood.”–An exploration of the first person perspective of Mitchell, K. J., Finkelhor, D., and Wolak, J. (2001). Risk factors for and impact
barriers to disclosing child sexual abuse. Child. Youth Serv. Rev. 113:104999. of online sexual solicitation of youth. Am. Med. Assoc. 285, 3011–3014. doi:
doi: 10.1016/j.childyouth.2020.104999 10.1001/jama.285.23.3011
Hamilton-Giachritsis, C., Hanson, E., Whittle, H., Alves-Costa, F., Pintos, A., and Muniz, C. N., Fox, B., Miley, L. N., Delisi, M., Cigarran, G. P., and Birnbaum,
Metcalf, T. (in press). Technology assisted child sexual abuse: professionals’ A. (2019). The effects of adverse childhood experiences on internalizing
perceptions of risk and impact on children and young people. Child Abuse Negl. versus externalizing outcomes. Crim. Justice Behav. 46, 568–589. doi: 10.1177/
104651. doi: 10.1016/j.chiabu.2020.104651 0093854819826213
Hamilton-Giachritsis, C., Hanson, E., Whittle, H., and Beech, A. (2017). Everyone Nilsson, D., Gustafsson, P. E., Larsson, J. L., and Svedin, C. G. (2010). Evaluation
Deserves to be Happy and Safe – A Mixed Methods Study Exploring How Online of the linköping youth life experience scale. J. Nerv. Ment. Dis. 198, 768–774.
and Offline Child Sexual Abuse Impact Young People and How Professionals doi: 10.1097/nmd.0b013e3181f4acb6
Respond to It. Available online at: https://bit.ly/2x5fcWY (accessed September Nilsson, D. K., Gustafsson, P. E., and Svedin, C. G. (2012). Lifetime
13, 2020). polytraumatization in adolescence and being a victim of bullying. J. Nerv.Ment.
Hjelen Stige, S., Halvorsen, J. E., and Solberg, E. T. (2020). Pathways to Dis. 200, 954–961. doi: 10.1097/nmd.0b013e3182718aa1
understanding–How adult survivors of child sexual abuse came to understand Paolucci, E. O., Genuis, M. L., and Violato, C. (2001). A meta-analysis of the
that they had been sexually abused. J. Child Sex. Abuse 29, 205–221. doi: published research on the effects of child sexual abuse. J. Psychol. 135, 17–36.
10.1080/10538712.2019.1643432 doi: 10.1080/00223980109603677
Hydén, M. (2014). The teller-focused interview: interviewing as a relational Rosenberg, M. (1965). Society and the Adolescent Child. Princeton, NJ: Princeton
practice. Qual. Soc. Work 13, 795–812. doi: 10.1177/1473325013506247 University Press.
International Centre for Missing and Exploited Children [ICMEC] (2017). Online Steinberg, L. (2011). Adolescence, 10th Edn. New York, NY: McGraw-Hill.
Grooming of Children for Sexual Purposes: Model Legislation & Global Review. Stern, A. E., Lynch, D. L., Oates, R. K., O’Toole, B. I., and Cooney, G. (1995).
Available online at: https://www.icmec.org/online-grooming-of-children-for- Self esteem, depression, behaviour and family functioning in sexually abused
sexual-purposes-model-legislation-global-review/ (accessed September 13, children. J. Child Psychol. Psychiatry 36, 1077–1089. doi: 10.1111/j.1469-7610.
2020). 1995.tb01352.x
Joleby, M., Landström, S., Lunde, C., and Jonsson, L. S. (2020a). Experiences and Swedish National Council for Crime Prevention (2019). Anmälda Brott: Tabeller –
psychological health among children exposed to online child sexual abuse – a Slutgiltig Statistik [Reported Crimes: Tables – Final Statistics]. Available online at:
mixed methods study of court verdicts. Psychol. Crime Law 1–23. doi: 10.1080/ http://www.bra.se/statistik/kriminalstatisik/anmälda-brott.html (accessed May
1068316X.2020.1781120 8, 2019).
Joleby, M., Lunde, C., Landström, S., and Jonsson, L. S. (2020b). ‘He Demanded Turner, H. A., Finkelhor, D., and Ormrod, R. (2010). The effects of adolescent
More Pictures’: Offender Strategies When Inciting Children to Engage in Online victimization on self-concept and depressive symptoms. Child Maltreat. 15,
Sexual Activity. Manuscript in preparation. 76–90. doi: 10.1177/1077559509349444

UNICEF (1989). Convention on the Rights of the Child. Available online at: http: Yardley, L. (2008). Demonstrating validity in qualitative psychology. Qual. Psychol.
//www.refworld.org/docid/3ae6b38f0.html (accessed September 13, 2020). 2, 235–251.
Valkenburg, P. M., and Peter, J. (2011). Online communication among adolescents:
an integrated model of its attraction, opportunities, and risks. J. Adolesc. Health Conflict of Interest: The authors declare that the research was conducted in the
48, 121–127. doi: 10.1016/j.jadohealth.2010.08.020 absence of any commercial or financial relationships that could be construed as a
Webster, S., Davidson, J., Bifulco, A., Caretti, V., Pham, T., Grove-Hills, J., et al. potential conflict of interest.
(2012). Final Report: European Online Grooming Project. Available online
at: https://www.researchgate.net/publication/257941820_European_Online_ The handling editor declared a past supervisory role with one of the authors SL.
Grooming_Project_-_Final_Report.
Whittle, H. C., Hamilton-Giachritsis, C., and Beech, A. R. (2013). Victims’ voices: Copyright © 2020 Joleby, Lunde, Landström and Jonsson. This is an open-access
the impact of online grooming and sexual abuse. Univers. J. Psychol. 1, 59–71. article distributed under the terms of the Creative Commons Attribution License
Wolak, J., and Finkelhor, D. (2013). Are crimes by online predators different from (CC BY). The use, distribution or reproduction in other forums is permitted, provided
crimes by sex offenders who know youth in-person? J. Adolesc. Health 53, the original author(s) and the copyright owner(s) are credited and that the original
736–741. doi: 10.1016/j.jadohealth.2013.06.010 publication in this journal is cited, in accordance with accepted academic practice.
Wolak, J., and Finkelhor, D. (2016). Sextortion: Findings From a Survey of 1,631 No use, distribution or reproduction is permitted which does not comply with
Victims. New Hampshire: University of New Hampshire. these terms.

ORIGINAL RESEARCH
published: 08 January 2021
doi: 10.3389/fpsyg.2020.612424
Transgender Identity Is Associated

With Bullying Involvement Among
Finnish Adolescents
Elias Heino 1* , Noora Ellonen 2 and Riittakerttu Kaltiala 1,3,4
1
Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland, 2 Faculty of Social Sciences, Tampere
University, Tampere, Finland, 3 Department of Adolescent Psychiatry, Tampere University Hospital, Tampere, Finland, 4 Vanha
Vaasa Hospital, Vaasa, Finland
Background: During adolescence, bullying often has a sexual content. Involvement in

bullying as a bully, victim or both has been associated with a range of negative health
outcomes. Transgender youth appear to face elevated rates of bullying in comparison to
their mainstream peers. However, the involvement of transgender youth as perpetrators
of bullying remains unclear in the recent literature.
Objective: The aim of this study was to compare involvement in bullying between
transgender and mainstream youth and among middle and late adolescents in a general
population sample.
Methods: Our study included 139,829 students in total, divided between a
Edited by: comprehensive school and an upper secondary education sample. Associations
Angelo Brandelli Costa, between gender identity and involvement in bullying were first studied using cross-
Pontifical Catholic University of Rio
Grande do Sul, Brazil tabulations with chi-square statistics. Logistic regression was used to study multivariate
Reviewed by: associations. Gender identity was used as the independent variable, with cisgender as
Guilherme Welter Wendt, the reference category. Subjection to and perpetration of bullying were entered each in
Universidade Estadual do Oeste do
Paraná, Brazil
turn as the dependent variable. Demographic factors, family characteristics, internalizing
Heng Choon (Oliver) Chan, symptoms, externalizing behaviors, and involvement in bullying in the other role were
City University of Hong Kong, added as confounding factors. Odds ratios (OR) with 95% confidence intervals (95% CI)
Hong Kong
are given. The limit for statistical significance was set at p < 0.001.
*Correspondence:
Elias Heino Results: Both experiences of being bullied and perpetrating bullying were more
elias.heino@tuni.fi
commonly reported by transgender youth than by cisgender youth. Among transgender
Specialty section: youth, all involvement in bullying was more commonly reported by non-binary youth
This article was submitted to than those identifying with the opposite sex. Logistic regression revealed that non-
Gender, Sex and Sexualities,
a section of the journal binary identity was most strongly associated with involvement in bullying, followed by
Frontiers in Psychology opposite sex identity and cisgender identity. Transgender identities were also more
Received: 01 October 2020 strongly associated with perpetration of bullying than subjection to bullying.
Accepted: 09 December 2020
Published: 08 January 2021 Conclusion: Transgender identity, especially non-binary identity, is associated with both
Citation: being bullied and perpetrating bullying even when a range of variables including internal
Heino E, Ellonen N and Kaltiala R
stress and involvement in bullying in the opposite role are taken into account. This
(2021) Transgender Identity Is
Associated With Bullying Involvement suggests that bullying during adolescence may serve as a mechanism of maintaining
Among Finnish Adolescents. heteronormativity.
Front. Psychol. 11:612424.
doi: 10.3389/fpsyg.2020.612424 Keywords: transgender, minority, non-conforming, victimization, bullying

1 January 2021 | Volume 11 | Article 612424
Heino et al. Association of Transgender Identity and Bullying
INTRODUCTION Both sexual minority and transgender youth may differ

from the mainstream by gender expression not conforming
Gender identity refers to an individual’s innate sense of being to traditional male and female roles (i.e., for males by being
male, female or an alternative gender (Bockting, 1999). Distinct feminine or for females by being masculine), which could render
from gender identity, gender expression refers to an individual’s them susceptible to being bullied, a behavior commonly directed
various characteristics which, during a given period, are generally at peers perceived as different (Jones et al., 2018; Price-Feeney
viewed as masculine or feminine (Coleman et al., 2012). et al., 2018). When comparing birth-assigned males and females
While various gender identities exist, the vast majority of in a school sample, Lowry et al. (2020) found that youth who
individuals present with cisgender identity, meaning that their described their appearance as gender non-conforming (i.e., males
gender identity is aligned with their birth-assigned sex. Gender believing they were perceived as feminine or females believing
minorities are individuals whose gender identity differs to various they were perceived as masculine) were violently victimized more
degrees from their birth-assigned sex. We refer to all gender often than those youth who described themselves as gender
minorities as transgender. This encompasses those who identify conforming and that the association was stronger among male
with the opposite sex and those whose gender identity aligns students. van Beusekom et al. (2020) likewise found that gender
with both or neither sex or varies (non-binary or gender non- non-conformity was associated with general victimization and
conforming gender identity). homophobic name calling and that the associations were stronger
Bullying is defined as aggressive behavior in which a pupil among males. Further, among transgender youth, those who
or group of pupils intentionally harm victims in various ways, perceived themselves as gender non-conforming were bullied
usually over a period of time, and is usually characterized by more frequently than those transgender youth who perceived
a power imbalance between the victim and the bully or bullies themselves as gender conforming, and also within a transgender
(Olweus, 1993; King et al., 1996). Bullying may assume various sample, the association between gender non-conformity and
forms, such as physical violence, verbal abuse, spreading rumors, experiences of being bullied was particularly strong among
exclusion from peer groups or then sexual gestures or remarks birth-assigned boys (Gower et al., 2018). In summary, it
(Olweus, 2013). Cyberbullying extends the scope of bullying seems that transgender youth as a whole are bullied more
to various information technologies, such as social media and often than their cisgender peers and that among transgender
mobile phones (Lindfors et al., 2012). A considerable share of populations bullying is more common among those who present
bullying among adolescents is of a sexual nature (Ashbaughm and as gender non-conforming.
Cornell, 2008) and often refers scornfully to homosexuality and The association between gender non-conformity and being
gender-non-conforming self-expression (Toomey et al., 2012). bullied may originate from heterosexism, a phenomenon
Adolescents’ involvement in bullying is common (Kaltiala- describing the effort to govern traditional masculine and
Heino and Fröjd, 2011; Lessne et al., 2016), and whether this feminine roles in society based on the assumption that
is as a victim or as a perpetrator, it has well documented heterosexuality is the superior sexual orientation and the norm
negative associations with health and educational trajectories (Chesir-Teran, 2003; Toomey et al., 2012). In the same vein,
(Chan and Wong, 2015a). Being bullied has been associated, for the stronger association of perceived gender non-conformity
example, with depression and suicidal ideation (Liang et al., 2007; and being bullied among natal males could be explained by
Kaltiala-Heino and Fröjd, 2011; Heikkilä et al., 2013) and school males’ stronger tendency to safeguard traditional masculine roles
truancy and impaired academic performance (Wormington et al., (van Beusekom et al., 2020). Behavior deviating from culturally
2016). Being a bully has likewise been associated with depression accepted masculine norms in boys is less readily tolerated than
(Klomek et al., 2008; Kaltiala-Heino et al., 2009; Kaltiala-Heino behavior deviating from the expected feminine behavior in girls
and Fröjd, 2011) and suicidal ideation (Kaltiala-Heino et al., 1999; (Ristori and Steensma, 2016). Even though the status of sexual
Heikkilä et al., 2013), but also with delinquency and substance and gender minorities has recently improved in many countries,
abuse (Liang et al., 2007). heterosexism is widespread (Dunn and Szymanski, 2017), thus
An abundance of research suggests that sexual minority adolescents not conforming to gender norms may be more
youth report being bullied 1.5–2 times more commonly than susceptible to bullying and harassment than their heterosexual
mainstream youth (Friedman et al., 2011; Abreu and Kenny, gender-conforming peers.
2018; Kurki-Kangas et al., 2019; McKay et al., 2019). Recent However, confounding by internal stress needs to be
research has also begun to unveil disparities in bullying considered when evaluating associations between transgender
involvement between gender minority and cisgender youth, identity and being bullied. Gender minority stress and resilience
particularly regarding disparities in being bullied. US-based (GMSR) theory (Hendricks and Testa, 2012; Testa et al.,
research indicates that transgender youth, in school samples, 2015) posits that gender minority people experience external
are bullied more often than their cisgender peers (Day et al., stress, such as discrimination and victimization (such as
2018; Eisenberg et al., 2019; Johns et al., 2019; Bishop et al., being bullied), but also internal stress related to internalized
2020). Transgender youth have been reported to more commonly transphobia and perceived stigma that predispose them to being
experience bullying related to gender or sexual orientation (Day constantly vigilant and anticipating discrimination. This may
et al., 2018) but also bullying related to weight and size (Bishop predispose to the development of depressive or hostile attribution
et al., 2020). In a clinical UK-based sample, almost 90% of bias (Morris, 2007; American Psychological Association, n.d.),
transgender youth reported being bullied (Witcomb et al., 2019). possibly leading to the perception of victimization by peers

when none was actually intended. Internal stressors may also (1) Is transgender identity associated with being bullied even
include concealment of one’s identity. Although hiding one’s when known correlates of involvement in bullying are
identity may reduce direct targeting by bullies, it may in controlled for?
turn create stress through identity non-affirmation and expose (2) Is transgender identity associated with perpetrating
to mental distress such as depression, known indeed to be bullying even when known correlates of involvement in
associated with transgender identity (Kaltiala-Heino et al., bullying are controlled for?
2018). Mental health problems may in turn further induce (3) Are the possible associations similar between opposite sex
negative attribution bias and experiences of being bullied identifying and non-binary youth?
and ostracized (Kaltiala-Heino and Fröjd, 2011). Therefore,
when studying associations between gender identity and being During adolescence, small differences in age may have a large
bullied, the role of mental distress needs to be accounted impact on development (Laursen and Hartl, 2013; Dahl et al.,
for in order to reveal possible independent associations 2018). Involvement in bullying decreases as adolescents grow
between transgender identity and being bullied. The role that older (Boulton and Underwood, 1992; Liang et al., 2007), and
transgender identity per se has in being bullied is important with maturation of sexuality (Cacciatore et al., 2019) and identity
for school policies to tackle bullying, and for health and development (Kroger et al., 2010), both older transgender youth
social policies. and their mainstream peers are likely more confident and more
Further, being bullied is commonly associated with able to handle diversity, which will likely also reduce involvement
perpetrating bullying (Cook et al., 2010; Shetgiri et al., 2012; in bullying among transgender youth. Thus, we finally ask:
Chan and Wong, 2015b). Those victimized themselves may (4) Are these associations similar among middle and late
reactively bully others or perpetrating bullying may be a way of adolescents?
defending oneself. On the other hand, aggressors often socialize
in antisocial groups where delinquency occurs, thus elevating We first expect to see that transgender adolescents report
the likelihood of being victimized themselves (Jennings et al., being bullied in excess in comparison to their cisgender peers,
2012). Perpetrating bullying may therefore arise from having but that the associations will grow weaker when confounding
been victimized or vice versa. by mental health correlates of bullying involvement and being
Thus, when studying the role of gender identity in being a bully perpetrator are controlled for. Second, we hypothesize
bullied among adolescents, perpetrating bullying needs to be that transgender youth will also report more perpetration
controlled for. Additionally, possible participation as a bully is of bullying than their cisgender peers. Third, in line with
an important problem in itself. To the best of our knowledge, heteronormative social control, we expect to see that the
the research so far has not explored bullying perpetrated among associations between gender identity and being bullied will be
gender minority youth (McKay et al., 2019). However, Dank et al. the strongest among non-binary/gender non-conforming youth.
(2014) found that transgender youth reported some of the highest And finally, we expect to find that the associations between
perpetration rates of sexual harassment perpetration. As bullying transgender identity and involvement in bullying will be weaker
among adolescents often has a sexual and heterosexist nature among older adolescents.
(Ashbaughm and Cornell, 2008; Toomey et al., 2012), similar
associations might be expected with bullying perpetration.
Elevated rates of bullying perpetration have also been found MATERIALS AND METHODS
among sexual minority populations in some studies (Berlan et al.,
2010; Eisenberg et al., 2015), who appear similar to transgender The School Health Promotion Study
youth when it comes to being bullied. The School Health Promotion Study (SHPS) of the National
To summarize, it appears that transgender youth are victims Institute for Health and Welfare is a school-based cross-sectional
of bullying more commonly than their cisgender peers, but anonymous survey designed to examine the health, health
research has not taken account of confounding by perpetrating behaviors, and school experiences of teenagers. The survey
bullying or mental health factors (Day et al., 2018; Gower et al., questionnaire is sent to every municipality in Finland. The
2018; Eisenberg et al., 2019; Johns et al., 2019; McKay et al., municipalities decide if the schools in their area will participate in
2019; Bishop et al., 2020; Lowry et al., 2020; van Beusekom the survey and the vast majority of schools do indeed participate.
et al., 2020). The possible associations between transgender The survey is run primarily for health policy and administrative
identity and perpetrating bullying are not known, leaving the purposes, and the data is available on request for purposes of
understanding of the associations between gender identity and scientific research. The main aim of the survey is to produce
this common problem incomplete. According to the research national adolescent health indicators that municipalities can
so far, it moreover remains unclear whether involvement in utilize in planning services and that can be used at national level
bullying is similar across various gender minority identities or to assess the effectiveness of health policies. The authors obtained
if opposite sex and non-binary identities differ in this respect. permission to use the data for scientific research but were not
Additionally, most of the literature on transgender youth and responsible for collecting it. The School Health Promotion Study
bullying originates from North America, a possibly culturally has received ethical approval from Tampere University Hospital
different setting from Northern Europe. In this context, we ask ethics committee and the ethics committee of the National
and aim to answer the following questions: Health Institute.

The survey is conducted among 8th and 9th graders Internalizing and Externalizing Symptoms
of comprehensive school and second-year students in upper Internalizing symptoms studied were depression and generalized
secondary education (upper secondary school and vocational anxiety. Depression was measured with two screening questions:
school) which follow completion of the 9 years of comprehensive “During the past month, have you often been bothered by
school. Survey participants in 2017 numbered 139,829. Of these, feeling down, depressed, or hopeless?” (yes/no) and “During the
48.9% (68,333) reported that they were male and 50.4% (70,539) past month, have you often been bothered by little interest or
that they were female. Of all respondents, 0.7% (957) did not pleasure in doing things?” (yes/no). These two questions have
report their sex, and these were excluded from further analyses. shown good psychometric properties in detecting depression in
Of the respondents, 52.7% were in comprehensive school grades adolescents (Richardson et al., 2010). In the analyses, a sum
8 or 9, 25.0% were attending upper secondary school, and 23.3% score of these items was used as continuous variable. Generalized
vocational school. The age of respondents in the comprehensive anxiety symptoms were elicited by the GAD-7, a self-report
school sample was [mean (SD)] 14.83 (0.82) years, those in
upper secondary school 16.84 (0.83) years and those in vocational
school 17.29 (2.43) years. Of the respondents, 3.5% (n = 4,940)
reported that they were 21 years old or older. These were excluded TABLE 1 | Descriptive statistics.
from further analyses. Descriptive information of the sample
Comprehensive education Upper secondary education
is given in Table 1. See section “Implausible, Likely Facetious
Responding” for final sample size. Demographic N (%) M (SD) N (%) M (SD)
variables
Measures Sex
Sex and Gender Identity Girl 36 123 (51.3%) 30 453 (50.8%)
The respondents were first asked “What is your sex?” with Boy 34 276 (48.7%) 29 520 (49.2%)
response alternatives “boy” and “girl” in the comprehensive Age 14.83 (0.82) 17.94 (2.17)
school survey, and “male”/“female” in the upper secondary Mother’s education
education response forms. This was intended to elicit the Only basic 3 815 (6.0%) 2 938 (5.2%)
respondent’s sex as noted in their identity documents and was Other 59 705 (94.0%) 53 580 (94. 8%)
the opening question of the whole survey. Later, in the section Father’s education
of the survey addressing health, respondents were asked about Only basic 5 520 (8.9%) 5 191 (9.4%)
their perceived gender as follows: “Do you perceive yourself Other 56 813 (91.1%) 50 315 (90.6%)
to be. . .,” with response options “a boy/a girl/both/none/my Family structure
perception varies.” According to sex and perceived gender, the Nuclear family 47 039 (69.5%) 38 699 (65.9%)
respondents were categorized into one of three gender identities: Other 20 682 (30.5%) 20 053 (34.1%)
cisgender identity (indicated male sex and perceives himself as At least one parent unemployed in past 12 months
a boy, or female sex and perceives herself as a girl), opposite Yes 20 736 (31.0%) 18 384 (31.5%)
sex identification (male sex, perceives herself as a girl; or female No 46 229 (69.0%) 39 972 (68.5%)
sex, perceives himself as a boy), and other/non-binary gender Difficulties to communicate with parents
identity (independent of sex: perceived to be both a boy and a Yes 4 902 (7.3%) 3 713 (6.4%)
girl, perceived to be neither a boy nor a girl, variable). No 61 946 (92.7%) 54 671 (93.6%)
Drinking alcohol weekly
Yes 2 790 (4.1%) 5 847 (9.9%)
Bullying
No 65 843 (95.9%) 53 349 (90.1%)
Bullying or being bullied was elicited using two questions derived Depression* 3.0 (1.5) 3.0 (1.5)
from a World Health Organization study on youth health (King GAD-7** 3.8 (4.7) 3.9 (4.6)
et al., 1996). The questions are based on Olweus’ definition of Gender identity
bullying (Olweus, 1993) that have been widely accepted as a basis Cisgender 66 687 (95.7%) 57 540 (96.5%)
for bullying research. Bullying was first defined as follows: “We Opposite sex 504 (0.7%) 313 (0.5%)
say a student is being bullied when another student (or group of Non-binary 2 483 (3.6%) 1 792 (3.0%)
students), say or do nasty things to him or her. It is also bullying gender
when a student is being teased repeatedly in a way she or he Bullied someone
does not like. But it is not bullying when two students of about Yes 1 717 (2.5%) 750 (1.3%)
the same strength quarrel or fight.” Respondents were then asked No 68 125 (97.5%) 58 884 (98.7%)
how frequently they had been bullied during the ongoing school Been bullied
term, and how frequently they had bullied others: many times Yes 3 438 (4.9%) 1 093 (1.8%)
a week, about once a week, less frequently, and not at all. In No 66 631 (95.1%) 58 789 (98.2%)
the analyses, responses to these questions were dichotomized to
*Range of depression was 2–8. **Range of GAD-7 was 0–21. The GAD-7 items
about once a week or many times a week (= frequently) vs. less describe the most prominent diagnostic features of the DSM IV generalized anxiety
frequently or not at all. disorder.

questionnaire designed to identify probable cases of generalized this, respondents reporting implausibly young age for being
anxiety disorder and to assess symptom severity. The GAD-7 enrolled in the grades studied (<13 years), implausible shortness
items describe the most prominent diagnostic features of the or height (extreme outliers) or who were calculated to have
DSM IV generalized anxiety disorder. The GAD-7 elicits how extreme BMI (< 10 or > 40) or reporting both extremely
often, during the last 2 weeks, the respondent has been bothered poor hearing, sight and mobility were classified as mischievous
by each of the seven core symptoms of generalized anxiety responders (for a detailed description, see Kaltiala-Heino and
disorder. Response options are “not at all,” “for several days,” Lindberg, 2019). Being classified as a mischievous respondent
“for more than half the days,” and “nearly every day,” scored, was strongly associated with reporting transgender identity in
respectively as 1, 2, 3, and 4. The GAD-7 has been shown to this data (Kaltiala-Heino and Lindberg, 2019). These respondents
be a reliable and valid measure for detecting generalized anxiety (2.7%) were excluded from further analyses. Thus, the data in
disorder in primary care and general population (Tiirikainen the analysis was from 130,372 respondents, of whom 96.1%
et al., 2019). In the analyses the sum score of these seven items were classified with cisgender identity, 0.6% with opposite sex
was used as continuous variable. identification, and 3.3% with other/non-binary gender identity.
Externalizing behaviors were represented, in addition to Descriptive statistics of the variables are presented in Table 1.
perpetrating bullying, by frequent consumption of alcohol.
Alcohol consumption was elicited as follows: “How often do Statistical Analyses
you use even small amounts of alcohol, for example half a Associations between gender identity and involvement in
can of beer or more?” with response options “once a week or bullying were first studied using cross-tabulations with
more often/once or twice a month/about once a month/less chi-square statistics. Logistic regression was used to study
frequently/not at all.” In the analyses the responses were multivariate associations. Gender identity was used as the
dichotomized to once a week or more often (= frequently) vs. all independent variable, with cisgender as the reference category.
other alternatives. (1) being bullied and (2) bullying others were entered each in turn
as the dependent variable. As covariates, in the first model age
Family Variables and sex were added, in the second model family characteristics
Family variables used were mother’s and father’s education (basic were added and finally, in the third model, internalizing
education, i.e., comprehensive school) only vs. at least upper symptoms, externalizing behaviors, and involvement in bullying
secondary education, family structure [living with both parents in the other role (as a bully when being bullied was studied, and
(= nuclear family) vs. any other family constellation], parental vice versa) were added. Odds ratios (OR) with 95% confidence
unemployment (none vs. one vs. both parents unemployed intervals (95% CI) are given. Due to the large size of the data we
or laid off during past 12 months) and difficulties in parent- set the limit for statistical significance at p < 0.001. The analyses
adolescent communication (never able to discuss important were run separately for the comprehensive school and upper
things with parents vs. can talk with parents at least sometimes). secondary education groups.
Family variables were controlled for because they have a strong
association with involvement in bullying (Knaappila et al., 2018).
RESULTS
Implausible, Likely Facetious Responding
It has been demonstrated that some adolescents deliberately Prevalence of Involvement in Bullying
mispresent themselves in survey studies, exaggerating their Overall, reported prevalence of experiences of being bullied was
belonging to minorities as well as their problem behaviors, higher in the comprehensive school sample than in the upper
symptoms, and psychosocial problems (Cornell et al., 2012; secondary education sample (4.9% vs. 1.8%). Similarly, reported
Robinson-Cimpian, 2014). Due to this, the proportion of those prevalence of bullying others was higher in the comprehensive
reporting belonging to minorities appears implausibly high, school sample (2.5% vs. 1.3%) (Table 1).
and associations between minority status and psychosocial Experiences of being bullied were most commonly reported
problems are overestimated. In relation to gender identity, by non-binary students, followed by opposite sex identifying and
such overestimation may risk a perception in society that cisgender students in both samples (Table 2).
gender variant youth are victims rather than active subjects
participating in building the contemporary adolescent
community. Particularly in light of the excessive media TABLE 2 | Experiences of bullying and bullying others according to gender
coverage of gender identity issues (Marchiano, 2017), gender identity,% (n).
identity is likely to be a topic which tempts adolescents to give
Cisgender Opposite sex Non-binary gender p
facetious responses.
Excluding respondents reporting unlikely combinations of Comprehensive education
extreme responses outside the focus of present interest on topics Been bullied 4.3 (2 886) 12.8 (64) 16.5 (404) <0.001
theoretically not related to the variables of interest for the Bullied others 2.0 (1 356) 8.9 (44) 10.5 (256) <0.001
actual study questions has been shown to be an appropriate Upper secondary education
method for controlling for such facetious responding (Robinson- Been bullied 1.6 (899) 5.8 (18) 8.5 (152) <0.001
Cimpian, 2014; Kaltiala-Heino and Lindberg, 2019). In line with Bullied others 1.0 (582) 4.9 (15) 7.9 (140) <0.001

Prevalence of perpetrating bullying followed a similar pattern. Notably, the final models also revealed a strong correlation
Bullying others was most commonly reported by non-binary between being bullied and bullying others, particularly in the
students, followed in both samples by opposite sex identifying upper secondary education sample (comprehensive education
and cisgender students (Table 2). sample, [OR (95% CI) = 9.20 (7.96–10.64), p = 0.000]; upper
secondary education sample, [OR (95% CI) = 46.97 (37.84-58.31),
p = 0.000] (Table 3; Models 3a and 3b ).
Relationship Between Gender Identity
and Being Bullied
Table 3 presents the associations between gender identity and Other Correlates of Being Subjected to
being bullied in the comprehensive school and upper secondary Bullying
education samples before and after controlling for relevant In addition to perpetrating bullying, certain family variables,
confounding. Among the comprehensive school sample, opposite depression, and anxiety were positively associated with being
sex identification first yielded over twofold odds while non- bullied among both samples. In both samples, a negative
binary identity yielded over fourfold odds for being bullied association emerged between natal female sex and being bullied
(Table 3; Model 1a∗ ). In the upper secondary education sample, (Table 3; Models 3a and 3b ).
a similar pattern emerged but with even stronger associations
(Table 3; Model 1b∗∗ ).
In each model presented in Table 3, more covariates The Relationship Between Gender
were added and controlled for. The associations between Identity and Perpetrating Bullying
opposite sex identification and non-binary identity with being Table 4 presents the association between gender identity and
bullied grew stronger when age and sex were added into the perpetrating bullying among the comprehensive school and
model but diminished when confounding by family variables, upper secondary education samples before and after controlling
alcohol consumption, and finally mental health variables and for confounding. Throughout our models, the associations
perpetrating bullying were controlled for. Nevertheless, in the between gender minority identities and perpetrating bullying
final model (Table 3; Model 3a ), after controlling for the were stronger than the associations between gender minority
aforementioned covariates, a statistically significant association identities and being bullied.
between non-binary identity and being bullied persisted among Initially in the comprehensive school sample, all transgender
the comprehensive school sample [OR (95% CI) = 1.98 (1.69– youth had over fourfold odds for perpetrating bullying (Table 4;
2.32), p = 0.000]. Among the upper secondary education Model 1a∗ ). In comparison to the comprehensive school sample,
sample, the association likewise only persisted among non-binary the odds for perpetrating bullying were lower for the opposite sex
youth [OR (95% CI) = 1.99 (1.50–2.62), p = 0.000] (Table 3; identifying and higher for the non-binary identifying youth in the
Model 3b ). upper secondary education sample (Table 4; Model 1b∗ ).
TABLE 3 | Regression analysis of being bullied.
Model 1a * Model 2a Model 3a Model 1b ** Model 2b Model 3b
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Gender ID (ref. cisgender)

Opposite sex 2.86 (2.02–4.04) 2.67 (1.89–3.82) 1.66 (1.13–2.49) 3.77 (2.10–6.78) 3.29 (1.82–5.94) 2.13 (1.01–4.17)
Non-binary 4.22 (3.67–4.86) 3.52 (3.04–4.06) 1.98 (1.69–2.32) 5.21 (4.14–6.55) 4.35 (3.44–5.49) 1.99 (1.50–2.62)
Official gender female (ref. male) 0.76 (0.72–0.85) 0.73 (0.68–0.80) 0.54 (0.50–0.60) 0.64 (0.55–0.74) 0.60 (0.52–0.69) 0.64 (0.54–0.75)
Age 0.96 (0.90–1–03) 0.94 (0.88–1.01) 0.83 (0.78–0.91) 1.04 (0.95–1.13) 0.99 (0.90–1.08) 0.94 (0.85–1.03)
Mother only basic education (ref. other) 1.22 (1.03.1.45) 1.17 (0.97–1.40) 1.82 (1.42–2.36) 1.58 (1.18–2.10)
Father only basic education (ref. other) 1.10 (0.95-1.28) 1.01 (0.87-1.19) 1.26 (1.00-1.58) 1.18 (0.92-1.50)
Nuclear family (ref. no) 1.09 (0.99–1.19) 0.97 (0.88–1.06) 1.19 (1.03–1.39) 1.11 (0.95–1.30)
Parental unemployment (ref. no) 1.36 (1.25–1.49) 1.22 (1.13–1.33) 1.35 (1.17–1.57) 1.25 (1.07–1.47)
Communication difficulties with parents (ref. no) 2.51 (2.24–2.82) 1.31 (1.15–1.49) 2.45 (2.01–3.02) 1.37 (1.09–1.74)
Alcohol weekly (ref. no) 1.30 (1.10–1.55) 1.30 (1.05–1.60)
Depression (continuous) 1.18 (1.14–1.23) 1.16 (1.08–1.24)
GAD-7 (continuous)*** 1.08 (1.06–1.09) 1.07 (1.04–1.09)
Bullied others at least once a week (ref. no) 9.20 (7.96–10.64) 46.97 (37.84–58.31)
Statistically significant values (p < 0.001) presented in bold face. *Estimates in unadjusted model: Opposite sex OR (95% CI) = 2.73 (CI 1.93–3.86), non-binary OR (95%
CI) = 4.09 (3.55–4.70). **Estimates in unadjusted model: Opposite sex OR (95% CI) = 3.51 (1.96–6.30), non-binary OR (95% CI) = 5.01 (4.02–6.34). ***Range of GAD-7
was 0–21. The GAD-7 items describe the most prominent diagnostic features of the DSM IV generalized anxiety disorder.

TABLE 4 | Regression analysis for perpetration of bullying.
Model 1a * Model 2a Model 3a Model 1b ** Model 2b Model 3b
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Gender ID (ref. cisgender)

Opposite sex 6.13 (4.09–9.19) 5.56 (3.69–8.39) 3.91 (2.47–6.19) 5.62 (2.85–11.12) 5.03 (2.54–9.99) 3.15 (1.38–7.19)
Non-binary 6.00 (4.98–7.24) 4.83 (3.98–5.86) 2.58 (2.07–3.21) 8.61 (6.71–11.05) 7.24 (5.60–9.35) 4.01 (2.91–5.52)
Official gender female (ref. male) 0.28 (0.25–0.32) 0.26 (0.23–0.30) 0.25 (0.21–0.29) 0.18 (0.14–0.23) 0.17 (0.13–0.21) 0.19 (0.15–0.25)
Age 1.13 (1.03–1.24) 1.09 (1.00–1.20) 1.04 (0.94–1.15) 1.07 (0.96–1.20) 1.02 (0.90–1.14) 0.96 (0.84–1.08)
Mother only basic education (ref. other) 1.36 (1.07–1.72) 1.24 (0.96–1.59) 2.01 (1.45–2.78) 1.51 (1.01–2.21)
Father only basic education (ref. other) 1.20 (1.06–1.36) 1.29 (1.03–1.60) 1.12 (0.84–1.50) 0.96 (0.69–1.34)
Nuclear family (ref. no) 1.21 (1.06–1.38) 1.12 (0.98–1.28) 1.13 (0.94–1.37 1.05 (0.85–1.29)
Parental unemployment (ref. no) 1.21 (1.06–1.37) 1.08 (0.94–1.23) 1.17 (0.98–1.41) 1.03 (0.84–1.27)
Communication difficulties with parents (ref. no) 2.71 (2.28–3.21) 1.59 (1.31–1.92) 2.62 (2.03–3.38) 1.83 (1.35–2.49)
Alcohol weekly (ref. no) 3.99 (3.34–4.77) 2.85 (2.28–3.56)
Depression (continuous) 1.02 (0.95–1.08) 0.96 (0.87–1.06)
GAD-7 (continuous)*** 1.04 (1.02–1.06) 1.03 (0.99–1.06)
Been bullied at least once a week (ref. no) 8.90 (7.68–10.30) 45.68 (36.73–56.82)
Bold face indicates statistically significant values (p < 0.001). *Estimates in unadjusted model: Opposite sex OR (95% CI) = 4.77 (3.21–7.10), non-binary OR (95%
CI) = 5.10 (4.2–6.13). **Estimates in unadjusted model: Opposite sex OR (95% CI) = 4.31 (2.20–8.44), non-binary OR (95% CI) = 7.57 (5.93–9.66). ***Range of GAD-7
was 0–21. The GAD-7 items describe the most prominent diagnostic features of the DSM IV generalized anxiety disorder.
The associations between opposite sex identification and DISCUSSION

non-binary identity with perpetrating bullying grew stronger
when age and sex were added into the model but diminished In this study we analyzed the association of gender minority
when confounding by family variables, alcohol consumption, identity with involvement in bullying among a large population-
and finally mental health variables and perpetrating bullying based sample of adolescents. We analyzed whether the
were controlled for. association of gender identity and involvement in bullying
In the comprehensive school sample, the association between differed among opposite sex and non-binary identifying youth or
gender identity and perpetrating bullying nevertheless persisted among middle and late adolescents.
as statistically significant among both opposite sex identifying We firstly found that in our large, nationally representative
[OR (95% CI) = 3.91 (2.47–6.19), p = 0.000] and non-binary sample, being bullied was generally associated with transgender
youth [OR (95% CI) = 2.58 (2.07–3.21), p = 0.000] although the identity, and with non-binary identity in particular. This finding
association was stronger among opposite sex identifying youth is in line with the existing literature, which indicates that
(Table 4; Model 3a ). In the upper secondary education sample, experiences of being bullied are more common among gender
the association persisted statistically significant only among nonminority than mainstream youth (Day et al., 2018; Eisenberg
binary youth [OR (95% CI) = 4.01 (2.91–5.52), p = 0.000] et al., 2019; Johns et al., 2019; Bishop et al., 2020). Various
(Table 4; Model 3b ). factors could explain this disparity. Transgender youth may
Notably, the final models also revealed a strong association differ from their peers in that their behavior or appearance
in both samples between being bullied and bullying others deviates from traditional feminine and masculine roles. This
[comprehensive education sample, OR (95% CI) = 8.90 (7.68– could partly explain elevated rates of being bullied as bullying is
10.30), p = 0.000; upper secondary education sample, OR (95% often targeted at those perceived to deviate from the mainstream
CI) = 45.68 (36.73–56.82), p = 0.000]. The association was (Jones et al., 2018; Price-Feeney et al., 2018). More specifically,
stronger among the upper secondary education sample (Table 4; relating to sexual orientation and gender identity, bullying
Models 3a and 3b ). sexual and gender minorities could also stem from heterosexism,
which refers to efforts to maintain traditional masculine and
Other Correlates of Perpetrating Bullying feminine roles in society (Chesir-Teran, 2003; Toomey et al.,
Positive associations between perpetrating bullying and 2012). On the other hand, internal stress, as described in gender
difficulties communicating with parents and weekly alcohol minority stress and resilience theory (Hendricks and Testa,
consumption were found in both samples. In the comprehensive 2012; Testa et al., 2015), could result in constant vigilance
school sample, anxiety (but not depression) was also positively and anticipation of being victimized through the development
associated with perpetrating bullying. In both samples, a negative of hostile or depressive attribution bias thus predisposing
association between natal female sex and perpetrating bullying transgender youth to detect victimization by their peers where
was found (Table 4; Models 3a and 3b ). none was actually intended.

Secondly, we found that transgender identity was generally involvement in bullying decreases (Boulton and Underwood,
associated with perpetrating bullying and that the association 1992; Liang et al., 2007; Coulter et al., 2018). In line with this,
was stronger than that of transgender identity and being bullied. involvement in bullying in our data was reported less commonly
To the best of our knowledge, past research has not examined by the older adolescents in the upper secondary education
perpetration of bullying among gender minority youth, thus sample across all gender identities. The association between
rendering comparisons to prior research impossible. In a study opposite sex identification and being bullied also leveled out
by Dank et al. (2014), however, it was reported that the few when confounding was controlled for in both samples. However,
transgender young people in their study were the ones most likely regardless of lower reported prevalence, the association between
to perpetrate dating violence among their sample. non-binary identity and perpetration of bullying was stronger
Such aggressive behavior could arise from being victimized among the older than among the younger adolescents in our
or having witnessed victimization of other gender or sexual study. It might be that those adolescents who still remain involved
minorities (Eisenberg et al., 2016), as a coping mechanism or in bullying at an older age represent adolescents with the most
avenue through which one could release negative feelings. developmental challenges. This finding could be seen to lend
On the other hand, adolescence in general is a mentally support to the notion that among transgender youth the possibly
challenging time (Paus et al., 2008) during which adolescents more complex nature of non-binary identity (in comparison to
struggle with a series of developmental tasks such as forming peer opposite sex identifying or cisgender youth) is indeed related to
relations and coming to grips with their sexuality (Havighurst, additional developmental challenges.
1948; Seiffge-Krenke and Gelhaar, 2008). The added complexity Additionally, while involvement in bullying was less prevalent
due to the emergence and further development of transgender among the older students of our study, the correlation between
identity could cause extra stress for adolescents. In this being bullied and being a bully grew stronger. This is likewise
context, perpetrating bullying could be seen as sign of acting in agreement with the assumption that when involvement in
out, perhaps due to transgender adolescents’ own unresolved bullying becomes less common as age increases, those who
developmental issues. remain involved likely represent adolescents with the most
Thirdly, non-binary identity was more strongly associated developmental challenges. Being both a bully and a victim
with involvement in bullying than opposite sex identity. Past (bully-victim) is known to correlate with greatest amount
research has found elevated rates of being subjected to bullying of mental health problems and developmental difficulties
among youth (Lowry et al., 2020; van Beusekom et al., 2020) and (Forero et al., 1999).
transgender youth (Gower et al., 2018) who perceive themselves
as more gender non-conforming (i.e., masculine females or
feminine males) than youth with no such perception. Non-binary Strengths and Weaknesses of the
identifying youth particularly may display gender expression that Present Study
does not conform to either masculine or feminine roles, and this Our study has several strengths. Our large sample was
may make them vulnerable to being bullied either due to simply an unselected, population-based sample representative of
being different from the mainstream, or as a result of heterosexist Finnish middle and late adolescents. This enhances the
control. We found, however, that not only being bullied but also generalizability of our results.
engaging in bullying was even more common among non-binary There are indications even between European countries
(perception of gender conforms to both or neither sex or it varies) of variation in transgender youth’s peer relationships and
than among opposite sex identifying youth. psychological functioning (de Graaf et al., 2018; van der Star
It may be that the process of gender identity formation is et al., 2018). One could speculate that such differences are even
a more complex process among non-binary youth than those greater between European and North American adolescents. As
young people identifying with the opposite sex. Such differences most research on gender identity and involvement in bullying
could stem from the nature of non-binary identity itself, as originates in the United States, we feel our study in a Northern
perceived gender may fluctuate, or align with both or neither European setting is a useful addition to the existing literature
traditional gender roles. This could delay the achievement of on the important subject of involvement in bullying and
so-called transgender identity milestones, or factors associated transgender identity.
with the formation of transgender identity, such as first living We controlled in our analyses for a wide range of confounding
in the gender role felt within (Wilkinson et al., 2018) as factors closely related to involvement in bullying and gender
young people struggle with their still unresolved gender identity. minority identity. This allowed us to examine more closely the
This internal turmoil due to uncertainty about one’s own relationship between transgender identity and involvement in
identity, could, for example, impede the formation of peer bullying. This is a strength of our study.
relationships, a key part of adolescent development (Laursen As has been recommended (Reisner et al., 2014; Eisenberg
and Hartl, 2013). This could exacerbate internal stress and et al., 2017), we identified transgender youth with two separate
predispose non-binary youth to mental health symptoms such as questions located far apart from each other in the study
depression, which are known to relate to involvement in bullying questionnaire (“two-step method”). Due to the large sample
(Kaltiala-Heino and Fröjd, 2011). size, we were additionally able to separate opposite sex
Finally, regarding age differences, the existing literature shows identifying youth from non-binary youth, rather than grouping
that as adolescents mature and progress toward adulthood, all transgender youth as one in our analyses.

Involvement in bullying was elicited using questions derived suggest that the development of transgender identity (and non-
from WHO’s Youth Study (King et al., 1996). The WHO binary identity in particular) is an additional stress for youth as
questions have since then been used in numerous studies across they navigate the already developmentally challenging years of
countries (for review see Kaltiala-Heino and Fröjd, 2011) which adolescence toward adulthood.
makes data elicited with them comparable with earlier research. Future studies should focus on including gender minority
This is a strength of our study. specific measures in study questionnaires. Such measures could
Our study also has several weaknesses. In spite of our large include various gender identities and for example gender
sample, the number of transgender youth reporting perpetrating minority specific stressors named in the GMSR theory, such as
bullying was on the smaller side, although we still feel we reached living in the desired role. Such measures could help uncover in
adequate cell sizes for statistical validity. more detail the association between bullying involvement and
In the present study, a secondary data was used. The data was various gender identities per se.
not planned nor collected by us, and we were therefore unable Programs that promote gender diversity should be
to influence the way certain topics of interest were elicited. As implemented in schools and in larger context in the society
a result, the way experiences of bullying were elicited in the with the aim of reducing heteronormativity and promoting the
study questionnaire made it impossible to distinguish between acceptance of gender diversity.
different types of bullying behavior in which adolescents had been Teachers, parents and health care workers must consider that
involved, such as traditional school bullying or cyberbullying, or gender minority youth are not necessarily only victims but also
physical and verbal bullying and exclusion. perpetrators of bullying.
Additionally, whether respondents were living in their
desired gender roles was not elicited in the questionnaire.
This inhibited additional comparisons regarding involvement DATA AVAILABILITY STATEMENT
in bullying among those who conceal their gender identity vs.
those who do not. The GMSR theory suggests concealment The data analyzed in this study is subject to the following
of one’s experienced gender identity (for example not living licenses/restrictions: The data belongs to the Finnish Institute
in the desired gender role) is a stressor that could possibly for Health and Welfare and is available for researchers by
negatively affect mental health of gender minority people. application. Requests to access these datasets should be directed
One could thus speculate that living in the desired gender to pauliina.luopa@thl.fi.
role could in fact reduce mental health symptoms such as
depression, thus decreasing bullying involvement, a behavior
associated with mental health issues. On the other hand, living ETHICS STATEMENT
in the desired gender role could manifest as behavior or
appearance deviating from traditional masculine and feminine The studies involving human participants were reviewed and
roles (such as natal girls using boys’ restrooms or natal approved by the Tampere University Hospital ethics committee
boys having a more feminine appearance) thus predisposing (Tampere University Hospital, Tampere, Finland) and National
youths to bullying, a behavior commonly directed to those Institute of Health and Welfare ethics committee (National
who deviate from the mainstream. Lastly, as the study Institute of Health and Welfare, Helsinki, Finland). Written
was a cross-sectional one, caution must be exercised when informed consent from the participants’ legal guardian/next of
interpreting the results as causality cannot be determined kin was not required to participate in this study in accordance
from such data. with the national legislation and the institutional requirements.
AUTHOR CONTRIBUTIONS
CONCLUSION
RK conceived the idea of the study and supervised the project.
Transgender identity and non-binary identity in particular, is NE and RK designed the statistical analyses while NE carried
associated with both being bullied and bullying others even when them out. EH wrote the manuscript with help from RK and
a range of variables, including internal stress and involvement in NE. All authors contributed to the final manuscript, each with
bullying in the opposite role, are taken into account. This could a specific focus.
REFERENCES Ashbaughm, L. P., and Cornell, D. G. (2008). Sexual harassment and bullying beha-
viors in sixth-graders. J. Sch. Violence 7, 21–38. doi: 10.1300/J202v07n02_03
Abreu, R., and Kenny, M. (2018). Cyberbullying and LGBTQ youth: a systematic Berlan, E. D., Corliss, H. L., Field, A. E., Goodman, E., and Austin, S. B. (2010).
literature review and recommendations for prevention and intervention. Sexual orientation and bullying among adolescents in the growing up today
J. Child Adolesc. Trauma 11, 81–97. doi: 10.1007/s40653-017-0175-7 study. J. Adolesc. Health 46, 366–371. doi: 10.1016/j.jadohealth.2009.10.015
American Psychological Association (n.d.). Hostile Attribution Bias. Available Bishop, A., Overcash, F., McGuire, J., and Reicks, M. (2020). Diet and physical
online at: https://dictionary.apa.org/hostile-attribution-bias (accessed activity behaviors among adolescent transgender students: school survey
September 15, 2020). results. J. Adolesc. Health 66, 484–490. doi: 10.1016/j.jadohealth.2019.10.026

Bockting, W. (1999). From construction to context: gender through the eyes of the Forero, R., McLellan, L., Rissel, C., and Bauman, A. (1999). Bullying behaviour and
transgendered. SIECUS Rep. 28, 3–7. psychosocial health among school students in New South Wales. Australia: cross
Boulton, M. J., and Underwood, K. (1992). Bully/victim problems among middle sectional survey. BMJ 319, 344–348. doi: 10.1136/bmj.319.7206.344
school children. Br. J. Educ. Psychol. 62, 73–87. doi: 10.1111/j.2044-8279.1992. Friedman, M. S., Marshal, M. P., Guadamuz, T. E., Wei, C., Wong, C. F., Saewyc,
tb01000.x E., et al. (2011). A meta-analysis of disparities in childhood sexual abuse,
Cacciatore, R., Korteniemi-Poikela, E., and Kaltiala, R. (2019). The steps of parental physical abuse, and peer victimization among sexual minority and
sexuality—a developmental, emotion-focused, child-centered model of sexual sexual nonminority individuals. Am. J. Public Health 101, 1481–1494. doi:
development and sexuality education from birth to adulthood. Int. J. Sex. Health 10.2105/AJPH.2009.190009
31, 319–338. doi: 10.1080/19317611.2019.1645783 Gower, A. L., Rider, G. N., Coleman, E., Brown, C., McMorris, B. J., and
Chan, H. C. O., and Wong, D. S. W. (2015a). Traditional school bullying and Eisenberg, M. E. (2018). Perceived gender presentation among transgender and
cyberbullying in Chinese societies: prevalence and a review of the whole-school gender diverse youth: approaches to analysis and associations with bullying
intervention approach. Aggress. Violent Behav. 23, 98–108. doi: 10.1016/j.avb. victimization and emotional distress. LGBT Health 5, 312–319. doi: 10.1089/
2015.05.010 lgbt.2017.0176
Chan, H. C. O., and Wong, D. S. W. (2015b). The overlap between school Havighurst, R. J. (1948). Developmental Tasks and Education. Chicago, IL:
bullying perpetration and victimization: assessing the psychological, familial, University of Chicago Press.
and school factors of Chinese Adolescents in Hong Kong. J. Child Family Stud. Heikkilä, H. K., Väänänen, J., Helminen, M., Fröjd, S., Marttunen, M., and Kaltiala-
24, 3224–3234. doi: 10.1007/s10826-015-0125-7 Heino, R. (2013). Involvement in bullying and suicidal ideation in middle
Chesir-Teran, D. (2003). Conceptualizing and assessing heterosexism in high adolescence: a 2-year follow-up study. Eur. Child Adolesc. Psychiatry 22, 95–102.
schools: a setting-level approach. Am. J. Commun. Psychol. 31, 267–279. doi: doi: 10.1007/s00787-012-0327-0
10.1023/A:1023910820994 Hendricks, M., and Testa, R. (2012). A conceptual framework for clinical work with
Coleman, E., Bockting, W., Botzer, M., Cohen-Kettenis, P., DeCuypere, G., transgender and gender nonconforming clients: an adaptation of the minority
Feldman, J., et al. (2012). Standards of care for the health of transsexual, stress model. Prof. Psychol. Res. Pr. 43, 460–467. doi: 10.1037/a0029597
transgender, and gender-nonconforming people, Version 7. Int. J. Transgend. Jennings, W., Piquero, A., and Reingle, J. (2012). On the overlap between
13, 165–232. doi: 10.1080/15532739.2011.700873 victimization and offending: a review of the literature. Aggress. Violent Behav.
Cook, C. R., Williams, K. R., Guerra, N. G., Kim, T. E., and Sadek, S. (2010). 17, 16–26. doi: 10.1016/j.avb.2011.09.003
Predictors of bullying and victimization in childhood and adolescence: a meta- Johns, M. M., Lowry, R., Andrzejewski, J., Barrios, L. C., Demissie, Z., McManus,
analytic investigation. Sch. Psychol. Q. 25, 65–83. doi: 10.1037/a0020149 T., et al. (2019). Transgender identity and experiences of violence victimization,
Cornell, D., Klein, J., Konold, T., and Huang, F. (2012). Effects of validity screening substance use, suicide risk, and sexual risk behaviors among high school
items on adolescent survey data. Psychol. Assess. 24, 21–35. doi: 10.1037/ students - 19 states and large Urban School Districts, 2017. Morb. Mortal. Wkly.
a0024824 Rep. (MMWR) 68, 67–71. doi: 10.15585/mmwr.mm6803a3
Coulter, R. W. S., Bersamin, M., Russell, S. T., and Mair, C. (2018). The effects Jones, L. M., Mitchell, K. J., Turner, H. A., and Ybarra, M. L. (2018). Characteristics
of gender- and sexuality-based harassment on lesbian, gay, bisexual, and of bias-based harassment incidents reported by a national sample of U.S.
transgender substance use disparities. J. Adolesc. Health 62, 688–700. doi: 10. adolescents. J. Adolesc. 65, 50–60. doi: 10.1016/j.adolescence.2018.02.013
1016/j.jadohealth.2017.10.004 Kaltiala-Heino, R., Bergman, H., Työläjärvi, M., and Frisén, L. (2018). Gender
Dahl, R. E., Allen, N. B., Wilbrecht, L., and Suleiman, A. B. (2018). Importance of dysphoria in adolescence: current perspectives. Adolesc. Health Med. Therap.
investing in adolescence from a developmental science perspective. Nature 554, 9, 31–41. doi: 10.2147/AHMT.S135432
441–450. doi: 10.1038/nature25770 Kaltiala-Heino, R., and Fröjd, S. (2011). Correlation between bullying and clinical
Dank, M., Lachman, P., Zweig, J. M., and Yahner, J. (2014). Dating violence depression in adolescent patients. Adolesc. Health Med. Therap. 2, 37–44. doi:
experiences of lesbian, gay, bisexual, and transgender youth. J. Youth Adolesc. 10.2147/AHMT.S11554
43, 846–857. doi: 10.1007/s10964-013-9975-8 Kaltiala-Heino, R., Fröjd, S., and Marttunen, M. (2009). Involvement in bullying
Day, J. K., Perez-Brumer, A., and Russell, S. T. (2018). Safe schools? transgender and depression in a 2-year follow-up in middle adolescence. Eur. Child Adolesc.
youth’s school experiences and perceptions of school climate. J. Youth Adolesc. Psychiatry 19:45. doi: 10.1007/s00787-009-0039-2
47, 1731–1742. doi: 10.1007/s10964-018-0866-x Kaltiala-Heino, R., and Lindberg, N. (2019). Gender identities in adolescent
de Graaf, N. M., Cohen-Kettenis, P. T., Carmichael, P., de Vries, A. L. C., Dhondt, population: methodological issues and prevalence across age groups. Eur.
K., Laridaen, J., et al. (2018). Psychological functioning in adolescents referred Psychiatry 55, 61–66. doi: 10.1016/j.eurpsy.2018.09.003
to specialist gender identity clinics across Europe: a clinical comparison study Kaltiala-Heino, R., Rimpelä, M., Marttunen, M., Rimpelä, A., and Rantanen,
between four clinics. Eur. Child Adolesc. Psychiatry 27, 909–919. doi: 10.1007/ P. (1999). Bullying, depression, and suicidal ideation in Finnish adolescents:
s00787-017-1098-4 school survey. BMJ 319, 348–351. doi: 10.1136/bmj.319.7206.348
Dunn, T., and Szymanski, D. (2017). Heterosexist Discrimination and LGBQ King, A., Wold, B., Tudor-Smith, C., and Harel, Y. (1996). The health of youth.
activism: examining a moderated mediation model. Psychol. Sex. Orientat. A cross-national survey. WHO Regional Publications. Eur. Series 69:222.
Gend. Divers. 5, 13–24. doi: 10.1037/sgd0000250 Klomek, A. B., Sourander, A., Kumpulainen, K., Piha, J., Tamminen, T., Moilanen,
Eisenberg, M. E., Gower, A. L., and McMorris, B. J. (2016). Emotional health of I., et al. (2008). Childhood bullying as a risk for later depression and suicidal
lesbian, gay, bisexual and questioning bullies: does it differ from straight bullies? ideation among Finnish males. J. Affect. Disord. 109, 47–55. doi: 10.1016/j.jad.
J. Youth Adolesc. 45, 105–116. doi: 10.1007/s10964-015-0316-y 2007.12.226
Eisenberg, M. E., Gower, A. L., McMorris, B. J., and Bucchianeri, M. M. (2015). Knaappila, N., Marttunen, M., Fröjd, S., Lindberg, N., and Kaltiala-Heino, R.
Vulnerable bullies: perpetration of peer harassment among youths across sexual (2018). Socioeconomic trends in school bullying among Finnish adolescents
orientation, weight, and disability status. Am. J. Public Health 105, 1784–1791. from 2000 to 2015. Child Abuse Neglect 86, 100–108. doi: 10.1016/j.chiabu.2018.
doi: 10.2105/AJPH.2015.302704 09.011
Eisenberg, M. E., Gower, A. L., McMorris, B. J., Rider, G. N., Shea, G., and Coleman, Kroger, J., Martinussen, M., and Marcia, J. E. (2010). Identity status change during
E. (2017). Risk and protective factors in the lives of transgender/gender adolescence and young adulthood: a meta-analysis. J. Adolesc. 33, 683–698.
nonconforming adolescents. J. Adolesc. Health 61, 521–526. doi: 10.1016/j. doi: 10.1016/j.adolescence.2009.11.002
jadohealth.2017.04.014 Kurki-Kangas, L., Marttunen, M., Fröjd, S., and Kaltiala-Heino, R. (2019). Sexual
Eisenberg, M. E., Gower, A. L., Nic Rider, G., McMorris, B. J., and Coleman, E. orientation and bullying involvement in adolescence: the role of gender, age,
(2019). At the intersection of sexual orientation and gender identity: variations and mental health. J. Sch. Violence 18, 319–332. doi: 10.1080/15388220.2018.
in emotional distress and bullying experience in a large population-based 1488136
sample of U.S. Adolesc. J. LGBT Youth 16, 235–254. doi: 10.1080/19361653. Laursen, B., and Hartl, A. C. (2013). Understanding loneliness during
2019.1567435 adolescence: developmental changes that increase the risk of perceived

social isolation. J. Adolesc. 36, 1261–1268. doi: 10.1016/j.adolescence.2013. Seiffge-Krenke, I., and Gelhaar, T. (2008). Does successful attainment of
06.003 developmental tasks lead to happiness and success in later developmental
Lessne, D., Cidade, M., Gerke, A., Roland, K., and Sinclair, M. (2016). Student tasks? A test of Havighurst’s (1948) theses. J. Adolesc. 31, 33–52. doi: 10.1016/j.
Victimization in U.S. Schools: Results From the 2013 School Crime Supplement adolescence.2007.04.002
to the National Crime Victimization Survey. Available online at: https://nces.ed. Shetgiri, R., Lin, H., and Flores, G. (2012). Identifying children at risk for being
gov/pubsearch/pubsinfo.asp?pubid=2016145 (accessed November 22, 2016). bullies in the United States. Acad. Pediatr. 12, 509–522. doi: 10.1016/j.acap.
Liang, H., Flisher, A. J., and Lombard, C. J. (2007). Bullying, violence, and risk 2012.06.013
behavior in South African school students. Child Abuse Neglect 31, 161–171. Testa, R. J., Habarth, J., Peta, J., Balsam, K., and Bockting, W. (2015). Development
doi: 10.1016/j.chiabu.2006.08.007 of the gender minority stress and resilience measure. Psychol. Sex. Orientat.
Lindfors, P., Kaltiala, R., and Rimpela, A. (2012). Cyberbullying among Finnish Gend. Divers. 2, 65–77. doi: 10.1037/sgd0000081
adolescents—A population based study. BMC Public Health 12:1027. doi: 10. Tiirikainen, K., Haravuori, H., Ranta, K., Kaltiala-Heino, R., and Marttunen,
1186/1471-2458-12-1027 M. (2019). Psychometric properties of the 7-item Generalized Anxiety
Lowry, R., Johns, M. M., and Robin, L. E. (2020). Violence victimization, substance Disorder Scale (GAD-7) in a large representative sample of Finnish
use disparities, and gender-nonconforming youth. Am. J. Prev. Med. 58, e159– adolescents. Psychiatry Res. 272, 30–35. doi: 10.1016/j.psychres.2018.
e169. doi: 10.1016/j.amepre.2019.12.021 12.004
Marchiano, L. (2017). Outbreak: on transgender teens and psychic epidemics. Toomey, R. B., McGuire, J. K., and Russell, S. T. (2012). Heteronormativity, school
Psychol. Perspect. 60, 345–366. doi: 10.1080/00332925.2017.1350804 climates, and perceived safety for gender nonconforming peers. J. Adolesc. 35,
McKay, T., Lindquist, C. H., and Misra, S. (2019). Understanding (and Acting On) 187–196. doi: 10.1016/j.adolescence.2011.03.001
20 years of research on violence and LGBTQ + Communities. Trauma Violence van Beusekom, G., Collier, K. L., Bos, H. M. W., Sandfort, T. G. M., and Overbeek,
Abuse 20, 665–678. doi: 10.1177/1524838017728708 G. (2020). Gender nonconformity and peer victimization: sex and sexual
Morris, S. (2007). Attributional biases in subclinical depression: a schema−based attraction differences by age. J. Sex Res. 57, 234–246. doi: 10.1080/00224499.
account. Clin. Psychol. Psychother. 14, 32–47. doi: 10.1002/cpp.512 2019.1591334
Olweus, D. (1993). Bullying at School: What We Know And What We Can Do. van der Star, A., Pachankis, J., and Bränström, B. (2018). LGBT bullying at school
Hoboken, NJ: Blackwell Publishing. across 28 European countries: the impact of bullying and structural stigma
Olweus, D. (2013). School bullying: development and some important challenges. on later life satisfaction. Eur. J. Public Health 28(Suppl._4):cky213.054. doi:
Annu. Rev. Clin. Psychol. 9, 751–780. doi: 10.1146/annurev-clinpsy-050212- 10.1093/eurpub/cky213.054
185516 Wilkinson, L., Pearson, J., and Liu, H. (2018). Educational attainment of
Paus, T., Keshavan, M., and Giedd, J. N. (2008). Why do many psychiatric disorders transgender adults: does the timing of transgender identity milestones matter?
emerge during adolescence? Nat. Rev. Neurosci. 9, 947–957. doi: 10.1038/ Soc. Sci. Res. 74, 146–160. doi: 10.1016/j.ssresearch.2018.04.006
nrn2513 Witcomb, G. L., Claes, L., Bouman, W. P., Nixon, E., Motmans, J., and Arcelus,
Price-Feeney, M., Jones, L. M., Ybarra, M. L., and Mitchell, K. J. (2018). J. (2019). Experiences and psychological wellbeing outcomes associated with
The relationship between bias-based peer victimization and depressive bullying in treatment-seeking transgender and gender-diverse youth. LGBT
symptomatology across sexual and gender identity. Psychol. Violence 8, 680– Health 6, 216–226. doi: 10.1089/lgbt.2018.0179
691. doi: 10.1037/vio0000219 Wormington, S. V., Anderson, K. G., Schneider, A., Tomlinson, K. L., and
Reisner, S. L., Conron, K. J., Tardiff, L. A., Jarvi, S., Gordon, A. R., and Austin, Brown, S. A. (2016). Peer victimization and adolescent adjustment: does
S. B. (2014). Monitoring the health of transgender and other gender minority school belonging matter? J. Sch. Violence 15, 1–21. doi: 10.1080/15388220.2014.
populations: validity of natal sex and gender identity survey items in a U.S. 922472
national cohort of young adults. BMC Public Health 14:1224. doi: 10.1186/1471-
2458-14-1224 Conflict of Interest: The authors declare that the research was conducted in the
Richardson, L. P., Rockhill, C., Russo, J. E., Grossman, D. C., Richards, J., McCarty, absence of any commercial or financial relationships that could be construed as a
C., et al. (2010). Evaluation of the PHQ-2 as a brief screen for detecting major potential conflict of interest.
depression among adolescents. Pediatrics 125:1097. doi: 10.1542/peds.2009-
2712 Copyright © 2021 Heino, Ellonen and Kaltiala. This is an open-access article
Ristori, J., and Steensma, T. D. (2016). Gender dysphoria in childhood. Int. Rev. distributed under the terms of the Creative Commons Attribution License (CC BY).
Psychiatry 28, 13–20. doi: 10.3109/09540261.2015.1115754 The use, distribution or reproduction in other forums is permitted, provided the
Robinson-Cimpian, J. (2014). Inaccurate estimation of disparities due to original author(s) and the copyright owner(s) are credited and that the original
mischievous responders: several suggestions to assess conclusions. Educ. Res. publication in this journal is cited, in accordance with accepted academic practice. No
43, 171–185. doi: 10.3102/0013189X14534297 use, distribution or reproduction is permitted which does not comply with these terms.

ORIGINAL RESEARCH
published: 12 January 2021
doi: 10.3389/fpsyg.2020.609958
Young People’s Experiences of

Attending a Theater-in-Education
Program on Child Sexual Exploitation
Hannah May 1 , Juliane A. Kloess 2* , Kari Davies 2 and Catherine E. Hamilton-Giachritsis 1
1
Department of Psychology, University of Bath, Bath, United Kingdom, 2 Centre for Applied Psychology, School
of Psychology, University of Birmingham, Birmingham, United Kingdom
Child sexual exploitation and abuse (CSEA) has grave implications for the mental health
and wellbeing of children and young people. It has been linked to a wide range of
difficulties which may extend into adulthood. School-based prevention programs that
aim to raise awareness (and thereby have the potential to prevent CSEA) are popular,
however, have historically lacked robust and consistent evaluation. The purpose of
the present study was therefore to explore young people’s experiences of attending a
school-based theater-in-education program, and the impact this had on their awareness
Edited by: and understanding of CSEA. Four focus groups of between four to six participants
Nadia Marie Wager,
each were conducted with young people from two co-educational State schools in the
University of Huddersfield,
United Kingdom United Kingdom. The approach of Template Analysis was used to analyze the data,
Reviewed by: and revealed a number of themes related to the superordinate themes of “Information
Martha Frías Armenta, and Detail Delivered” and “Format and Timing.” The results suggest that participants
University of Sonora, Mexico
Peter James Hills,
gained new awareness and understanding of aspects related to CSEA, including other
Bournemouth University, forms of (criminal) exploitation, as well as how to avoid harm and what to do “if bad
United Kingdom
things happen.” Participants further reported that the theater performance/live element
*Correspondence:
of the program was particularly impactful, feeling that this was delivered to them at the
Juliane A. Kloess
J.A.Kloess@bham.ac.uk right time, but suggesting that younger people would also benefit from the important
messages. In addition, areas for improvement were identified in terms of the delivery
Specialty section:
of the program, and the issue of victim blaming. Findings are discussed with a view to
Forensic and Legal Psychology, practical implications and directions for future research.
Frontiers in Psychology Keywords: child sexual exploitation, child sexual abuse, internet safety education, theater in education,
awareness raising, school-based prevention, relationship and sex education
Received: 24 September 2020
Accepted: 14 December 2020
Published: 12 January 2021
INTRODUCTION
Citation:
May H, Kloess JA, Davies K and Child sexual exploitation and abuse (CSEA) involves an individual or group “taking advantage
Hamilton-Giachritsis CE (2021) Young
of an imbalance of power to coerce, manipulate or deceive a young person (i.e., a person below
People’s Experiences of Attending
a Theater-in-Education Program on
the age of 18 years) into sexual activity (a) in exchange for something the victim needs or wants,
Child Sexual Exploitation. and/or (b) for the financial advantage or increased status of the perpetrator” (Department for
Front. Psychol. 11:609958. Education, 2017, p. 5). CSEA takes place in the physical world and via Internet technologies, with
doi: 10.3389/fpsyg.2020.609958 over 9,000 incidents of sexual offenses against children between October 2017 and September 2018

May et al. Experiences of a CSE Education Program
involving an online element, including rape, sexual assault, and For the purpose of the present article, preventative measures
grooming (National Society for the Prevention of Cruelty to and strategies aimed at “educating” children and young people
Children [NSPCC], 2019). about CSEA will be termed school-based sexual abuse prevention
The accurate detection and recording of CSEA remains a programs (SSAPPs). They may be delivered in a range of
huge challenge, given that it is usually shrouded in secrecy, formats, including videos, role plays, structured exercises, and
resulting in low rates of disclosure. Official, up-to-date prevalence group discussions. While Brown and Saied-Tessier (2015)
rates of CSEA are therefore hard to establish, and retrospective suggest that SSAPPs are the most common form of CSEA
data from various samples are relied upon in order to provide prevention, the great majority of research which has explored
estimates (Office for National Statistics, 2020). According to data their effectiveness was conducted outside the United Kingdom.
collected as part of the United Kingdom Adverse Childhood In order for these programs to be effective, the authors argue
Experiences (ACEs) studies, the prevalence rate for England is that they should comprise at least four sessions or more,
6.3% (Bellis et al., 2014), and the prevalence rate for Wales covering a range of topics, including healthy relationships,
is 10% (Public Health Wales NHS Trust, 2015). Similarly, a consent, online safety, and where to go for help. SSAPPs
retrospective survey of 24,899 adults in the general population Topping and Barron (2009) cautioned against taking apparently
(aged between 18 and 74 years) in England and Wales for the year positive SSAPP outcomes at face value, and highlighted that
ending March 2019 revealed that 7.5% of the sample had reported effect sizes often paint a different picture, with actual outcomes
experiences of sexual abuse which had been committed against varying considerably. They expressed concern over the 22
them before the age of 16 years (Office for National Statistics, studies included in their review generally lacking valid and
2020). Internationally, in their meta-analysis of 217 studies, reliable outcome measures, having minimal replicability, and
Stoltenborgh et al. (2011) reported a global prevalence rate of reporting no measures of fidelity. It was therefore merely
12%, based on the analysis of hundreds of samples consisting of a possible to calculate effect sizes for 11 of the 22 studies by
total of approximately 10 million individuals. focusing on knowledge and skills around safety as outcomes
Evidence suggests that CSEA has profound consequences which showed large variation (d = 0.14–1.40). The authors also
for children’s and young people’s physical and psychological noted that there was evidence of negative outcomes for some
wellbeing which often continues into adulthood. In a systematic participants, such as a fear of strangers, and embarrassment
literature review of 14 reviews (comprising 587 studies), Maniglio and wariness around touch. However, these were predominantly
(2009) concluded that there was evidence to suggest that those reported by adults (and not young people themselves), and were
who experience CSEA are at significant risk of developing short in duration.
a variety of mental health difficulties, including psychosis, A Cochrane review of SSAPPs found that when compared to
personality disorder, posttraumatic stress, and substance abuse. a control group, the programs increased children’s and young
In an umbrella review of negative outcomes linked to experiences people’s protective behavioral skills (measured in a pass/fail
of CSEA, Hailes et al. (2019) estimated that up to 10% of common simulated grooming scenario) immediately post-intervention
mental health problems in the general population, including (Walsh et al., 2015). It was also found that SSAPPs produced
depression, anxiety, eating disorders and post-traumatic stress increases in knowledge of CSEA prevention concepts (i.e.,
disorder, could be prevented if CSEA was eliminated. Finally, body ownership, private parts, distinguishing appropriate and
it also has to be acknowledged that these figures demonstrate inappropriate touch, and types of secrets, as well as whom to
that we are dealing with a public health issue, with substantial tell), as assessed by means of both vignettes and questionnaires.
implications not only for associated support services, but also for These effects were sustained at 6 month follow up. Young people
society as a whole. who had taken part in SSAPPs were more likely to disclose
Recognition of the long-term impact of CSEA, and the experiences of CSEA, while at the same time acknowledging
difficulty of detecting it, has led to growing emphasis on that this may be impacted by the clustering of participants in
preventative measures in order to reduce young people’s schools/classes. However, the review concluded that there was
vulnerability to being exposed to and experiencing CSEA (Beckett insufficient evidence to support the long-term effect of SSAPPs
et al., 2017). Research with survivors of CSEA suggests that in terms of reducing the incidence of CSEA in participants.
failing to report such experiences may be related to the victim Based on their meta-analysis, Bovarnick and Scott (2016)
believing that they are in a real and loving relationship with took a more critical position by suggesting that even the
the perpetrator (Quayle et al., 2012; Whittle et al., 2014; Beckett most effective programs are unlikely to change how children
et al., 2017). The subject most often talked about by young and young people actually behave, especially if they are “one-
people interviewed in Quayle et al.’s(2012) study was how “what off ” measures. They did acknowledge, however, that SSAPPs
seemed in some ways so normal or desirable turned into the may increase children’s and young people’s knowledge of and
opposite” (p. 50). Educational preventative measures that enable confidence around aspects related to CSEA (such as power
children and young people to distinguish legitimate relationships imbalance), as well as having the potential to challenge attitudes
from inappropriate and abusive ones may therefore help tackle around gender and relationships that contribute to and underpin
CSEA more widely. In a recent NSPCC study, young people harmful sexual behavior (e.g., consent). The authors concluded
highlighted that “online safety” needed to be part of a broader by recommending more intensive programs of longer duration,
education about healthy relationships and consent (rather than and advised that these should be tailored to meet the specific
being delivered on its own) (Hamilton-Giachritsis et al., 2017). needs of the relevant school in order to be most effective.

In 2019, the NSPCC published a report detailing an SD = 5.67), while this remained unchanged in the control group
evaluation of their Protect and Respect Child Sexual Exploitation (M = 42.42, SD = 8.67).
Programme (Williams, 2019). This program includes group- In the Cochrane review by Walsh et al. (2015), three out of
based education work that is delivered in schools. Children the 24 SSAPPs involved some element of theater in education.
were referred into these groups following concerns raised by While the findings of the review were generally positive, neither
staff about their potential risks of experiencing CSEA. The the studies included in the review nor the review itself isolated
evaluation employed a predominantly qualitative methodology, the impact of theater in education from other interventions.
and data derived from interviews with NSPCC practitioners In an integrative review of SSAPPs specifically, Fryda and
suggest that young people engaged best with these groups Hulme (2014) found that six out of the 23 identified programs
when they were based around their own life experiences, and used theater, with the most common mode of delivery being
by having the opportunity to share these and ask questions. film. Group discussion and role play, both elements of theater
Overall, practitioners reported that they had observed a positive in education, were part of 10 and 12 SSAPPs, respectively.
impact for young people in terms of their awareness and Overall, the studies included in the review highlighted positive
understanding of CSEA-related risks (although no additional outcomes for children in terms of disclosures of abuse, perception
outcome measure of this appears to have been collected), of risk, and self-protection skills, with the most frequently
however, they were in disagreement as to whether this awareness measured outcome being knowledge gain. However, in light
and understanding would translate into a real-life reduction of the limitations across the studies, these findings are to be
of risk in terms of experiencing CSEA (Williams, 2019). It interpreted with caution. The authors also suggest that the
is important to note that our knowledge of SSAPS and their variation across studies in terms of measures that were used to
effectiveness is not yet well understood, and therefore care must capture changes in children represents conflicting views of which
be taken to ensure that children and young people are protected aspects/factors may significantly reduce children’s and young
from experiencing unintended adverse outcomes as a result of people’s risk of experiencing CSEA.
taking part in them. In recent years, there has been growing recognition that
One particular approach to SSAPP is theater in education, statutory education for children and young people on healthy
which is a process of using performance, workshops and role play relationships, sex and consent is lacking in the United Kingdom.
to encourage young people to explore topics that they may feel More specifically, an inquiry by Barnardo’s (2014) found that
reluctant or ill-equipped to discuss (Sawney et al., 2003; Wooster, the young people who responded had received inadequate
2016). It aims to prompt safe communication around these teaching around healthy relationships and sex, concluding
topics, and develop young people’s capacity to make informed that the provision of high-quality education on relationships
decisions (Sawney et al., 2003). Sawney et al. (2003) suggest and sex in schools was vital. In response to these identified
that it can be difficult to measure the effectiveness of theater in deficits, from September 2020, it became a statutory requirement
education due to its dynamic and co-constructed nature, as well for all primary school children to receive Relationships
as existing disparity in terms of what may constitute meaningful Education, and for all secondary school children to receive
change and impact. Nevertheless, there is evidence to suggest Relationships and Sex Education. This is enshrined in The
that theater in education is effective in: (i) increasing awareness Relationships Education, Relationships and Sex Education and
of sexually transmitted diseases in 14–15 years old graduate Health Education (England) Regulations 2019 under Sections
students (Lightfoot et al., 2015); (ii) further developing young 34 and 35 of the Children and Social Work Act 20171 . Schools
people’s understanding of healthy vs. abusive relationships in a are permitted flexibility in determining how they choose to
group of 12–13 years old pupils (Bell and Stanley, 2006); and deliver this education (Department for Education, 2019), with
(iii) enhancing the impact of sexual health education, based on one option being theater-in-education programs.
reflections of facilitators about their experience of delivering the It therefore seems timely to present our findings from a
program (Gordon and Gere, 2016). study that sought to explore young people’s experiences of
The use of this approach to delivering SSAPPs therefore seems attending a school-based theater-in-education program, as part
to be appropriate. However, in terms of long-term change, few of which aspects of unhealthy relationships and sex were
studies have directly examined the role of theater in education covered. The company whose program was evaluated as part
in terms of reducing the risk of children and young people of the present study uses a theater-in-education approach to
experiencing CSEA. In a group randomized control trial by Krahé delivering their SSAPP. It is based in the United Kingdom, and
and Knappert (2009), one group of German school children provides a range of theater-in-education programs to educate
watched a play about how to manage abusive interactions children and young people about various important aspects,
involving adults (n = 44), with another group watching a including relationships, consent and exploitation. The primary
recording of the performance (n = 55), and a third group acting aim of the study presented here was to explore young people’s
as a control group (n = 49). Both intervention groups showed experiences of attending a school-based theater-in-education
significant increases in skills for dealing with abusive interactions program, and the impact this had on their awareness and
(i.e., distinguishing good/bad touch and secrets, getting help, understanding of CSEA by conducting focus groups with pupils
and rejecting unwanted touch), which was measured at 2 and who had attended the relevant performance and participated in
30 weeks post-intervention. Those who saw the play showed a
significant increase in skills (M = 48.73, SD = 6.44 and M = 48.55, 1
https://www.legislation.gov.uk/ukpga/2017/16/part/1/chapter/4/enacted

subsequent workshops2 . The study therefore aimed to answer the actors informed the pupils that, if interested, they were able
following questions: to take part in a research study that involved providing some
feedback about their experience of attending the program. At
1. How did young people experience attending the program?
the end of the workshop part of the program3 , the researchers
2. What did young people find most helpful/useful?
told the pupils about the possibility of contributing to a set of
3. What do young people know now that they did not know
focus groups, and handed out information sheets to anyone who
before?
expressed an interest.
4. What were young people’s views of the characters?
5. What was the impact of attending the program on young
people?
Consent and Assent
Consent forms were distributed to interested young people as
6. What additional elements did young people think would
consent by parents/guardians was required in order for young
have been beneficial to include?
people to be able to take part in the study. Young people were
7. Did young people think the program was delivered in the
asked to take the consent forms home to get them signed by their
right format and at the right time?
parents/guardians, and hand them back to a member of teaching
staff who then liaised with the researchers over organizing a
MATERIALS AND METHODS date and time for the focus groups to take place. Prior to the
commencement of the focus group discussions, young people
Ethics were reminded of the purpose of the study, invited to ask any
Full ethical approval for the study was granted by the Science, questions, and asked to sign an assent form.
Technology, Engineering, and Mathematics Ethical Review
Committee at the University of Birmingham, and the Psychology Data Collection
Research Ethics Committee at the University of Bath. The The focus group discussions followed a semi-structured interview
researchers adhered to the British Psychological Society’s (2018) schedule, asking participants about their experience of attending
throughout the study. the theater-in-education program. The researchers allowed
participants’ contributions and discussion in response to
Sample/Participants questions to reach a natural conclusion before continuing
A total of four focus groups were conducted at two co- with the interview schedule. Clarification was sought from
educational State schools in the United Kingdom. At the first participants, where necessary. The discussions were audio-
school, the program was delivered to Year 10 pupils (aged 14– recorded using a Dictaphone and transcribed verbatim by a
15 years) in February 2019, with two focus groups (FG1 and FG2) professional transcription service.
taking place 4 days later. At the second school, the program was To the researchers’ knowledge, no disclosures of sexual
delivered to Year 9 pupils (aged 13–14 years) in June 2019, with exploitation and abuse were made by participants. One
one focus group (FG3) taking place 9 days later, and the second participant in FG3 made statements which alluded to the
focus group (FG4) taking place 15 days later. In both schools, the possibility of a friend having had experiences similar to those
program formed part of a dedicated day for pupils to learn about depicted in the program. In line with the company’s safeguarding
healthy relationships and sex. More specifically, Focus Group 1 policy, the researcher informed the designated member of
(n = 5) consisted of three females and two males; Focus Group teaching staff upon completion of the focus group to ensure that
2 (n = 4) consisted of three females and one male; Focus group this could be followed up in accordance with the school’s policies
3 (n = 4) consisted of four females; and Focus Group 4 (n = 6) and procedures around safeguarding.
consisted of four females and two males.
Data Analysis
Procedure The transcribed focus group discussions were analyzed using
When schools contacted the theater-in-education company, they Template Analysis. Template Analysis is a qualitative data
were asked whether they would be interested in taking part in analysis approach for thematically grouping and analyzing text. It
a research study. If they agreed, they were put in touch with results in a list of codes (i.e., a “template”), with hierarchical codes
the PI (second author) to receive further information about the signifying themes that were identified in the data. It differs from
study, and to discuss relevant organizational aspects for the day other forms of thematic analysis in that some of these codes are
of the program, and the running of the focus groups, respectively, defined by the researcher in the form of a preliminary template
including the distribution and collection of parent/guardian prior to analyzing the text. These are subsequently expanded
consent forms and young people’s assent forms. and modified throughout the process of data analysis (King,
On the day of the program, the researchers visited the schools 3
The program consists of two elements—one is in the form of a performance,
and attended the program for its duration. At the start of the one is in the form of a workshop. For the purpose of the present study, analysis
program, they were introduced to the pupils by the actors. The was conducted on any data relevant to the performance element. This element
focused on child sexual exploitation and abuse (CSEA). Any data in focus group
2
It should be noted that the young people who attended the program, and took discussions that were concerned with the workshop element (covering county lines
part in our study, were teenagers aged between 13 and 15 years, and consequently and other forms of criminal exploitation) were not analyzed as part of the study
received the program quite late in their relationship and sex education. presented here.

2004). Transcripts were imported into NVivo12, a qualitative data and more. . .you need to be more mature in the way you do things,
analysis software, with the purpose of facilitating the process of because you might make one small mistake or just reply to someone,
analysis (see Figure 1 for an overview of the process). and that could be the worst mistake you ever made (FG2).
Across the focus group discussions, there was a sense that this
Epistemological Position new awareness of potential risk was being translated into wary
Template Analysis is compatible with a range of epistemological and less trusting behavior by participants:
stances that may be adopted as one undertakes qualitative
research (Brooks et al., 2015). For the purpose of the present Just be careful, even if it’s not online and it’s meeting somebody in
study, the researchers adopted a realist position with a view real life. . .it’s just being really careful who you trust and how quickly
to discover factual information about the experiences of you trust somebody before you know more about them (FG3).
young people who attended a school-based theater-in-education Much of the behavioral change participants mentioned
program, and the impact this had on their awareness and was focused around the use of social media or internet
understanding of CSEA. communication platforms, which was identified as a common
pattern across all focus group discussions:
RESULTS Be suspicious, like if a boy. . .if a man started texting you, just block
them or remove them or report them, say if it was Instagram or
The original research questions were borne in mind during something, just try and block them, report them, so you don’t have
the process of data analysis, as well as the interpretation to get yourself into a situation like Cath’s5 situation (FG4).
and organization of themes. In order to help make sense Like when it comes to having private accounts on social media, only
of the data derived from the focus group discussions, the accept the people you know because then you know what they’re
themes were amalgamated into two overarching categories, like and you can trust them more because you already know them
namely “Information and Detail Delivered” and “Format and in person (FG3).
Timing.” Themes and subthemes relevant to these categories
are reported below. Theme 2: What to Do if Bad Things Happen
This theme summarizes young people’s understanding of what
Information and Detail Delivered they think they should do if they spot any signs of CSEA
This category presents young people’s views of the content and (involving either themselves or friends/peers), as well as what
level of detail provided as part of the program. Participants’ may happen if they chose to seek help. It includes the subthemes
responses indicated that they learnt about the following aspects “How to seek help and what to expect” and “Helping each other.”
from the program: (i) avoiding harm; (ii) what to do if bad things Participants across all four focus groups talked about where they
happen; (iii) characteristics of victims and perpetrators; (iv) could go for help and/or advice, including Childline6 , Umbrella7 ,
healthy vs. abusive relationships; and (v) the various forms CSEA teachers, and family members. In addition to this practical
can take. Participants also reported that they would have liked knowledge, participants in three of the focus groups emphasized
more detail/information about related topics, such as consent. the importance of having reassurance about confidentiality
Seven themes (including 24 subthemes) are captured within this should they come forward and/or wish to speak to someone:
superordinate theme (see Figure 2). Themes are denoted by the When you’re growing up, you always know the people that you can
following subheadings, with further subthemes referred to in the turn to, but I think it’s good to have extra information, and maybe
body of the text4 . societies that you don’t know at all that you can go to anonymously
and like talk about it with because, sometimes, it’s things that you
Theme 1: Avoiding Harm don’t want to share with anyone you know because maybe you don’t
This theme summarizes a sense of young people becoming aware want to be judged or you think it’s going to change their opinion of
of the risks and the potential consequences of interactions, along you. But if you can go anonymously and say, “This is happening,”
with a suggestion of how their behavior may change as a result, and get help for that, that’s really good (FG3).
particularly with regard to social media use. It includes the
In two of the focus groups (each from a different school and
subthemes “Becoming aware” and “Being wary (Being careful on
as such representing a different year group), participants spoke
social media).” As one participant noted: “It [the performance]
about being able to spot the signs of someone struggling and/or
makes you aware of like. . .makes you aware of what might
having issues (both in themselves and friends/peers) after seeing
happen” (FG4). The extracts presented here were typical of many
the program, and how this shared recognition would enable them
responses where the word “aware” was used alongside a non-
to support one another:
specific sense of negative consequences:
say if certain. . .someone was to like text me or something, and I told
Just being more aware and like there’s. . .there’s people that claim
my friend, they would have that in their minds then too. I would
that they’re someone else and that you need to have a better mindset
5
4
For example, the theme “Avoiding Harm” contains two subthemes, namely The name of the victim character was changed in order to preserve the anonymity
“Becoming aware” and “Being wary (Being careful on social media).” “Being wary” of the theater-in-education company.
6
contained a further second-order subtheme (“Being careful on social media”). https://www.childline.org.uk/
7
Second-order subthemes appear in brackets. https://umbrella.uk.net/

The researchers devised ‘templates’ of codes based on the interview

questions and the themes that seemed likely to arise from these.
Transcripts were imported into NVivo12, and Researcher A analysed FG1

by assigning a code or descriptive label to each section within the
transcript. This resulted in a modified version of the original template, with
various codes added or amended in order to incorporate the content of the
transcript (FG1). For example, the codes ‘Explanations for risky behaviour’
and ‘Other sources of information’ were added to capture the information
some participants drew on more broadly.
Researcher B developed a separate template and coded FG1 and FG2 by

hand for the purpose of ensuring inter-rater reliability.
The research team met to compare and combine the two templates. The
comparison revealed that Researcher A and Researcher B had largely coded
the FG1 transcript consistently. Only two main areas (‘Explanation for
risky behaviours’ and ‘Views about characters’) highlighted differences -
the descriptive labels of subcodes were consistent, but the labels of the
themes representing the subcodes differed. This was discussed within the
research team, and the labels ‘Explanation for risky behaviours’ and ‘Views
about characters’ were agreed upon.
Researcher A proceeded to code the transcripts from FG3 and FG4, using
the amended template, making further minor modifications to it, where
necessary.
The FG2 transcript (originally analysed by Researcher B) was imported

into NVivo in order to complete the data set, however, due to an error in
file transfer, Researcher A had to partially re-code it. This resulted in some
further changes to the template, which were discussed within the research
team in order to reach agreement that the final set of codes accurately
reflected the focus group discussions.
Once the template had been finalised, the codes were organised into
themes that corresponded with the research questions.
FIGURE 1 | Overview of the steps by King (2004) used to guide the process of template analysis.

FIGURE 2 | Mindmap of themes and subthemes for the superordinate theme “information and detail delivered.”
have it, and other people would have it, so it’s like it’s distributed a In two of the focus group discussions, there was some debate
lot through our age group (FG1). around the degree of blame the victim deserved, with participants
often contradicting both each other and at times themselves:
Theme 3: Victims I feel like she deserves 20% of the blame, and he deserves 85%. [. . .]
This theme summarizes a common pattern across all four focus So, you can also put 80% of the blame on him, 5% on Cath, and then
group discussions which represents young people’s views of the that other 15% can just go to the school, her parents, and people like
victim’s character. It includes the subthemes “Victim blaming,” that who haven’t told her (FG1).
“Impact of abuse” (“Societal victim blaming”), and “Views of
I don’t think it was all down to him. I think, in the first place,
the victim” (“Naïve or stupid?,” “Innocent and uninformed,”
like social media is obviously a big platform that people talk about
“Pity and sympathy,” “Wanted things from perpetrator”), which with all this thing happening about vulnerability and stuff, but she
capture a range of often contradictory perspectives about Cath’s shouldn’t have done it in the first place. But even if she did, or if
actions and the causes behind them. she was put in that place without her even doing that, then it was
obviously his fault because he’s manipulated her. . .which is why it’s
Victim Blaming his fault (FG2).
Some participants in each of the focus groups made statements
which might be considered examples of victim blaming. They
Impact of Abuse
thereby held Cath at least partially responsible for being abused. The majority of participants appeared to be anticipating that
Many participants made reference to the warning signs the CSEA would have far-reaching consequences for the victim, and
perpetrator had displayed, and how they themselves had “known” spoke of Cath’s life being “ruined” or “wasted.” In two focus group
that “there was something wrong with him” (FG1). Some discussions, participants’ perceptions of these consequences
quotes illustrate how some participants appeared to suggest that seemed to be influenced by societal attitudes toward victims of
Cath’s failure to recognize these signs, and/or her “irresponsible” CSEA:
behavior, made her at least somewhat culpable for what followed: Because he said, he said something like, “Nobody’s going to want
damaged goods”—that’s what he said. And like that obviously hit
I personally think that the fact that she’s 14 and she went into a her because she knows it’s true. So, in that situation, she’d probably
bar and had like drinks, it was irresponsible in a way because she think to herself, “Is it better for me to just leave it, rather than being
shouldn’t have been like met someone online [. . .] but I did feel demeaned in society, or should I go out there and try and work
like really sympathy for her because she’s like obviously thought this harder to get out of it?” (FG2).
person trusted her and that. . .but then again, she was just being a
bit irresponsible, in my opinion. I did feel bad for her but she was a Views of Victim
bit irresponsible (FG3).
Participants expressed a variety of different views of and feelings
toward Cath, including: (a) pitying her, (b) seeing her as innocent,
I have empathy for her because he took it too far, but then I kind of
don’t because she went and met up with him and still stayed with (c) debating whether she was naïve or stupid, and (d) thinking
him after he lied about his age, and, em, even what he did the first that she wanted things from the perpetrator (see Table 1).
time, and still carried on (FG4). Interestingly (in light of the victim blaming demonstrated above),
participants in two of the focus groups suggested that it was not
However, while some participants’ apparent attribution of Cath’s fault that she lacked information about CSEA, and that
blame to Cath did not seem to shift, there was a sense that they others were to blame for failing to give her the relevant awareness
felt more empathy for her as the story progressed: and understanding that might have protected her.
At the start, I was kind of like. . .like I was kind of angry with her, Theme 4: Perpetrators
like what is she doing, like why is she talking to random people, and This theme summarizes the negative views young people
then, as it went on, I was like, yes, she made a mistake, but then it expressed about perpetrators of CSEA, and the character in the
shouldn’t have got this far, and I felt really bad for her (FG2). performance, respectively. It includes the subthemes “Blaming

TABLE 1 | Themes and quotes identified in relation to views of the victim. TABLE 2 | Themes and quotes identified in relation to views of the perpetrator.
Quotes Quotes
Pity and I pitied her. I mean, it was just sad, like she wasted so much Blaming the I really dislike him because like he’s like, what, 20, he knows like
sympathy time for her. . .like of her life, just because of one mistake she perpetrator faking his age is wrong, especially to a 14-year-old, because he
made. (FG1) knows that they’ll fall for it because of how young they are, so,
Because I generally think that she was just like, you know, trying obviously, he knew that was wrong. So, like I don’t really like
to get. . .trying to be happy and all, but like. . .things like didn’t him for doing that. (FG3)
go the way that she wanted, so I kind of felt sorry for her really. I don’t think she got herself in—like she was naïve in that she
(FG3) didn’t. . .she didn’t know to like. . .she didn’t block him
Innocent and She’s still like an innocent child who doesn’t understand like straightaway. However, it’s still his fault, like you can’t blame the
uninformed what’s going on in the world.” (FG3) girl. (FG1)
P1: Yeah. Because I don’t think anyone told Cath what, em, Dislike and Dehumanizing. To [make like] a 14-year-old girl have sex with
people would do to her. disgust men so he can get money—I thought it was absolutely
P2: Yeah, she hasn’t been educated correctly. (FG1) disgusting! Horrible! (FG3)
Naïve or P1: She was a stupid teenager. I don’t like him. I don’t like him. (FG2)
stupid? P2: No, she wasn’t! Manipulative Him talking to her constantly and like just saying things and
P1: Yes, she was. and dishonest muddling up her own words and making her think randomly,
P2: She wasn’t stupid, she was naïve—there’s a difference. making her confused herself, she won’t know, in that situation,
P1: Yeah, innocent but stupid, so, you know. . .naïve. . .. like she won’t know whether she did say yes or no, like she will
P2: No, naïve and stupid are two different things. (FG1) just be confused because he’s just constantly confused her and
Wanted things There was like someone who was being there and like kind manipulated her. (FG2)
from of. . .being that person that she could go to because there was He was lying the whole time. (FG1)
perpetrator no one else for her to go to. So, it’s like. . .she saw him as like a
cool person in a way because he was offering her like things
that like she couldn’t go and get herself, like alcohol and things an unhealthy relationship because you’re constantly living in the
like that. (FG3) fear that you’re going to be judged or you’re going to be hurt or
I think that like. . .you see like. . .she wanted that. . .love from
something if you don’t, if you don’t agree to it, which is probably
somebody, like for someone to comfort her and all that stuff,
because you could tell by the way—like she was going. . .like
how she felt (FG2).
she was on her page looking, like going with the intention to
look for males like. . .and then. . .. (FG2)
Participants briefly discussed the importance of consent,
however, they mainly referred to it as an aspect that characterizes
a healthy relationship. As the following extract illustrates, the
the perpetrator,” “Dislike and distrust,” and “Manipulative and majority of participants did not go on to define it, nor explained
dishonest.” Participants expressed almost globally negative views how they thought it was established, but merely discussed it in
of the character, with responses indicating that they blamed terms of its absence and the ensuing consequences thereof, as well
him for his actions, disliked and distrusted him, and viewed as instances where consent was not given:
him as manipulative and dishonest (see Table 2). It was notable there’s got to be consent off both people for it to happen, otherwise it
that despite the prevalence of victim blaming, it was clear can get a bit serious within court and police and stuff like that so. . ..
that participants in three of the focus groups blamed the (FG4)
perpetrator too.
Participants in all of the focus groups identified trickery
and manipulation as key features of abusive relationships, with
Theme 5: Relationships
many participants expressing surprise that a victim could be
Young people predominantly spoke about negative, abusive, or
manipulated into such situations without the use of overt force:
non-consensual aspects of romantic relationships, as captured
in the subthemes “What an unhealthy relationship looks like” I obviously knew that things like this happened and I was like
(“Consent”) and “Abusive relationships” (“Getting tricked and aware that people can get involved in things they don’t want to get
manipulated,” “Significance of emotional connection”). It was involved in, but I thought it was more like people would be forced or
of particular note that they were able to offer little explanation pressured into it, rather than doing it because they wanted to protect
and/or understanding of what a “healthy relationship” entails, the person that was getting them involved because they’d like tricked
them that much. I thought that was really kind of. . .terrible (FG3).
and mostly defined a “healthy relationship” in terms of what it
is not, thereby focusing minimally on any positive relationship Related to this was a sense of realization amongst participants
aspects. While present across three of the focus group discussions, that victims may have feelings for their abusers, and that this is
“What an unhealthy relationship looks like” was merely referred how they are able to exploit them, resulting in victims agreeing to
to marginally: things they would never otherwise have agreed to:
Because I think, if you feel like you can talk to someone and tell them When I watched the performance, I knew the majority of the
that you don’t want to or you do, that’s a healthy relationship, but if things that was going on, like I like understood it and I already
you’re constantly living in fear and thinking, oh, when I get home, was like aware of it, but the one thing that like I definitely got
I can’t. . .I can’t do something or I can’t do this because someone’s informed about was about, em, the way someone can use someone
going to say something or I’m going to be judged or something, that’s else to do something for them, like have sex with multiple people

and. . .because they know that that one person has like an emotional TABLE 3 | Themes and quotes identified in relation to suggestions for
connection to them (FG3). more information.
Quotes
Theme 6: The Extent and Variety of CSEA
Participants in two of the focus groups reported that they had Other forms And, you know, the people in this school very, very need a
learnt more about other potential forms of exploitation and of exploitation workshop or a drama piece on child pornography because these
and abuse little children are doing the most. They need lessons. . .. (FG1)
abuse through the program, explaining that they had known That’s also why I think we need to learn more information about
about single-perpetrator-to-single-victim abuse, but were not the other ones [forms of exploitation] because I wasn’t aware of
aware (prior to the program) of multiple victims being abused the, em, traveling with drugs [to other people] because they didn’t
by one perpetrator, and/or one victim being abused by multiple want to—again, I can’t remember what it’s called. But I wasn’t
aware of that at all, so that was something new to me, but that
perpetrators:
also made me realize, because it was also presented as one
. . .it doesn’t have to be the person at hand who is like abusing them that. . .like the most common, right now, especially for young
sexually or physically, it can be multiple people who they might not people, it really showed the fact that I feel like. . .sexual
exploitation isn’t presented too much, but it’s presented like a lot
even know (FG3).
more than everything else, and I feel like the others need to be
addressed just as much as sexual exploitation.” (FG3)
Theme 7: More Information Needed
Views of I thought. . .you know how he did the [hot-seating] for, em, Cath,
This theme summarizes the range of topics young people other we could have like done that on him to see what he would have
highlighted as wanting more information about. It consists characters done with. . .like why he was so desperate for money. (FG3)
of the subthemes “Other forms of exploitation and abuse,” I think I would have liked to have seen what her family thought
“Views of other characters,” and “More on consent.” There were about it after and like people in the community, so like how she
would go around. . . Like she said herself that people would still
various suggestions from participants regarding other forms make snide remarks or some people will try and help her, but I’d
of abuse they thought should be depicted, including county like to see it, like I’d like to see her carrying on with her life and
lines, radicalization/extremism, and indecent images of children. see how she gets through it or seeing how hard it is and how
Participants in two of the focus groups reported that they would her—I want to see like her family’s opinions as to how. . .if they
were being supportive or if they were like it’s also her fault. I
have liked to hear the views and experiences of other characters,
wanted to see that as well. (FG4)
such as the perpetrator and the victim’s parents (see Table 3). In
More on I think, even though that grooming was a very good topic, they
particular, and perhaps most significantly, consent was identified consent probably should have done like a different topic, like, like, em,
by some participants as an aspect deserving of and needing more consent and stuff like that, and more focus on consent and that
attention. As can be seen in the extracts in Table 3, participants because that’s stuff that happens more. . .. (FG1)
in two of the focus groups appeared to be confused about the Participant 1: Consent is a scary thing.
Researcher: Consent?
concept of consent and what it meant/represented.
Participant 1: Yeah, because women can withdraw consent
after—I can [. . .].
Format and Timing Researcher: Well, anyone can withdraw consent.
This category presents young people’s views of the accessibility, Participant 2: Anyone can withdraw—don’t just say women!
Participant 1: I know, I know, I know, but it’s mostly women that
appropriateness, and timeliness of the program, as well as
do it. (FG1)
how effective, enjoyable and believable they found it, and the Participant: I don’t get consent. . .like I don’t get it like. . . Not to
emotional impact it had on them. Six themes (including 20 like. . .I don’t want to sound rude, but like. . .
subthemes) are captured within this superordinate theme (see Researcher: That’s alright, go on?
Figure 3). Themes are denoted by the following subheadings, Participant: How do you just not say no? (FG2)
with further subthemes referred to in the body of the text.
Theme 1: Effectiveness ensured that everyone (including teachers) was exposed to the
This theme summarizes young people’s views of the effectiveness same information:
of the format of the program. It consists of two subthemes, if that happened to me, I’d talk to, they also have seen it as well.
namely “Use of theater helped” (“Connecting with characters,” I would have it, and other people would have it, so it’s like it’s
“Better than just being told,” “Memorable,” “Visualizing”) and distributed a lot through our age group (FG2).
“Useful to have it at school.” A view strongly endorsed across all
focus group discussions was the positive impact of using a live Theme 2: Enjoyment
theater performance to tell the story. Several participants made Participants in three of the focus groups expressed that the
reference to being able to “visualize” the story (due to the live performance had been “entertaining” (FG1) or enjoyable: “it was
element of the program), and that this along with the emotional a good show and all” (FG3).
connection to the characters made them more likely to remember
what they had seen. Furthermore, there was a sense that this Theme 3: Believability
facilitated participants putting themselves in the victim’s shoes This theme summarizes young people’s views on how believable
(see Table 4). they thought the program was. It consists of two subthemes,
Participants in one focus group reported that it was helpful namely “Believable” and “Not believable enough” (“It’s different
to see the program in the context of the school day, as this in real life,” “They made it look too easy”). While participants

FIGURE 3 | Mindmap of themes and subthemes for the superordinate theme “format and timing.”
in three focus groups felt that the program was believable, Theme 4: Emotional Impact
they thought that some aspects were unrealistic. For example, This theme summarizes a range of emotions young people
participants in two focus groups highlighted the evident described feeling when watching the program. It consists of
discrepancy between the perpetrator character’s true age and the four subthemes, namely “Upset and sad,” “Angry,” “Worried
age he claimed to be: “His age. Like he’s saying he’s 17 but he looks and scared,” and “Upsetting but necessary” (see Table 5).
like he’s 55. . . No one’s going to believe that.” (FG1). The same A common pattern identified across all focus group discussions
participants also reported that the performance made it look as was participants reporting to have found the performance
though it was easy to avoid perpetrators: upsetting and/or sad, as well as scary and/or worrying. In
two of the focus group discussions, a pattern of anger
I think also that, em, in the drama, it was too easy for him to
was also identified. However, there was a strong sense
do it. Like people need to understand that they ain’t just going to
do that. You’re going to like say, “Oh yeah, I wouldn’t let them
that the elements of the performance that triggered those
take me out like that.” They need to understand that they will emotional responses were necessary in order for the program to
work harder (FG1). be effective:
Children can’t tell what that is, like they don’t know it’s grooming
A pattern that was identified in three of the focus group
until you actually tell them what grooming is (FG1).
discussions was an acknowledgment by participants that reality
would be different to watching a play. More specifically, Participants in three focus groups felt that the delivery of the
participants realized that the warning signs that may have seemed program had been sufficiently sensitive, and again seemed to
obvious to them as an audience would be more difficult to spot in suggest that a degree of reality was important in delivering the
real life: “If like, obviously, watching it, you can kind of maybe see message:
where it’s going, but if you’re in that situation, you don’t have a
I think it was like quite sensitive in the way it was delivered. . .
clue what’s going to be happening in the future” (FG3).
But also, I think, with topics like this, there’s only a certain level
of sensitivity you can provide because, at the end of the day, it needs
to be a reality check. It needs to be. . .this happens and something
TABLE 4 | Themes and quotes identified in relation to use of theater. needs to be done about it, basically (FG3).
Quotes Theme 5: Not Taken Seriously

Connecting I feel it’s like easier to connect with the characters like in that While this theme was not very prominent across the focus group
with characters play, em, and it just made it much more effective. (FG2) discussions, there was a sense from some participants that the
I think just seeing it sort of performed, it just shows you what
happens and the emotional level of the—like, for example, what
was her name, Cath, Cath must have felt or how she [would be
TABLE 5 | Themes and quotes identified in relation to emotional impact.
scarred] for the rest of her life. (FG3)
Better than just Yeah. It’s better if you see it than rather reading it or like Quotes
being told listening to it because, like she said, you can visualize it and see
what they’re thinking and what they’re feeling. But when you’re Upset and sad I don’t know, I just felt really upset, like I was on the verge of
like, em, when you’re reading it, it’s like very hard to picture in tears. I was like. . .how can someone be so. . .sly and malicious
your head what’s happening and how they feel and stuff. (FG2) to manipulate someone like that? (FG2)
Visualizing It’s easier to visualize then. (FG1) It kind of made me feel like upset because. . .not because
Actually seeing the characters react, it just changes how like the personally but because I empathized with all the characters and
situation plays out. (FG1) I felt bad for them. (FG3)
Memorable But like. . .you only get this play like once, and the play is Angry It made me feel angry because they’re horrible people. (FG3)
effective because you. . .you have it once and you visualize it, I didn’t feel as like angry at him [character in the workshop] than
you see it, and it’s a good thing because you don’t forget it. I was with [the character of the perpetrator]. (FG2)
Like I could remember that he said she’s damaged goods, but Worried and Well, I was scared at one point. (FG1)
that only came to me because I actually remembered the play, scared It made me scared because it’s. . .like anybody that that can
and it kind of. . .and it kind of touched me. (FG2) happen to, so. . .. (FG4)

program had not been taken seriously by everyone. They reported they can be more careful, be more prepared, and not get themselves
that this may have been due to different reasons, represented into situations like this (FG3).
in the subthemes of “Funny and cringe” and “Some people are
immature” (see Table 6).
DISCUSSION
Theme 6: Timing
This theme summarizes the range of opinions expressed by young Findings from the present study suggest that young people who
people about the timeliness of attending the program. It consists attended a school-based theater-in-education program developed
of two subthemes, namely “Do it earlier (But not too early)” new awareness and knowledge around CSEA, including victims,
and “Now is the right time.” The question around timing of the perpetrators, unhealthy, and abusive relationships, as well as how
program generated considerable discussion in all of the focus to avoid harm and what to do “if bad things happen.” Young
groups, with participants generally agreeing that now was a good people demonstrated an increased awareness of these aspects
time to have seen the program, while also suggesting that it could during the focus group discussions, and clearly identified possible
be shown to younger children as well. Views on how much earlier ways of changing their behavior (by means of the subthemes
young people should receive the program were more varied. One “Being wary” and “Being careful on social media”), however,
participant endorsed seeing the program both now and later: “I unfortunately, it was not possible to determine whether this
think it was a good time now, but I think they should warn us like increased awareness translated into actual behavioral change.
again, so like for people who just didn’t take it seriously” (FG3). Young people reported that they would have liked to receive more
information about the concept of consent, and the views and
Do It Earlier (but Not Too Early) experiences of other characters, especially the perpetrator and
While several responses cautioned against the program being the victim’s parents. Furthermore, their responses to questions
delivered too early [i.e., “Not to primary schools.” (FG1)], about healthy relationships and consent suggest that they were
the majority of participants endorsed showing it to children predominantly defining these in terms of what they are not
younger than themselves, and suggested that if Cath had had (i.e., by referring to aspects characteristic of unhealthy and
more awareness of and information about CSEA when she was abusive relationships).
younger, this might have led to a different outcome for her: In terms of the format and timing of the program, the
majority of young people reported that the theater-in-education
Cath, she was technically in the situation in our year group. She
element of it helped them to connect with and remember the
didn’t know about it before that. So, if she was maybe like 8, 9, like so
it could start from there, and then having more conversations about story, as well as making it more real—“better than just being
it, it could progress until the full understanding, and like prevent told.” Young people generally felt that it was beneficial for the
stuff from happening, it would have had more of an effect for her to program to take place during school time, with everyone having
make the right decisions (FG2). to attend, and pupils and teachers thereby receiving and being
exposed to the same information. Young people in all the focus
Now Is the Right Time groups endorsed that they had received the program at the right
Despite some debate on the issue of timing, participants in all time. The majority of pupils (from both Year 9 and Year 10)
focus groups endorsed that the time at which they had seen the also agreed that it would be useful to have this information
program was right. There was no suggestion that the timing earlier (e.g., at the start of secondary school), suggesting that it
was poor and/or inappropriate, and participants in all focus should be shown to children younger than themselves. While
groups referred to the particular stage they are at in their lives, the focus group discussions featured some debate on how early
emphasizing the relevance of the program to this: was “too early,” United Kingdom prevalence data highlight
that around half of CSEA experienced by young people occurs
This is the age where things like this can actually start happening to prior to secondary school age (Office for National Statistics,
people we know. Like it can happen younger, it can happen older,
2020). This signifies that targeting younger year groups may
but around this age is when things can start, so it’s good to kind of
indeed be effective. In addition, some young people talked about
prepare people with the knowledge so then, as things go forward,
experiencing negative emotional responses/states, while at the
same time acknowledging that this was necessary and fitting
TABLE 6 | Themes and quotes identified in relation to not taken seriously. given the purpose and subject matter of the program. Others
reported that they and/or their classmates found the performance
Quotes “funny” and/or “cringe worthy,” suggesting that some young
Funny and That was the funniest part of the whole play: she goes, “I love people had not taken it seriously.
cringe you,” and he goes “I know.” (FG1) Overall, our findings suggest that a school-based theater-
Yeah, like. . .em, like half of it, the first half of it was like. . .they in-education program can lead to increased awareness and
thought it was like a bit funny and like, you know, cringe understanding of CSEA in young people, as well as what to do
[laughing]. (FG3)
“if bad things happen.” According to Fryda and Hulme (2014),
Some people Some people took it more seriously than others. (FG1)
knowledge gain is the most commonly captured outcome for
are immature I feel like. . .not. . .like out of the whole year, there’s going to be
two or three people that don’t maybe take the ideas on-board
SSAPPs undoubtedly due to the relative ease of measurement.
as well as everybody else because of maturity issues or However, Williams (2019) highlights that there is no indication
whatever, but you’re always going to get that. (FG3) of whether improving young people’s knowledge of CSEA leads to

an actual reduction in victimization. More specifically, Finkelhor highlight that several standardized measures have been used to
et al. (1995) argue that education that aims to prevent CSEA does evaluate SSAPPs, including the Children’s Knowledge of Abuse
not necessarily stop abuse from occurring, and that even when Questionnaire (Tutty, 1992) (used by Daigneault et al., 2012),
young people understand preventative measures and strategies and the “What If ” Situations Test (Saslawsky and Wurtele,
they may still go on to experience CSEA (Pelcovitz et al., 1992). 1986) (used by Chen et al., 2012; Daigneault et al., 2012) in
While the program involved in the present study appears to adapted forms. Adapting a standardized measure may be a
have contributed to young people’s increased awareness and means through which programs can measure their impact more
understanding of CSEA (and their hypothesizing about potential meaningfully, as well as enabling more accurate comparisons
behavioral change in the future), we were not able to determine across different interventions.
whether this translated into actual behavioral change. There is no fixed guidance as to what information SSAPPs
It is of particular note that the level of victim blaming present should contain. Topping and Barron (2009) identified a number
in young people’s responses during the focus group discussions of key aspects that appear across various SSAPs. Some of these
was unexpected, given the program’s aim of increasing young (e.g., being able to recognize CSEA, and knowing what to do
people’s empathy for victims of CSEA. Despite the fact that many when one experiences it) were evident in the program involved
expressed pity and sympathy for the victim in the program, in the present study. However, the authors emphasize that there
and a strong dislike and/or disgust for the perpetrator, the is a lack of evidence concerning the “differential effectiveness”
majority of young people still at least partially blamed her for of the key aspects they identified, and therefore called for more
being abused. However, there were interesting nuances in young research to be conducted into this. Complicating this issue further
people’s responses in that they reflected on the fact that some is more recent evidence that contradicts established narratives
manipulation may not be as overt, and therefore more subtle as of how perpetrators operate online, such as (i) perpetrators’
a result. While some young people blamed the victim for not main goal for engaging children and young people in sexually
heeding warning signs and engaging in “risky” behavior, others exploitative and abusive interactions online being “cybersex”
also felt that responsible others had failed in not providing her (rather than a physical meeting); (ii) the majority of perpetrators
with the awareness and understanding that would have allowed not using deception and/or hiding their true age online; and
her to protect herself. In addition, some young people recognized (iii) perpetrators not necessarily engaging potential victims in
that warning signs of CSEA may be much harder to identify and a friendship-forming and/or relationship-forming stage as part
spot in real life, however, interestingly, this was not reflected in of sexually exploitative and abusive interactions online (Kloess
their views/perceptions of the victim. et al., 2015). In addition, there appears to be a misperception
In particular, a certain level of awareness of the victim- among the general public and professionals of what constitutes
blaming attitudes held by society at large was evident in the “online grooming,” and sexual exploitation and abuse of children
discussions by some young people. This suggests that some via internet technologies, respectively. More specifically, research
of the attitudes and views they endorsed may originate from has shown that professionals who support young people in a
exposure to common societal discourses around sexual violence therapeutic capacity sometimes perceive “online abuse” to be
and consent. Unfortunately, British young people and adults alike less impactful and of less urgent concern than “offline abuse.”
are presented with many examples of victim-blaming narratives However, the same piece of research also discovered that “online
in the media, at times even from public sector organizations such abuse” can have just as much impact on young people as “offline
as the police and the judicial system (e.g., Halliday, 2019; Petter, abuse,” with additional psychological effects due the unique
2019). Challenging these is therefore likely to be a substantial and elements of the online environment, such as being in constant
enduring task for those who deliver school-based intervention contact with the perpetrator (particularly at night, which leads
programs, and we must also be realistic about how much to lack of sleep and subsequent exhaustion), and enduring fear
change is achievable in relation to this in young people, when a that explicit images may be distributed and made public online
substantial proportion of the adult world continue to subscribe to (Hamilton-Giachritsis et al., 2017).
these attitudes and views. In addition, research has demonstrated While the majority of SSAPPs focus on increasing awareness
a link between victim blaming and traditional views of gender and knowledge of a specific threat (e.g., CSEA) in school-aged
roles (Ben-David and Schneider, 2005), with people being more children and young people, other programs have concerned
likely to blame a woman for being sexually assaulted if they themselves with teaching pupils more generic life skills. More
perceive her to be non-conforming to stereotypically female specifically, in reviewing SSAPPs aimed at promoting internet
behaviors and traits (e.g., Kunst et al., 2018). This would suggest safety in the US, Finkelhor (2014) argues that teaching children
that confronting attitudes and perceptions around traditional and young people such skills (e.g., conflict management,
gender roles may be a route through which SSAPPs may be able to consequence anticipation, refusal techniques, and help-seeking)
open up a dialogue about these issues with school-aged children is a more effective preventative measure than targeted internet
and young people. safety education. This suggests that a skills-based approach may
In reviewing SSAPP evaluation methods, Fryda and Hulme be of more value in helping children and young people to
(2014) noted that it was rare for programs to use standardized avoid and/or safely navigate interactions online than any type
measures, usually devising their own and/or heavily adapting of program that seeks to promote internet safety. It is also
those in existence. The same is true of the company involved in important to note that standalone SSAPPs are unlikely to achieve
the present study, who developed their own questionnaires based long-term impact, and that more intensive programs of longer
on the aims of the program. However, Fryda and Hulme (2014) duration and/or repeated exposure have greater effectiveness

(Bovarnick and Scott, 2016). As such, it is important not to present study increased young people’s ability to spot signs
overstate their influence or value. In their review, Whittle et al. of CSEA, maintain their own safety, and feel confident in
(2014) also highlight the importance of family systems and seeking help and getting support. Nevertheless, no conclusions
secure attachment relationships, and emphasize that insecure can be drawn as to whether this will translate into actual
attachment relationships contribute to increasing vulnerability to behavioral change. So far, evaluations of SSAPPs (especially
CSEA in children and young people. Naturally, the issue of family those in the United Kingdom) have predominantly focused
dynamics and attachment relationships is difficult to mediate in on increasing awareness and understanding, as well as shifting
the context of SSAPPs. attitudes, rather than assessing behavioral change in children
Furthermore, Brown and Saied-Tessier (2015) make an and young people. While this is undoubtedly challenging to
important point by noting that despite the popularity of SSAPPs, achieve, it is necessary in order to justify the current spending
prevention should not be made the sole responsibility of children on SSAPPs, especially in light of the recent introduction of
and young people. The move to situate preventative measures the statutory requirement for both primary and secondary
and strategies within SSAPPs has been criticized by some for the schools to offer relationships and sex education. With this, there
inherent assumption that children and young people have the is the potential that the number of providers commissioning
ability and power to avoid being exploited and abused (Williams, them is going to rise. As such, future research would benefit
2019). The lack of conclusive evidence that SSAPPs prevent from exploring behavioral change in young people, as well
experiences of CSEA led Eaton and Holmes (2017) to suggest that as how young people of male/non-binary genders, diverse
they should not be seen as “preventative” (although they may still cultural/ethnic backgrounds, and differing intellectual abilities,
have a positive impact). Eaton (2017) takes an even more critical experience SSAPPs. In addition, evidence as to which aspects
stance by arguing that CSEA does not occur because children and and/or topics included in SSAPPs are most useful and effective
young people lack awareness, knowledge and/or understanding would allow existing programs to tailor their content accordingly,
thereof, but because perpetrators exploit and abuse them, and and thereby be in line with evidence-based practice. This would
that similar preventative measures and strategies do not stop be best explored by means of a large-scale quantitative study
adults from being abused. She further cautions that we must that ideally also records how attendance at SSAPPs translates
retain a critical perspective on organizations who stand to benefit into actual behavioral change in young people. Last, but not
from the narrative that mere education can prevent CSEA. least, future research would benefit from identifying effective
While the present study offers unique insights into young strategies to combat victim blaming in school-aged children
people’s experiences of attending a school-based theater-in- and young people.
education program, and the perceived impact this had on their
awareness and understanding of CSEA, there are a number of
limitations that require acknowledging. Firstly, young people DATA AVAILABILITY STATEMENT
self-selected to take part in the focus group discussions, which
may suggest that they were generally more engaging, or had been The datasets presented in this article are not readily available
impacted by the program. This group of young people is therefore due to the focus group discussions representing personal
not representative of the year groups of Years 9 and 10 overall. perspectives and experiences of children. It is therefore not
Secondly, given the very few male participants in the present deemed appropriate for the data to be made publicly available.
study, it was not possible to determine differences between male Requests to access the datasets should be directed to JK,
and female young people in terms of the attitudes they held, the J.A.Kloess@bham.ac.uk.
views they endorsed, as well as the level of impact the program
had on them. In light of existing research demonstrating a link
between victim blaming and traditional views of gender roles, this ETHICS STATEMENT
would have been interesting to explore in more depth. Thirdly,
the present study did not consider the needs of young people The studies involving human participants were reviewed
with learning disabilities, or those in non-mainstream education and approved by the Science, Technology, Engineering, and
(e.g., pupil referral units). Research indicates that young people Mathematics Ethical Review Committee at the University of
with learning disabilities are more vulnerable to experiencing Birmingham, and the Psychology Research Ethics Committee
CSEA, and at a disadvantage with regard to accessing support at the University of Bath. Written informed consent to
and protection (Franklin et al., 2015). Finally, Brown and Saied- participate in this study was provided by the participants’ legal
Tessier (2015) point out that SSAPPs rarely deal with intra- guardian/next of kin.
familial CSEA or harmful sexual behavior engaged in by young
people, despite evidence suggesting that these are significant areas
of concern. Taken together, this adds to the complexity in terms AUTHOR CONTRIBUTIONS
of how to identify and evaluate the most useful and effective
content for SSAPPs. JK conceived and designed the project, with support from
There are a number of complex factors and issues that CH-G. HM and KD collected the data, with support from JK.
impact on the interpretation of our findings. However, overall, HM and KD analyzed the data under the supervision of JK
they tentatively indicate that the program involved in the and CH-G. HM completed the majority of the write-up, with

contributions from JK, KD, and CH-G. JK and CH-G edited the and the Institute of Global Innovation at the University
different versions of the manuscript. JK prepared the manuscript of Birmingham.
for submission to Frontiers in Psychology. All authors were
involved in agreeing the final coding template.
ACKNOWLEDGMENTS
FUNDING The authors would like to express their immense gratitude and
appreciation to the theater-in-education company and the two
The study presented here was supported in part schools for their assistance, time, and effort in supporting the
by pump-priming funding from the School of Psychology study and facilitating the process of data collection.
REFERENCES Eaton, J. (2017). Why ‘CSE Awareness’ Will Never Prevent CSE. Available
online at: https://victimfocusblog.com/2017/05/29/why-cse-awareness-will-
Barnardo’s (2014). Report of The Parliamentary Inquiry Into The Effectiveness never-prevent-cse/ (accessed May 29, 2017).
of Legislation for Tackling Child Sexual Exploitation and Trafficking Eaton, J., and Holmes, D. (2017). Working Effectively to Address Child
Within the UK. Available online at: https://www.sericc.org.uk/pdfs/3339_ Sexual Exploitation: An Evidence Scope. Available online at: https://www.
cseparliamentaryinquiryreport.pdf (accessed March 27, 2020). researchinpractice.org.uk/media/2613/cse_evidencescopeupdateoct17_v2.pdf
Beckett, H., Holmes, D., and Walker, J. (2017). Child Sexual Exploitation: (accessed April 06, 2020).
Definition & Guide for Professionals: Extended Text. Available online Finkelhor, D. (2014). Commentary: cause for alarm? Youth and internet risk
at: https://uobrep.openrepository.com/bitstream/handle/10547/623178/UOB- research – a commentary on Livingstone and Smith. J. Child Psychol. Psychiatry
RIP-CSE-GuidanceFeb2017.pdf?sequence=3&isAllowed=y (accessed February 55, 655–688. doi: 10.1111/jcpp.12260
20, 2020). Finkelhor, D., Asdigian, N., and Dziuba-Leatherman, J. (1995). Victimization
Bell, J., and Stanley, N. (2006). Learning about domestic violence: young people’s prevention programs for children: a follow-up. Am. J. Public Health 85, 1684–
responses to a healthy relationships programme. J. Sex Edu. 6, 237–250. doi: 1689. doi: 10.1016/0145-2134(94)00112-8
10.1080/14681810600836356 Franklin, A., Raws, P., and Smeaton, E. (2015). Unprotected, Overprotected:
Bellis, M., Hughes, K., Leckenby, N., Perkins, C., and Lowey, H. (2014). National Meeting the Needs of Young People With Learning Disabilities Who
household survey of adverse childhood experiences and their relationship with Experience, or Are at Risk of, Sexual Exploitation. Available online at:
resilience to health-harming behaviors in England. BMC Medicine 12:72. doi: https://www.barnardos.org.uk/sites/default/files/uploads/Unprotected%
10.1186/1741-7015-12-72 2C%20overprotected%20-%20meeting%20the%20needs%20of%20young%
Ben-David, S., and Schneider, O. (2005). Rape perceptions, gender role attitudes, 20people%20with%20learning%20disabilities%20who%20experience%
and victim perpetrator acquaintance. Sex Roles 53, 385–399. doi: 10.1007/ 2Cor%20are%20at%20risk%20of%2C%20sexual%20exploitation%202015.pdf
s11199-005-6761-4 (accessed May 18, 2020).
Bovarnick, S., and Scott, S. (2016). Child Sexual Exploitation Prevention Fryda, C., and Hulme, P. (2014). School-based childhood sexual abuse prevention
Education: A Rapid Evidence Assessment. Available online at: https: programs. J. Sch. Nurs. 31, 167–182. doi: 10.1177/1059840514544125
//uobrep.openrepository.com/bitstream/handle/10547/623175/FINAL-REA-1. Gordon, R., and Gere, D. (2016). Sex squad: engaging humour to reinvigorate
pdf?sequence=3&isAllowed=y (accessed April 07, 2020). sexual health education. Sex Edu. 16, 1–13. doi: 10.1080/14681811.2015.
British Psychological Society (2018). Code of Ethics and Conduct. Available 1120193
online at: https://www.bps.org.uk/sites/www.bps.org.uk/files/Policy/Policy% Hailes, H., Yu, R., Danese, A., and Fazel, S. (2019). Long-term outcomes of
20-%20Files/BPS%20Code%20of%20Ethics%20and%20Conduct%20% childhood sexual abuse: an umbrella review. Lancet Psychiatry 6, 830–839.
28Updated%20July%202018%29.pdf (accessed August 13, 2020). doi: 10.1016/S2215-0366(19)30286-X
Brooks, J., McCluskey, S., Turley, E., and King, N. (2015). The utility of template Halliday, J. (2019). Nottinghamshire Police ‘Victim Blaming’ Women Who Walk
analysis in QUALITATIVE psychology research. Qual. Res. Psychol. 12, 202– Alone. The Guardian. Available online at: https://www.theguardian.com/
222. doi: 10.1080/14780887.2014.955224 uk-news/2019/nov/27/nottinghamshire-police-victim-accused-of-blaming-
Brown, J., and Saied-Tessier, A. (2015). Preventing Child Sexual Abuse: Towards A women-who-walk-alone (accessed November 19, 2019).
National Strategy for England. Available online at: https://learning.nspcc.org.uk/ Hamilton-Giachritsis, C. E., Hanson, E., Whittle, H., and Beech, A. R. (2017).
media/1102/preventing-child-sexual-abuse-towards-a-national-strategy.pdf Impact of Online and Offline Child Sexual Abuse: “Everyone Deserves to be
(accessed February 20, 2020). Happy and Safe”. Available online at: https://learning.nspcc.org.uk/media/
Chen, Y., Fortson, B., and Tseng, K. (2012). Pilot evaluation of a sexual abuse 1123/impact-online-offline-child-sexual-abuse.pdf (accessed February 20,
prevention program for Taiwanese Children. J. Child Sexual Abuse 21, 621–645. 2020).
doi: 10.1080/10538712.2012.726699 King, N. (2004). “Using templates in the thematic analysis of text,” in Essential
Daigneault, I., Hébert, M., McDuff, P., and Frappier, J. (2012). Evaluation of a Guide to Qualitative Methods in Organizational Research, eds C. Cassell and G.
sexual abuse prevention workshop in a multicultural, impoverished urban area. Symon (London: Sage), 256–270. doi: 10.4135/9781446280119.n21
J. Child Sexual Abuse 21, 521–542. doi: 10.1080/10538712.2012.703291 Kloess, J. A., Seymour-Smith, S., Hamilton-Giachritsis, C. E., Long, M. L., Shipley,
Department for Education (2017). Child Sexual Exploitation: Definition and D., and Beech, A. R. (2015). A qualitative analysis of offenders’ modus operandi
Guide for Practitioners. Available online at: https://www.gov.uk/government/ in sexually exploitative interactions with children online. Sex. Abuse 29, 563–
publications/child-sexual-exploitation-definition-and-guide-for-practitioners 591. doi: 10.1177/1079063215612442
(accessed January 17, 2020). Krahé, B., and Knappert, L. (2009). A group-randomized evaluation of a theatre-
Department for Education (2019). Relationships Education, Relationships based sexual abuse prevention programme for primary school children in
and Sex Education (RSE) and Health Education: Statutory Guidance for Germany. J. Community Appl. Soc. Psychol. 19, 321–329. doi: 10.1002/casp.
Governing Bodies, Proprietors, Head Teachers, Principals, Senior Leadership 1009
Teams, Teachers. Available online at: https://assets.publishing.service.gov. Kunst, J., Bailey, A., Prendergast, C., and Gundersen, A. (2018). Sexism, rape
uk/government/uploads/system/uploads/attachment_data/file/908013/ myths and feminist identification explain gender differences in attitudes toward
Relationships_Education__Relationships_and_Sex_Education__RSE__and_ the #metoo social media campaign in two countries. Media Psychol. 22:5. doi:
Health_Education.pdf (accessed February 19, 2020). 10.1080/15213269.2018.1532300

Lightfoot, S. F., Taboada, A., Taggart, T., and Burtaine, A. (2015). ‘I learned to http://sheu.org.uk/content/it-opened-my-eyes-using-drama-theatre-
be okay with talking about sex and safety’: assessing the efficacy of a theatre- education-deliver-sex-and-relationship-education (accessed January 17,
based HIV prevention approach for adolescents in North Carolina. Sex Edu. 15, 2020).
348–363. doi: 10.1080/14681811.2015.1025947 Stoltenborgh, M., van IJzendoorn, M., Euser, E., and Bakermans-Kranenburg, M.
Maniglio, R. (2009). The impact of child sexual abuse on health: a systematic review (2011). A global perspective on child sexual abuse: meta-analysis of prevalence
of reviews. Clin. Psychol. Rev. 29, 647–657. doi: 10.1016/j.cpr.2009.08.003 around the world. Child Maltreat. 16, 79–101. doi: 10.1177/107755951140
National Society for the Prevention of Cruelty to Children [NSPCC] (2019). Over 3920
9,000 Police-Recorded Online Child Sexual Abuse Offences. Available online Topping, K., and Barron, I. (2009). School-based child sexual abuse prevention
at: https://www.nspcc.org.uk/what-we-do/news-opinion/web-exploited-by- programs: a review of effectiveness. Rev. Edu. Res. 79, 431–463. doi: 10.3102/
child-sex-offenders/ (accessed January 28, 2020). 0034654308325582
Office for National Statistics (2020). Child abuse extent and nature, Tutty, L. M. (1992). The ability of elementary school children to learn child sexual
England and Wales: Year ending March 2019. Available online at: abuse prevention concepts. Child Abuse Negl. 16, 369–384. doi: 10.1016/0145-
https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/ 2134(92)90046-T
articles/childabuseextentandnatureenglandandwales/yearendingmarch2019 Walsh, K., Zwi, K., Woolfenden, S., and Shlonsky, A. (2015). School-based
(accessed April 07, 2020). education programmes for the prevention of child sexual abuse. Cochrane
Pelcovitz, D., Adler, N. A., Kaplan, S., Packman, L., and Krieger, R. (1992). The Database of Syst. Rev. 4:CD004380. doi: 10.1002/14651858.CD004380.
failure of a school-based child sexual abuse prevention program. J. Am. Acad. pub3
Child Adolesc. Psychiatry 31, 887–892. doi: 10.1097/00004583-199209000- Whittle, H. C., Hamilton-Giachritsis, C. E., and Beech, A. R. (2014). In
00017 their own words: young peoples’ vulnerabilities to being groomed and
Petter, O. (2019). ‘The Price You Pay Could Be Rape’: Police Campaign sexually abused online. Psychology 5, 1185–1196. doi: 10.4236/psych.2014.
Accused of ‘Victim Blaming’. The Independent. Available online at: 510131
https://www.independent.co.uk/life-style/west-yorkshire-police-rape-prie- Williams, M. (2019). The NSPCC’s Protect & Respect Child Sexual Exploitation
free-drinks-drugs-twitter-party-animals-victim-blaming-a8822736.html Programme: A Discussion of The Key Findings From Programme Implementation
(accessed March 4, 2019). And Service Use. Available online at: https://learning.nspcc.org.uk/media/1699/
Public Health Wales NHS Trust (2015). Adverse Childhood Experiences and a-discussion-key-findings-programme-implementation-service-use.pdf
Their Impact on Health-Harming Behaviours in the Welsh Adult Population. (accessed April 20, 2020).
Available online at: http://www2.nphs.wales.nhs.uk:8080/PRIDDocs.nsf/ Wooster, R. (2016). Theatre in Education in Britain. London: Bloomsbury Methuen
7c21215d6d0c613e80256f490030c05a/d488a3852491bc1d80257f370038919e/ Drama. doi: 10.5040/9781472591517
\protect\T1\textdollarFILE/ACE%20Report%20FINAL%20(E).pdf (accessed
April 20, 2020). Conflict of Interest: The authors declare that the research was conducted in the
Quayle, E., Jonsson, L., and Lööf, L. (2012). Online Behaviour Related to Child absence of any commercial or financial relationships that could be construed as a
Sexual Abuse: Interviews With Affected Young People. Available online potential conflict of interest.
at: https://www.childrenatrisk.eu/robert/public/Interviews_with_affected_
young_people.pdf (accessed January 17, 2020). Copyright © 2021 May, Kloess, Davies and Hamilton-Giachritsis. This is an open-
Saslawsky, D. A., and Wurtele, S. K. (1986). Educating children about sexual abuse: access article distributed under the terms of the Creative Commons Attribution
implications for pediatric intervention and possible prevention. J. Pediatr. License (CC BY). The use, distribution or reproduction in other forums is permitted,
Psychol. 11, 235–245. doi: 10.1093/jpepsy/11.2.235 provided the original author(s) and the copyright owner(s) are credited and that the
Sawney, F., Sykes, S., Keene, M., Swinden, L., and McCormick, G. (2003). original publication in this journal is cited, in accordance with accepted academic
It Opened My Eyes: Using Theatre in Education to Deliver Sex and practice. No use, distribution or reproduction is permitted which does not comply
Relationship Education: A Good Practice Guide. Available online at: with these terms.

SYSTEMATIC REVIEW
published: 09 February 2021
doi: 10.3389/fpsyg.2021.606731
Prioritizing Indecent Image

Offenders: A Systematic Review and
Economic Approach to Understand
the Benefits of Evidence-Based
Policing Strategies
Susan Giles* and Laurence Alison
Department of Psychology, Institute of Population Health, University of Liverpool, Liverpool, United Kingdom
In 2013, there were an estimated 50,000 individuals involved in downloading and sharing
indecent images of children (IIOC) in the United Kingdom (UK). This poses challenges for
limited police resources. We argue that police officers can make most effective use of
limited resources by prioritizing those offenders who pose the greatest risk of contact
offending, by nature of demonstrable pedophilia, hebephilia or dual offending status
and thus, those at highest risk must be dealt with first. What is currently lacking is
Edited by:
Noora Ellonen, a clear idea of the potential scale of the problem in socio-economic terms and why,
University of Tampere, Finland therefore, it is so important that evidence-based approaches to offender detection and
Reviewed by: investigation continue to be a top priority for funders and policy makers. A systematic
Joshua D. Behl,
literature review was undertaken to address two related questions. First, what is the scale
Flagler College, United States
Peter Simonsson, of the problem in the UK, in terms of the number of pedophilic and hebephilic individuals
Temple University, United States who pose a risk of contact offending against a child? Second, what is the potential
Jason Roach,
University of Huddersfield, socio-economic burden generated by the national IIOC suspect pool if left unattended
United Kingdom to by targeted police action? Applying population estimates of pedophilia and hebephilia
*Correspondence: to the male population (16–89 years), we estimate there are between 2,365–5,991
Susan Giles
males with paedophila and 12,218–30,952 males with hebephilia who are likely contact
spgiles@liverpool.ac.uk
offenders. Applying average prevalence and incidence based costing methods to a
Specialty section: conservative estimate of one victim per offender, the combined socio-economic burden
from these persons could amount to £236-£597 million (incident costs) increasing to
a section of the journal £2.9-£7.3 billion (lifetime costs; £3.3-£8.3 billion including QALY measures). Applying
Frontiers in Psychology the same costs to CEOP (2013) estimate of 50,000 IIOC offenders we estimate that
Received: 15 September 2020 between 6,000 and 27,500 dual offenders could have already committed past contact
Accepted: 12 January 2021
Published: 09 February 2021
offenses, contributing an economic burden of between £97–£445 million (incident costs)
Citation:
increasing to £1.2–£5.4 billion (lifetime costs; £1.4–£6.2 billion including QALY measures).
Giles S and Alison L (2021) Prioritizing Future contact offenses could contribute a further burden of £16–£18.6 million (incident
Indecent Image Offenders: A
costs) increasing to £198–£227 million (lifetime costs; £226–£260 million including QALY
Systematic Review and Economic
Approach to Understand the Benefits measures). Drawing upon these findings, we argue for the benefits of a research-informed
of Evidence-Based Policing prioritization approach to target IIOC offenders.
Strategies.
Front. Psychol. 12:606731. Keywords: indecent images of children, child sexual abuse, child sexual abuse material, risk assessment, arrest
doi: 10.3389/fpsyg.2021.606731 prioritization, police decision making

Giles and Alison Proritizing IIOC Offenders
INTRODUCTION that IIOC offending does not tend to lead to contact sexual
offenses but that the converse is not true; specialization is
The number of indecent image of children (IIOC) offenses, in most evident for IIOC offenders but contact sexual offenders
recent years, has created a growing workload for the police against a child will “cross over” to IIOC offending. Further,
officers investigating them. Convictions for IIOC offenses are frequency of pornography use (including offline IIOC) has been
increasing (Sentencing Council, 2012) and police forces are shown to contribute to the prediction of crime and violent
struggling to manage the demand and risks posed by this (including sexual) recidivism for child contact sexual offenders
offending group. In the United States of America (USA), (Kingston et al., 2008). This suggests a general paraphilic
for example, Waters testified to United States Congress that lifestyle for some high risk dual offenders and that police
despite 500 000 national offenders being identified, only 2% resources should be focused on dual offenders. Set against this,
of cases were being investigated due to a lack of police there is still a recognized risk that those with IIOC offenses
resources (Committee on the Judiciary, 2008). He described the may cross over to contact offending. In their meta-analysis,
situation as “catastrophic” as the magnitude of the problem Babchishin et al. (2014) propose that the psychological factors
had already overwhelmed USA law enforcement’s forensic and that differentiate those who do and do not act on their interests
investigative infrastructure. In the United Kingdom (UK), should be related to propensities for rule violation. Those most
the Child Exploitation and Online Protection Center (CEOP, at risk for cross over offenses would be expected to have high
2013) received 8,000 reports in 2012 that contained a total levels of pedophilia, high levels of anti-sociality, have access
of 70,000 still and moving IIOC. This represents a two- to children, and have few psychological barriers to acting on
fold increase from the previous year. They estimated that their deviant impulses. Those least at risk would experience the
in 2012, there were around 50,000 individuals involved in converse. These findings provide useful insights into potential
downloading and sharing IIOC in the UK (CEOP, 2013). decision criteria, although a limitation of Babchishin et al.
More recently, 39 police forces recorded 17,521 obscene (2014) meta-analysis is that up to half of IIOC offenders could
publication offenses against children in the year ending March be expected to have undetected offline offenses (Seto et al.,
2019 and identified 552 unique victims within IIOC (Elkin, 2011). As such, the profile of IIOC offenders is recognized as
2020). These figures help to provide some indication of diverse and the management of offenders with IIOC offenses
the size of the task now faced by British and international should carefully consider the existence of concomitant contact
investigating authorities. sexual offenses.
Safeguarding children is at the forefront of the policing The scale of the IIOC sex offending problem, both now
agenda; with the aim of protecting children from abuse in and in the long term, has prompted multidisciplinary research
the first place (where possible) and reducing the likelihood of in the area, with collaborations between academics and police
reoffending and repeat victimization. CEOP (2013) claim that forces to help improve detection and direct the allocation of
20% of the images they received in 2012 had been assessed policing resources to the most concerning cases. Researchers
as being self-generated by offenders; raising concerns about such as Steel (2009), Hughes et al. (2006) Brennan, Hammond
contact offending and the need to identify and safeguard and colleagues (Hammond et al., 2009; Brennan and Hammond,
these children. Set against this, meta-analysis by Seto et al. 2011, 2017) are beginning to explore the investigative and
(2011) confirms that there is a considerable subgroup of inferential value of examining pedophilic-related search terms
offenders who offend online only (by downloading and sharing and other online behavior amongst IIOC offenders. Hammond
images) and who are unlikely to present a risk of contact et al. (2009) and Brennan and Hammond (2017) for example,
offending. This article argues that evidence based policing suggest that there is evidence of quite discrete paraphilic
approaches can help prioritize those offenders at highest risk sexual interests to be found in offenders’ online behavior.
of committing contact offenses and in so doing, demonstrate Their analysis of 119,869 search terms and 3,000,000 records
the most efficient use of limited public sector resources toward (representing one week of activity on a P2P site) revealed
safeguarding goals. seven class taxonomy of paraphilic use; indicating the presence
Seto et al. (2011) found approximately one in six internet of distinct paraphilic sub-communities in the P2P network
offenders are expected to have a history of contact sexual offenses. space (Hammond et al., 2009). This included pedophilic,
Criticality is important here; researchers report substantial hebephilic, gerontophilic, bestiality, sadistic, rape and incest sub-
variation in reported prevalence rates dependent upon what communities. The concern here is that whilst there is a wide
method is used. Prevalence rates reduce to one in eight when range of offenses, pedophilic individuals are at increased risk
using official data and increases to one in two when using self- of contact offenses against children. Research has demonstrated
reports. As Seto et al. (2011) meta-analysis is primarily based the serial nature of pedophilia, the large number of children
on studies involving official data it is likely that a prevalence abused by each pedophilic individual and the under reporting
rate of one in six is also an underestimate. These figures do not of assaults (e.g., Abel and Harlow, 2001; Hall and Hall,
take account of undetected offenders. An analysis of undetected 2009). Hammond et al. (2009) suggest that effective targeting
offenders by Neutze et al. (2012) suggests a higher rate of strategies, such as those that draw upon the action of distinct
undetected dual (offline and online) offending than would be online sub-communities, may be a more effective strategy
suggested by Seto et al. (2011) meta-analysis. Research examining for targeting offenders than a random or “scatter gun” (pg.
specialization and “cross over” (Howard et al., 2014) suggest 7) approach.

An alternative approach espoused by the Fighting MATERIALS AND METHODS

International Internet Pedophilia (FIIP) project, a collaboration
between Kent Police and academics at the University of The broad objective of the work is to estimate the potential scale
Liverpool, is to prioritize the investigation of known IIOC of the problem in the UK by establishing offending and economic
offenders who share features with dual offenders (those who figures that could form the basis of sensible cost estimates. As
have committed IIOC and contact offenses). This approach has the body of literature has not been comprehensively reviewed
led to the development of the Kent Internet Risk Assessment for these purposes, a scoping review was preferred. This would
Tool (KIRAT: Long et al., 2016). KIRAT does not purport allow a preliminary assessment of the size and scope of the
to predict future risk or re-offending. It is a four-tiered existing literature and the extent to which it could be used to
decision tool used nationally by UK Police to prioritize IIOC answer the two research questions. Where gaps in knowledge
offenders, based on nine features most often associated with are identified, we can draw on the available evidence to make
dual offending, with upwards of 95% accuracy for higher informed decisions about derived estimates. As such, neither
risk offenders and <20% false positive rates for lower risk narrative nor quantitative synthesis were the primary aims of this
offenders. Recent cross-cultural validations have successfully paper. Quality assessment and data extraction that you might
demonstrated the relative homogeneity of KIRAT risk factors anticipate from systematic review or rapid evidence assessment
for contact offending across 24 countries. A further recidivism was therefore, limited. Rather, we follow in the economics
tool (CPORT) with investigative value has also been developed tradition; producing a systematic literature review to help provide
by Michael Seto and colleagues (Seto and Eke, 2015; Eke an overview of the available evidence from which we make critical
et al., 2019) and has recently been applied in a Spanish decisions about which figures to use in our cost estimates. The
context (Soldino et al., 2019). estimates that we provide should be seen as a “starting point”
The approaches outlined above demonstrate that it important of a much wider discussion about the benefits of police action
to have a conceptually driven and empirically supported in this area and how these benefits may be measured. Toward
approach about the links between online behavior and risk, this end, the first author undertook a systematic literature review
which guides practical action. Whilst lower risk cases still in August 2020, to identify relevant academic and gray literature
need to be processed through the legal system (Long et al., that would inform our research questions.
2016) police officers can facilitate optimal attainment of harm
reduction goals by prioritizing those offenders who pose the Search Strategy
greatest risk of contact offending, by nature of demonstrable The search strategy included setting objectives and specifying
pedophilia or dual offending status and thus, those at highest inclusion and exclusion criteria for reviews, designation of search
risk must be dealt with first. A random approach would terms and databases to be searched. The following databases
arguably be less effective beyond the general deterrence that were examined for published material 2000–2020; Cochrane
it may provide. library, Web of Science, Scopus, PubMed, APA PsychINFO and
What is currently lacking is a clear idea of what the National Criminal Justice Reference System (NCJRS). There
potential scale of the problem is in socio-economic terms. followed conduct of searches, de-duplication, application of
Child sexual abuse victimization impacts society (in terms screening procedures and compilation of target lists for further
of costs of supporting victims) along with the long term review. Material would be excluded if full text was unavailable
economic development of victims (in terms of mental well- in English, or unavailable through the University of Liverpool
being, education and employment). Economic metrics that online library services. Where possible, the search results were
seek to establish the disease burden associated with child exported into reference management software, Endnote Online.
sexual abuse encapsulate this interdependence by establishing Search teams were applied to online searches (Google/Google
costs to society along with the less tangible emotional costs Scholar) to identify soft literature available in the public domain.
to victims. As such, we prefer the term socio-economic Decisions were made by the research team to select economic and
to describe our approach. By establishing the scale of the prevalence figures that would best provide a sensible basis for cost
problem in socio-economic terms we aim to demonstrate the estimates. In the discussion that follows we provide an account of
importance of evidence-based approaches to offender detection our search strategy, the relevant material that was found along
and investigation and that they should continue to be a top with a critical discussion of that material that led to our decision
priority for funders and policy makers. The aim of the current about which material (prevalence, contact offending, economic
article is to expand and develop upon this debate by drawing figures) to include in our cost estimates.
on published prevalence studies and econometrics to present
some of the economic burden that could be contributed by Estimating Victim Costs
IIOC offenders. It draws on a systematic review and economic A key objective of the review was to estimate the average victim
approach to examine two research questions. First, what is cost for child sexual assault including children who are sexually
the potential scale of the problem in the UK, in terms of assaulted during the commission of a first generation IIOC
the number of and socio-economic burden of males with offense (the offender records the abuse). If arrest prioritization
pedophilia and hebephilia who pose a risk of contact offending tools are effective by picking out contact offenders then these
against a child? Second, what is the socio-economic burden costs to the victim and the National Exchequer can be averted;
potentially generated by the national suspect pool of 50,000 constituting costs saved for future contact victims, Historical
IIOC offenders? victims can also be safeguarded from repeat victimization.

Material would be included if it provided an economic analysis important to select a study that incorporated both tangible and
of the impacts of sexual harm that could be extracted directly intangible costs. From these, Saied-Tessier (2014), Letourneau
from the published account. Econometrics emerging from sexual et al. (2018) and Heeks et al. (2018) were considered for inclusion
assaults against children and adults were both considered on in further economic analysis given their focus on either child
the basis that the latter is a more established field of study and sexual abuse, the full range of costs included and/or focus on the
could shed light on costs experienced by child victims. Material United Kingdom (UK).
would be excluded if it did not provide unit level costs (i.e., Saied-Tessier (2014) provides a comprehensive review of
costs per victim). The first author applied Boolean operators to psychological and physical health problems experienced by
search terms to generate search phrases (cost AND victim AND victims of child sexual abuse in her prevalence based assessment
sex∗ crime OR rape OR child sex∗ abuse OR indecent image∗ OR of child sexual abuse in 2014. She argues that costs per victim
internet sex∗ offend∗ OR online sex∗ offend OR contact sex often do not appear to exist for the UK and represents a much needed
OR groom∗ OR chat room off∗ or solicitation off∗ OR molest∗ area of research. This NSPCC report focuses on annual costs and
OR pedo∗ OR paedo∗), resulting in 523 items. These items whilst unit costs are provided, where available, individual level
were then screened for filtering, such that items were excluded data is arguably still too underdeveloped for the purposes of the
based on duplication, or on the basis that the full text did not present report. As such, Saied-Tessier (2014) was not utilized for
inform the research question. Subsequent screening of 50 items further analysis. As such, a decision was made to calculate lower
was undertaken by the first author. A great wealth of material bound costs from the prevalence based (annual) costs provided
was available around the impact of sexual harm on victims, the by Heeks et al. (2018); (UK) and upper bound costs from the
relative costs of reporting and non-reporting of sexual abuse, and incidence based (lifetime) cost estimates provided by Letourneau
factors affecting disclosure of abuse. Promising work is beginning et al. (2018) (United States).
to emerge around the vicarious costs associated with being a Heeks et al. (2018) provide sexual offense unit costs of victims
victim of an IIOC offense. However, this work is arguably too in the UK. The Home Office focus takes account of a wide range
underdeveloped for the purposes of the present paper as it relies of child sexual offenses (such as intercourse with a female under
on US legal case studies or potential restitution frameworks 13 years/16 years, incest and indecent assault against a minor)
(Jacques, 2011; Cassell et al., 2013). As such, these references in its definition of sexual offenses and the figures below can be
were excluded. Twenty three references provided unit costs for applied to cases of child sexual abuse. However, the total burden
victims of sexual assault, including 8 that provided a focus on costs are based on responses from adults (16–59 years old) to
child abuse or child sexual abuse. Of the remaining 23 references, the Crime Survey for England and Wales. Adults complete a
3 were discarded on the basis that more recent updates were Self-Completion Module that asks questions about sexual abuse
available [i.e., Heeks et al. (2018) was retained whilst Brand and experienced during childhood. The Self Completion Module is
Price (2005), Dubourg et al. (2005) and Home Office (2011) were not used in the 10–15 year old survey as it is too sensitive. As such,
discarded]. Table 1 provides a summary of the remaining 20 prevalence and economic figures do not specifically take into
references. These references draw upon a range of economic and account offenses against children that have taken place 2015/16.
survey methods to establish sensible (and often conservative) cost There are limitations in applying adults cost to children though,
metrics. Three broad costing methods are employed. Incident arguably, it is a conservative strategy because children are likely
methods attempt to identify the cost per incident and can range to experience negative effects of child sexual abuse throughout
in scope from single costs (e.g., cost of medical or mental health their lifetime.
treatment) to total costs incurred. Prevalence methods draw out To provide a cost per incident, the figures from Heeks et al.
incident costs but examine the number of cases reported within (2018) were extracted and retained for subsequent cost estimates.
a year period to establish the total cost of disease incurred that Heeks et al. (2018) calculate the unit cost of rape as £39,360
year. Incidence methods calculate the lifetime costs of cases and other sexual offenses as £6,520 in 2015/16. These figures
first diagnosed in a particular year. The references in Table 1 increase to £43,214 and £7,158 in 2019 using Bank of England
vary in scope, but provide rich information on the range of Inflation Calculator (Bank of England Inflation Calculator)
tangible and intangible costs attributable to sexual harm. Tangible online “inflation” calculator with inflation averaged at 3.2% a
costs are often more readily quantifiable; they include those year. We do not know the exact extent to which child sexual abuse
costs that arise immediately or proximally to the offense, such involves acts of penetration but we can estimate this from the
as medical care, mental health services, criminal justice costs, most recent figures provided by the Office of National Statistics
victim/children’s services and loss of economic productivity. (2020)1 . Around one quarter of child sexual offenses recorded by
Intangible costs are the hardest to quantify; constituting the the police March 2018–2019 were rape offenses (ONS, 2020)1 .
human and emotional cost to victims. A range of measures Whilst one might argue this risks conflating figures as rapes
are employed (quality of life years adjusted, jury compensation might be more likely reported to the police than sexual assaults,
awards and willingness to pay) to try capture meaningful metrics responses to the anonymous CSEW Self Completion Module
on the reduced quality of life resulting from victimization. In reported by ONS (2020)1 found that roughly one third of those
Table 1 half of the references are solely focused on quantifying reporting any contact sexual offense before the age of 16 had
tangible costs, with seven incorporating intangible costs and
three focused specifically on the issue of intangible costs and how 1 Availableonline at: https://www.ons.gov.uk/peoplepopulationandcommunity/
they can be best be measured. For this research it was deemed populationandmigration/populationestimates (accessed September 1, 2020).

TABLE 1 | Unit costs of sexual abuse and child sexual abuse (published 2000–2020).
Criminal justice
Victim services
Intangible
Security
Health
Labor
References Country Offense Costing Unit cost
typea methodb
Cohen et al. (2004) USA SV WTP ✓ US resident willing to pay $126 per year for 10%
reduction in rape and sexual assault
Dolan et al. (2005) UK SV P ✓ ✓ Monetary value of discounted QALY losses £16,480
rape and £4,790 sexual assault. Intangible victim costs
£1,027 rape and £341 sexual assault
Fang et al. (2012) USA CA I ✓ ✓ ✓ ✓ Lifetime cost for non-fatal child maltreatment $210,012
Gelles and Perlman (2012) USA CA P ✓ ✓ ✓ ✓ National prevalence provided
Heeks et al. (2018) UK SV P ✓ ✓ ✓ ✓ ✓ ✓ Rape £39,360; and other sexual offenses £6,520
Hunt et al. (2017) USA SV In ✓ Average cost to taxpayers for legal services per
rape/sexual assault is $2000-$5000
Letourneau et al. (2018) USA CSA I ✓ ✓ ✓ ✓ ✓ ✓ female victim of non-fatal CSA $282,734; male victim
of non-fatal CSA
McCollister et al. (2010) USA SV In ✓ ✓ ✓ ✓ Combined tangible and intangible cost per rape/sexual
assault $240,776
New and Berliner (2000) USA CSA In ✓ Average amount billed for mental health treatment was
$2634.09
Olavarria-Gambi (2007) Chile SV P ✓ $1,368.8 medical assistance to victims of rape and
sexual assault
Peterson et al. (2017) USA SV I ✓ ✓ ✓ Lifetime economic burden of rape per adult female
$122,461
Peterson et al. (2018) USA SV I ✓ $730 average cost of lost productivity across victims of
intimate partner violence, sexual violence and stalking
Post et al. (2002) USA SV P ✓ ✓ ✓ ✓ ✓ Cost of rape and sexual assault $94,466 1996 USD
Saied-Tessier (2014) UK CSA P ✓ ✓ ✓ ✓ ✓ National prevalence provided
Shanahan and Donato (2001) Australia CSA In ✓ ✓ ✓ ✓ Tangible and intangible victim costs between
$119,340-$218,790
Smith et al. (2014) Australia SV P ✓ ✓ Incident costs for sexual assault $3,912
Tennessee et al. (2017) USA SV In ✓ $6,737 hospital billing for victims of sexual assault
Walby (2004) UK DV P ✓ Rape and assault by penetration £104,300; and
non-penetrative sexual assault £240
Wang and Holton (2007) USA CA P ✓ ✓ ✓ ✓ National prevalence provided
Yang et al. (2014) USA CSA P ✓ ✓ ✓ ✓ ✓ Estimated costs averaged $159,610 per sexual
violence incident for children, $110,937 rape of adult
and $389 sexual assault of adult
a CA, Child abuse; CSA, Child Sexual Abuse; DV, Domestic Violence; SV, Sexual Violence.
b I, Incidence; In, Incident; P, Prevalence; WTP, Willingness to Pay.
experienced rape and assault by penetration (including attempts). Similar to Saied-Tessier (2014), they calculate loss to quality of
As such, in calculating incident cost estimates we conservatively life separately from estimated total burden because of mixed
estimate 25% offenses at £43,214 and 75% at £7,158. opinion about the validity of converting QALY to a monetary
Letourneau et al. (2018) provide average lifetime costs for amount (e.g., see Dolan et al., 2005). Separately, they estimate
victims of non-fatal child sexual abuse in the USA in 2015. quality of life losses as $41,001 for female victims and $38,904
Drawing upon a wider range of measures than Heeks et al. for male victims.
(2018) such as education costs, they also draw on the available Converting these figures to the UK context required a number
child sexual abuse literature to establish life time effects on of decisions to be made. First, Letourneau et al. (2018) do not
health, criminal offending, suicide death and QALY losses. They differentiate different forms of child sexual abuse and so it is not
estimate that the lifetime cost for female victims of non-fatal possible to estimate costs based on rape or sexual assault (as in
child sexual abuse is $282,734. Estimates for male victims are Heeks et al., 2018). Set against this, Letourneau et al. (2018) draw
“artificially low” (pg. 419) at $74,691 due to a lack of reliable on an established literature that examines the impact of child
data pertaining to losses in male victim work productivity. sexual abuse in general. As such, we were confident that their

robust and replicable incidence based methods would not lead population is ∼3–5%. Seto (2008) agrees these as upper limits
to inflated life time estimates. The literature on harms emerging and has suggested more conservative estimates in recent years (1–
from different forms of abuse is arguably too under developed 3%; see Seto, 2018). The epidemiological problem stems from the
to meaningfully change their estimates. Second, Letourneau inability to fully explore the diagnosis of pedophilia or hebephilia
et al. (2018) establish total burden based on 75% of victims using survey methods. Studies, such as those outlined in Table 2,
being female. This is in spite of absent male work productivity include self-report measures of sexual fantasies, arousal, viewing
data significantly impairing male estimates (apart from a small child-adult sex and to a lesser extent behavior but they do not
difference in suicide death costs all other costs are equal). In this typically explore persistence, recurrence nor intensity of such
paper, we decided that differences between males and females thoughts, feelings and behaviors (Seto, 2018). As such, prevalence
in Letourneau were methodological artifacts and so would not is likely to be overestimated. That said, one could also argue there
be applied. As such, we would apply the value provided for are issues with self-disclosure and that these surveys identify
female victims. Third, their values are provided in 2015 USD. only those individuals willing to admit paraphilic thoughts.
Examining exchange rates provided by HM Revenue Customs An alternative view expressed in the literature is that surveys
Exchange Rates (2011) the average value of $1 was £0.6255 such as these might attract individuals who are more sexually
(average for the year to 31st March, 2015). As such, $282,734 was curious, open-minded or deviant (Ahlers et al., 2011; Santtila
calculated as £176,850 in 2015, this figure increases to £197,535 et al., 2015). Whilst recognizing these limitations, the prevalence
in 2019 using Bank of England online “inflation” calculator with rates described in Table 2 were considered for inclusion in
inflation averaged at 2.8% a year. This would be used to estimate economic analysis.
lower bound lifetime costs. Upper bound estimates would include Three studies were considered for inclusion based on (a) their
quality of life losses adjusted to $40,477 to account for 75% of focus on a wider range of ages, (b) large sample size and their
victims being female (as these figures were not affected by paucity attempts to identify a representative sample and/or (c) attempt
of male productivity data). This was calculated as £25,318 in to differentiate prevalence of pedophilia and hebephilia. In their
2015 and increased to £28,279 in 2019 using Bank of England stratified random online sample of the Czech population, Bartova
online “inflation” calculator with inflation averaged at 2.8% a et al. (2020) report a current sexual preference for pre-pubescent
year. As such, an upper bound estimate of lifetime cost would children amongst 0.6% of males (aged 18–88 years) and for
be £225,814. pubescent children amongst 3.1% of males. This is similar to
Santtila et al. (2015) who in their attempt to identify a 1 year
incidence of pedophilia in their analysis of 1,310 Finnish twins
Estimating Prevalence of Pedophilia and report a prevalence of 0.2% for pedophilia (children aged under
Contact Offending 12 years) and 3.1% for hebephilia (13–15 years). The age range of
To support review question 1, a systematic search of the literature their study is more limited, focused on males aged 33–43 years.
was undertaken to identify a) prevalence of pedophilia (interest In their online survey of 8,718 German men, Dombert et al.
in pre-pubescent children, generally aged 13 years and younger) (2016) explore lifetime sexual interest in pre-pubescent (children
and hebephilia (interest in pubescent children) in the general aged under 12 years) only, reporting 0.1% exclusive pedophilic
population, and b) rates of contact offending amongst pedophilic sexual preference and an upper estimate of 5.4% non-exclusive
and hebephilic individuals. pedophilic sexual interest. Though the findings from these three
In terms of prevalence, material would be included if studies are broadly similar, we decided to opt for Bartova et al.
it provided population-based prevalence rates that could be (2020) for a number of reasons. First, both Santtila et al. (2015)
extracted from published accounts. Clinical and forensic samples and Dombert et al. (2016) define paedophila as a sexual interest
were not retained. Community and convenience based samples in children under 12 years. This makes sense given many children
were retained for further inspection. Since Seto (2008) provides biologically mature earlier than 13 years, however, this does not
an outline of prevalence studies published 2000–2013 the first match the cut off criterion for prepubescence outlined in DSM
author made use of these resources (of which five studies V nor does it align with UK legislation (the Sexual Offenses Act
potentially met the study criteria) and searched for journal 2003 uses distinct differences between sexual assaults on children,
articles published post 2013. She applied Boolean operators to section 5–8 specifically distinguishes “rape and other offenses
search terms to generate search phrases (“prevalence” AND against children under 13 years”). Like Dombert et al. (2016)
“paedophil∗” OR “pedophil∗” OR “hebephil”), resulting in an and Bartova et al. (2020) explore a wider range of ages than
additional 219 items. These items were then screened for Santtila et al. (2015) but compared to the former Bartova et al.
filtering as described above. Seventeen items were downloaded (2020) provide separate estimates for pedophilia and hebephilia.
for further consideration. Only 8 of these items provided a As such, we proceed with the estimate that 0.6% of the UK
population based estimate of pedophilia or hebephilia with male population exhibit pedophilic sexual preferences and 3.1%
Santtila et al. (2010, 2015) drawing upon the same data set. exhibit a hebephilic sexual preference.
Three additional post 2,000 references were identified from Seto It is difficult to estimate the number of pedophilic and
(2018). Table 2 provides a summary of 11 references that met the hebephilic individuals that pose a risk of contact offending as
inclusion criteria. this is not always explicitly considered in research. Bartova et al.
American Psychiatric Association (2013) estimate that the (2020), for example, omitted to ask about contact offending
highest possible prevalence for pedophilic disorder in the male because of its criminal implications, preferring to ask whether

TABLE 2 | Outline of prevalence studies (published 2000–2020).
References Country Method Sample Prevalence
Abdullahi et al. (2015) Nigeria Survey Student: 871 randomly selected sample from 0.2% males reported pedophilia (symptoms lasted up
one university (n = 447 males, mainly 18–25 to 6 months/causes distress)
years)
Ahlers et al. (2011) Germany Survey Community: 367 males in a metropolitan city 9.5% had sexual fantasies about pre-pubescent child;
(from a representative sample of 6,000 men 6.0% had masturbation fantasies; 3.8% had
aged 40–79 years) experienced sexual contact with a pre-pubescent
child
Bartova et al. (2020) Czech Republic Survey Community: 10,044 stratified random online 0.6% disclosed a sexual preference in pre-pubescent
sample (n = 5,023 males, aged 18–88 years) children (causing 3.3% to confide in a health care
professional; 3.1% disclosed a sexual preference in
pubescent children (causing 11.5% to confide in a
health care professional; 0.9% sexually aroused by
pre-pubescent children (4.4% pubescent); 0.6%
disclosed “porn use” with respect to pre-pubescent
children (1.7% pubescent); 0.4% experienced sexual
fantasies about pre-pubescent children once a
week/every day (1.4% pubescent); 1% would definitely
have sex with a pre-pubescent child if it was legal
(4.4% pubescent)
Dawson et al. (2014) Canada Survey Student and online; 1,015 non-representative 0.6% report sex with a pre-pubescent child as sexually
sample (n = 305 males, average age 23 years) arousing; 0.9% rated sex with a pubescent child as
sexually arousing
Dombert et al. (2016) Germany Survey Community; representative sample of 8718 Estimate upper limit of 5.4% non-exclusive paedophilic
males aged 18–89 years disorder, with 4.1% reporting sexual fantasies involving
pre-pubescent children and 0.1% evidencing exclusive
pedophilia; 3.2% indicated sexual behavior involving
pre-pubescent children, with 1.7% indecent image
offenders, 0.8% reporting sexual contact with children
and 0.7% mixed offenders (1.5% combined contact
offending); 0.4% had paid a child for sexual services,
similarly 0.4% had intended child sex tourism
Joyal (2014) Canada Survey Online; non-representative sample of 1,516 (n = 1.8% ever fantasized about pre-pubescent children
717 males)
Santtila et al. (2010, Finland Survey Twin study: 1,310 males aged 33–43 years 0.2% reported sexual interest in children aged 12 or
2015) younger; 3.1% for children aged 13-15 years;
combined 1 year incidence estimated at 3.3%
Seto et al. (2015) Sweden Survey Schools; representative sample of 1,978 males 4.2% reported use of indecent images of children,
aged 17–20 years which was positive associated with self-reported
coercive sexual behavior involving pre-pubescent
children; 14% reported sexually coercive behavior
against a pre-pubescent child
Williams et al. (2009) Canada Survey Student; non-representative sample of 191 12% admitted any sexual fantasy about children; 5%
males (average age 19.7 years) admitted any sexual behavior involving children
Wurtele et al. (2014) USA Survey Student and online; 435 non-representative Nearly 10% reported some likelihood of having sex with
sample (n = 173 males, aged 18 years and children or viewing child pornography if they would not
older) be caught or punished; 4% reported some agreement
with each of following items, fantasize about having
sex with child, being sexually attracted to little children,
masturbation fantasies about sex with children
Bold values indicate prevalence of contact offending provided in reviewed papers.
respondents would engage in such behavior if it were legal. et al. (2009) and so were considered for inclusion. In line with
Four studies in Table 2 had attempted to address the issue of the purposes of the current paper, Ahlers et al. (2011) attempt
contact offending in their prevalence studies. Sexual contacts to extrapolate self-reported rates of contact offending to the
with children range from 1.5 to 14% with the larger prevalence general population. They apply observed prevalence (14/376)
figure exploring a broader range of sexually coercive behaviors but use the number of individuals invited to participate in the
amongst young adults (Seto et al., 2015). Ahlers et al. (2011) research as the denominator (14/6,000) resulting in a much more
and Dombert et al. (2016) provide representative samples from conservative contact offending prevalence rate of 0.23% (rather
a wider range of ages than Seto et al. (2015) and Williams than 3.8% of the sample). This translates to 1,800 and 69,000

males (aged 40–79 years) who have acted out their sexual interest RESULTS
on pre-pubescent children in the City of Berlin and Germany,
respectively. They do acknowledge that the absolute number Review Question 1: What Is the Potential
is likely to be considerably higher. Their figures are based on Scale of the Problem in the UK, in Terms of
self-disclosure amongst men aged 40–79 years, whereas they the Number of and Socio-Economic
acknowledge that 20-40 year olds are both sexually and criminally Burden of Pedophilic and Hebephilic
more active. Further, only non-exclusive heterosexual pedophilic
types appeared to participate in Ahlers et al. (2011) and so it
Individuals Who Pose a Risk of Contact
is unclear how well this prevalence would translate to exclusive Offending Against a Child?
nor homosexual pedophilic types. Dombert et al.’s (2016) sample Applying Bartova et al. (2020) prevalence rates of 0.6% for
includes those with an interest in males and females, they have pedophilia and 3.1% for hebephilia to mid-year population
also used a particularly in-depth methodology more closely estimates 2019 (www.ons.gov.uk) of adult males (16–89 years) in
aligned with DSM V. As such, we decided to use Ahlers et al.’s the UK (N = 26,275,165), there could be as many as 157,651
(2011) 3.8% as an upper bound estimate and Dombert et al.’s males with pedophilic sexual preferences and 814,530 males
(2016) 1.5% prevalence as a lower bound estimate of contact with hebephilic sexual preferences. To get a sense of those that
offending. Note, neither of these references provide estimates of pose a risk of contact offending we apply Ahlers et al.’s (2011)
contact offending for hebephilic individuals. Further, we elected prevalence of 3.8% as an upper bound estimate and Dombert
to use self-reported prevalence from the sample rather than the et al.’s (2016) prevalence of 1.5% as an lower bound estimate of
population-based sample approach taken by Ahlers et al. (2011). contact offending. The estimated number of contact offenders
There are many reasons why the remaining 5,624 individuals did multiplied by incident and lifetime victim costs are provided in
not participate in their study; no doubt ranging from repulsion to Table 3.
undisclosed contact offenses. We estimate that there could be between 2,365 and 5,991
males with pedophilia who are likely contact offenders. Applying
Heeks et al.’s (2018) costs of £43,214 (25%) and £7,158 (75%)
Past and Future Offending of IIOC this could mean between £38 million and £97 million socio-
Offenders economic burden (based on a conservative estimate of one victim
Since Seto and colleagues have already conducted a detailed per offender). Applying life time costs this could increase to
meta-analysis on past and future offending of IIOC offenders £467 million (£534 million, QALY) to £1.2 billion (£1.35 billion,
(Seto et al., 2011) we did not conduct further systematic search QALY). Applying the same assumptions to population estimates
of the literature. In terms of past offending, Seto et al. (2011) of males with hebephilia, we estimate between 12,218 and 30,952
found approximately one in six internet sex offenders have a males are likely contact offenders, translating to between £198
history of contact sexual offenses. Prevalence rates reduce to million and £500 million incident costs and between £2.4 billion
one in eight (12%) when using official data and increase to (£2.8 billion, QALY) and £6.1 billion (£7 billion, QALY) life
one in two when using self-reports (55%). In line with Seto time costs.
et al. (2011) findings we adopt 12 and 55% as lower and upper
bound risk estimates for past contact sexual offending. In terms
of reoffending, recidivism studies (Seto et al., 2011; 1.5–6 year Review Question 2: What Is the
follow up) indicate that internet sexual offenders have a lower Socio-Economic Burden Potentially
rate of sexual recidivism (4.6%) than contact sexual offenders
Generated by the National Suspect Pool of
(e.g., 13.7%). Where information is available on type of sexual
recidivism in Seto et al. (2011) meta-analysis, 2% recidivate 50,000 IIOC Offenders?
with a contact sexual offense and 3.4% recidivate with a further Historical Victims Are Potentially Identified and
IIOC offense. These figures are likely to be underestimates as Safeguarded
information is based on official criminal records in all samples. In line with Seto et al. (2011) findings we adopt 12 and 55% as
On the balance of evidence, dual offenders demonstrably show lower and upper bound risk estimates for past contact sexual
slightly higher rates of both internet and contact sexual offenses offending. Economic estimates are outlined in Table 4.
than internet sexual offenders (Goller et al., 2010; Graf and
Dittmann, 2011; Wakeling et al., 2011). As Wakeling et al. (2011)
study provides an analysis of a large UK data set–a “routine” Lower Bound Estimate Based on Official Records
correctional sample (N = 1,344 UK IIOC offenders), we elected It would be expected that at least 6,000 individuals in the UK who
to use this study to help inform upper bound estimates for dual download and trade IIOC online will have an official history of
offenders. Wakeling et al. (2011) found that dual offenders had a contact sexual offending. Applying the average cost per victim
two year proven sexual reoffending rate of 6.6% whilst internet (incident, lifetime and lifetime plus QALY) to a conservative
specialists had a reoffending rate of 1.6%. In the economic estimate of one prior victim per offender, the potential UK
analysis that follows we draw on both Seto et al. (2011) and suspect pool could have already contributed toward a socio-
Wakeling et al. (2011) to provide lower and upper bound economic burden between £97 million (incidence costs) and £1.2
reoffending estimates. billion-£1.35 billion (QALY)(lifetime costs).

TABLE 3 | Estimated socio-economic burden attributable to paedophilic and hebephilic contact sexual offending (one victim per offender).
Population Contact offending Contact offenders (n) Incident costsa Lifetime costsb Lifetime costs
prevalence (with QALY)c
Males with pedophilia Lower bound: 1.5% 2,365 £38,237,766 £467,170,275 £534,050,110
(n = 157,651) (Dombert et al., 2016)
Upper bound: 3.8% (Ahlers 5,991 £96,895,466 £1,183,432,185 £1,352,851,674
et al., 2011)
Males with hebephilia Lower bound: 1.5% 12,218 £197,589,496 £2,413,482,630 £2,758,995,452
(n = 814,530) (Dombert et al., 2016)
Upper bound: 3.8% (Ahlers 30,952 £500,555,744 £6,114,103,320 £6,989,394,928
et al., 2011)
a Estimate25% offenses at £43,214 (penetrative) and 75% at £7,158 (non-penetrative) in 2019 GBP.
b Estimated at £197,535 in 2019 GBP.
c Estimated at £225,814 in 2019 GBP.
TABLE 4 | Estimated socio-economic burden attributable to historical contact offending of IIOC suspects (estimated at one victim per offender).
IIOC suspects (N Offending prevalence Contact offenders (n) Incident costsa Lifetime costsb Lifetime costs
= 50,000) (with QALY)c
Past contact Lower bound: 12% 6000 £97,032,000 £1,185,210,000 £1,354,884,000

offending (Seto et al., 2011)
Upper bound: 55% 27,500 £444,730,000 £5,432,212,500 £6,209,885,000
(Seto et al., 2011)
Upper Bound Estimate Based on Self-Reports Upper Bound Estimates Based on Detection of
It might be expected that as many as 27,500 would disclose Dual Offenders
a history of contact sexual offending (including undetected If we assume more dual offenders amongst the suspect pool
cases). Again, applying the cost per victim of sexual assault to then we can apply Wakeling et al. (2011) prevalence rates.
a conservative estimate of one prior victim per offender, the Conservative application of their consolidated prevalence rate
potential UK suspect pool could have already contributed an (6.6% across all risk levels) to the national suspect pool suggests
economic burden between £445 million (incident costs) and £5.4 that 3,300 new offenses could occur within two years. This
billion–£6.2 billion (QALY) (lifetime costs). would be estimated to include 2,300 (4.6%) offenders committing
new internet offenses and 1,150 (2.3%) offenders committing
Potential Future Victims May Be Safeguarded new contact offenses. Under this model, there are potentially
Estimates associated with future offending are provided 1,150 future victims of contact sexual offenses which with a
in Table 5. conservative estimate of 1 victim per offender could contribute a
further socio-economic burden between £18.6 million (incident
costs) and £227 million–£260 million (QALY) (lifetime costs).
Lower Bound Estimate Based on Official Records
Given the conservative recidivism rates identified by Seto et al.
(2011) it would be anticipated that 4.6% of the 50,000 suspect DISCUSSION
pool would go on to commit a new sexual offense of some
kind within 1.5–6 years (n = 2,300). Specifically, 2% with a This article has presented two ways in which the socio-economic
contact sexual offense (n = 1,000) and 3.4% with a further IIOC burden that is potentially contributed by child sex offenders
offense (n = 1,700). Under this model, there are potentially can be estimated. The aim of this work was to demonstrate
1,000 future victims of contact sexual offenses. Heeks et al. the potential scale of the problem within the UK and to
(2018) annual costs can be predicted over the length of time provide a clear argument for evidence based policing strategies
included in follow up studies, as such, incident costs are estimated to help identify the riskiest offenders, in terms of their risk of
over 1.5–6 years. Combined with lifetime and lifetime (with contact offending.
QALY) future contact offending against 1 victim per offender In our first analysis, we consider the amount of pedophilic and
could contribute a further socio-economic burden between hebephiliac males in the UK who are likely contact offenders. At
£16 million (incident costs) and £198 million-£226 million one victim per offender, the combined socio-economic burden
(QALY) (incidence costs). from males with pedophilic and hebephiliac sexual preferences

TABLE 5 | Estimated socio-economic burden attributable to future offending of IIOC suspects (estimated at one victim per offender).
IIOC suspects (N Offending prevalence Recidivists (n) Incident costsa Lifetime costsb Lifetime costs
= 50000) (with QALY)c
Future contact Lower bound: 2% 1,000 £16,172,000 £197,535,000 £225,814,000

Upper bound: 2.3% 1,150 £18,597,800 £227,165,250 £259,686,100
(Wakeling et al., 2011)
Future internet Lower bound: 3.4% 1,700
Upper bound: 4.6% 2,300
(Wakeling et al., 2011)
could amount to somewhere in the region of £236-£597 million future contact offenses amongst the suspect pool of 50,000 IIOC
incident costs and £2.9-£7.3 billion lifetime costs (increased offenders we estimate that there could be between 1,000 and 1,150
to £3.3-£8.3 billion including QALY). These figures are likely future contact victims. These offenses could contribute a further
underestimates given the average number of children abused by socio-economic burden of between £16–£18.6 million incident
pedophiles [estimates ranging from 1.7 for homosexual incest costs and between £198–£227 million lifetime costs (increasing
pedophiles to 27 for bisexual pedophiles; Abel and Harlow to £226–£260 million including QALY). It should be stressed that
(2001)]. The results of our first review question suggests that these estimates are based on what would be expected based on
targeted police action toward those offenders with pedophilic official records. It is likely that there could be further undetected
and hebephiliac themes in their online behavior is a sound internet and contact sexual offenses, but these estimates are
investigative strategy. However, further work is needed to not available.
understand the differences between those who admit to using The results of our review have supported the argument
indecent images and those that admit to contact sexual offenses. that targeted police action toward those offenders with past
In this review, rates of indecent image use were low (Table 2) sex offending histories or those who share characteristics with
but contact offenses were even lower. This concurs with broader known dual offenders makes sound moral and fiscal sense as
meta-analysis of Seto et al. (2011) and Babchishin et al. (2014). these offenders are likely to cause more harm to victims in
As the work by Babchishin et al. (2014) demonstrates, IIOC real terms. There is also the potential that historical victims
only offenders have a higher rate of pedophilia than contact are identified and potentially safeguarded through targeted
sexual offenders against children. Dual offenders had higher police action, thereby contributing further to victim safety and
rates of pedophilia than IIOC only offenders and also pose a harm reduction goals. Whilst there is a mutual consensus that
greater risk of contact sexual offenses. As such, investigative action must be taken, academics and practitioners are also
strategies that focus on both evidence of pedophilia, hebephilia beginning to appreciate that indiscriminant arrests would not
and concomitant contact offending would appear to be the be effective in reducing the scope of the IIOC problem [beyond
optimal approach. the general deterrence purpose that they would serve, e.g.,
In our second analysis, we examined the national pool of Wolak et al. (2014)]. Viewing all internet IIOC offenders as high
50,000 IIOC offenders in the UK (CEOP, 2013) estimating that risk for contact offending or recidivism is unworkable because
between 6,000 and 27,500 dual offenders could have already the population is simply too diverse. There are also ethical
contributed to an economic burden of between £97–£445 million implications with an indiscriminant or random arrest strategy
incident costs and £1.2–£5.4 billion lifetime costs (increased in potentially tackling less harmful offending whilst not tackling
to £1.4–£6.2 billion including QALY). Given the problems in more harmful offending. Further, if a random strategy was made
defining prevalence of past contact sexual offenses it is likely public via the media and resulted in few convictions, because it
that the estimates are conservative and that there have been a did not pick out the most serious offenders or those who posed
larger number of victims for the highest risk cases. We know the greatest risk for recidivism, it could seriously undermine
very little about how many victims are harmed by detected IIOC public confidence in investigative competence in this area. Our
offenders, and virtually nothing about the victims of undetected review has suggested that prioritization based on aspects of
IIOC offenders. Bourke and Hernandez (2009) found an average behavior that are most correlated with pedophilia, hebephilia and
of 14.7 past victims for each IIOC offender entering cognitive dual offending is an effective strategy for identifying offenders
behavior therapy in prison. Although self-report studies such who pose the greatest risk of harm to children. Indeed, the
as this run the risk of conflating undetected victim prevalence application of Wakeling et al. (2011) figures are persuasive here.
and have been criticized (Seto, 2013) the findings help to If, by nature of prioritizing dual offenders, law enforcement
provide reasonable assurances that one victim per offender (for agencies detect more dual than internet only offenders more
both review questions) is a conservative estimate. In terms of victims can be safeguarded. Here, even a modest estimate of 150

additional victims saves millions in socio-economic terms along limited systematic reviews and meta-analyses available. Some
with the untold toll on victims’ lives. of the available evidence is based on small, non-representative,
In terms of practical implications for law enforcement, convenience samples, age restricted samples, broad range of
our review demonstrates the importance of evidence-led risk measures and definitions. We have attempted to identify those
prioritization tools. Currently, only one such tool has been studies that have large population based national samples of
developed and validated in investigative contexts. KIRAT (Long community males from a variety of ages and which utilize
et al., 2016) prioritizes individuals suspected of possessing, some discriminate operationalization of sexual interest. These
making, taking and/or distributing IIOC based on their risk of could be improved upon and we recognize that our choices
also committing contact offenses. KIRAT focuses on previous may impact the validity of our findings. Specifically, Bartova
criminal history, access to children, current online and offline et al. (2020) ask whether respondents “have such a preference”
behavior, and other relevant factors, and shows a 95% accuracy for intimate contact with pubescent or pre-pubescent children.
rate for high-risk offenders and a 20% false positive rate for lower This does not help to establish how exclusive that preference
risk offenders. Prior to KIRAT, the relative significance of risk was. Bartova et al. (2020) approach falls foul of the types of
variables were unknown – and over 150 variables identified by criticisms levied at surveys by Seto (2013). Dombert et al.
the research team that many officers might think discriminate (2016) methodology takes us closer to understanding exclusive
between contact and IIOC offenders have now been discarded. and non-exclusive pedophilic types but they did not include
It is the national standard in the United Kingdom, having been hebephilic sexual interests and so were not utilized in this study.
legally mandated for use in all 43 UK police forces by the National A second limitation is the application of pedophilic individuals’
Police Chiefs Council. KIRAT has many benefits for police force; contact offending prevalence to males with hebephilia. Further
it is simple to use, does not require clinical training, it is objective population-based studies aligning more closely to DSM-V
and immune to offenders’ forensic awareness since it is based criteria (including contact behavior and mirrored content for
largely on criminal history rather than indecent image preference hebephilia) would be clearly beneficial here. One final limitation
or search terms. Further work is needed to develop KIRAT and to is the application of pedophilia and hebephilia prevalence to the
help embed and sustain use within police forces internationally. male population aged 16–17 years of age. The studies cited in
A further promising tool, CPORT has been developed by Michael Table 2 had surveyed males over 18 years of age. We made the
Seto and colleagues (Seto and Eke, 2015; Eke et al., 2019) however decision to estimate national prevalence from 16 years as males
this recidivism tool would benefit from further validation in of this age can be diagnosed with pedophilia and pedophilic
investigative contexts. At present, its approach is predominantly disorder. And yet, we do not know the extent to which prevalence
risk assessment (who will go on to reoffend) rather than risk and rates of contact offending apply to this younger group. Set
prioritization (who is likely to have already committed contact against these limitations, the rates of pedophilia and hebephilia
offenses). Our review also demonstrates the importance of reported here are within the range of those reported by other
technology driven strategies that prioritize offenders evidencing researchers (e.g., Seto, 2013) and so we are confident that
pedophilia, hebephilia along with concomitant contact offending. we have not radically overestimated prevalence of pedophilia
Whilst academics are working with police forces to identify and hebephilia. Further work is needed on contact offending
suitable tools further academic research is needed to understand prevalence and, in the absence of UK studies, it would be helpful
the links between online behavior and risk of contact offending. to establish the external validity of international studies.
Existing research demonstrates language has limited potential There are a number of limitations with our cost estimates.
to differentiate different types of offenders (e.g., Broome et al., First, our estimates assume that victims disclose their abuse
2018) and the evidence base around indecent image content and and so use the services making up the bulk of tangible costs.
collections has proven inconclusive (e.g., Long et al., 2016). However, as few as 14% of sexual violence victims report
In addition to practical implications, this work contributes offenses to the police (Daly and Boughours, 2010). Further work
to the Evidence Based Policing literature on cost effectiveness. is needed to understand economic costs attributable to non-
Economic Evaluation is one of the five key aspects of reported offenses. Certainly, the innovative approaches being
policing intervention evaluation [as outlined in the College of developed to capture intangible costs would be insightful here,
Policing’s EMMIE model; Thornton et al. (2019)] and academic- as these costs would apply whether a victim disclosed or not.
practitioner partnerships are working in multi-disciplinary Second, Heeks et al. (2018) do not specifically identify costs of
settings to explore a range of ways by which the cost effectiveness sexual violence with child victims. As we have noted above, we
of police action can be assessed [Crime Harm Index, Sherman anticipate this has led to cost underestimation but further work
et al. (2016); QALY, Heaton and Tong (2016)]. Our paper is examining child related costs would be clearly beneficial. Third,
the first of its kind to provide an economic framework that we assume an equivalence between prevalence based annual costs
could be applied to establish the cost effectiveness of a range and cost per incident. This is not inconsistent with the economics
of evidence-led strategies to tackle online child sexual abuse. As literature but we should be categorically clear that our incident
such, we anticipate that this paper will be of interest to a range of estimates do not purport to be annual estimates. What we take
academics and practitioners across multiple disciplines. from this prevalence based approach is the cost that could be
There are a number of limitations with our approach in anticipated from each sexual offense and apply this as a cost
the current study. First, we rely on single studies as the basis per incident, irrespective of time frame. Fourth, the evidence
of estimates rather than meta-analyses. However, there are underlying assumptions in Letourneau et al. (2018) is essentially

US centric. Further and more detailed work would be needed highlights the additional impact that IIOC distribution has on its
to interrogate the equivalent literature base in the UK, drawing victims. Many victims report that additional distress is caused by
upon databases routinely used within Health Economics as well knowing that images of their abuse are circulated and viewed for
as Psychology. Fifth, whilst Letourneau et al. (2018) approach is sexual purposes (von Weiler et al., 2010) leading some authors
particularly in-depth some measures are under developed. For to conclude that this constitutes revictimization in real terms
example, educational impacts are measured using the typical as victims experience exacerbated PTSD symptoms including
costs of special education and this does not do justice to the flashbacks and panic attacks. The extent to which IIOC internet
profound educational impacts following sexual violence (e.g., offenders are made culpable for prolonged victim distress, by
Bolger, 2016). Further, Letourneau et al. (2018) estimates are downloading and sharing IIOC, is being hotly debated in the
based on victimization at age 11 years. This is a good starting US Supreme Court. Certainly, we might expect in time that the
point given the age preferences of males with pedophilia and human cost of being a victim during the commission of a sexual
hebephilia but further work is needed to explore variation in offense for IIOC purposes is more realistically calculated. With
victim costs as a function of victim age. Finally, we apply costs to further work, such as that started by Saied-Tessier (2014) and
each estimated past or future contact offense and this is despite through restitution awards we might gain further clarity about
there being a wide range of contact offenses amongst males with how to calculate victim costs for IIOC offenses.
pedophilia, hebephilia and dual offenders. Cohen and Galynker There is a more general point to be made here about
(2002), for example, found that pedophilic individuals are 2.5 wider community costs. Researchers and legal academics argue
times more likely to engage in physical contact with a child that, in very simple terms, social harms arise from the sexual
than voyeuristic or exhibitionistic activities. However, in a review objectification of children for adults’ sexual pleasure. Leary
article, Hall and Hall (2009) report that pedophilic individuals (2007) argues that the vicarious harm emerging from IIOC
typically engage in fondling and genital manipulation more than images manifests in many ways; offenders may use images for
intercourse. Exceptions to this include cases of incest, hebephilia sexual gratification, to groom children to be sexually molested,
and when children are physically coerced (Murray, 2000; Cohen to support the idea that adult-child sexually abusive relationships
and Galynker, 2002). We have tried to address this, to some are acceptable, to decrease the inhibitions of potential victims, to
extent, in the application of Heeks et al. (2018) by estimating demonstrate to victims how to sexually please a sexual offender,
only 25% of cases involve penetrative acts of rape. However, we to control victims, to barter/exchange on the internet, and for
acknowledge further scrutiny would be beneficial here. profit (pg. 13). In short, IIOC images create additional harm and
There are a number of ways in which the present study could costs to children not in IIOC images and to society in general.
be developed. Further work exploring victimization would be Whilst it is difficult to attach a socio-economic metric to this
beneficial. For example, Abel et al. (1987) found that pedophilic it does point toward the vicarious harms that the proliferation
individuals with male victims had a higher number of victims. of IIOC online may potentially cause, further justifying evidence
A more realistic assessment of victim numbers would drastically based and targeted policing strategies in this area.
alter the estimates provided here. Further, conspicuously absent Whilst the cost estimates we present here are alarming our
from this review is the population of contact sexual offenders intention is not to provoke moral panic, particularly in the
who are neither exclusively pedophilic nor hebephilic and yet discussion around persons with pedophilia and hebephilia.
pose a considerable risk of harm to children. In Seto (2009) Our cost estimates are based on the assumption that the
study only 50% of child sexual abuse offenders expressed a overwhelming majority of individuals with these sexual
sexual preference for children. This population contributes preferences (96.2–98.5%) do not contact offend. Rather, our
significant socio-economic burden and yet it was not deemed intention is to outline the very real risk that police offices and
feasible to gain prevalence and rates of contact offending from other public sector agencies work with every day, and which
the population-based approach taken in the present study. impact on the safety and safeguarding of children. As such, a
Investigating authorities are also beginning to appreciate the key issue, in our view, are evidence-based methods to assist in
considerable impact of “high harm” offenders; those online the triaging of this otherwise overwhelming task. The estimates
offenders who have little intention of meeting children in real life that we provide should be seen as a “starting point” of a much
but coerce children to engage in sexual acts through blackmail wider discussion about the benefits of police action in this
or coercion. Evidence is beginning to emerge on the impacts area and how these benefits may be measured. We argue that
of online child sexual abuse and how it can engender levels of evidence-led decision criteria can start to facilitate investigative
sexual harm experienced by online victims comparable to that goals by helping to prioritize high risk offenders and in doing
experienced by offline victims of child sexual offenses (Hanson, so, contribute most successfully to harm reduction. It is hoped
2017; Hamilton-Giachritsis et al., 2017; Jonsson et al., 2019). that by outlining potential impact in socio-economic terms that
Any measure of national socio-economic burden needs to take funders and policy makers would continue to recognize the fiscal
account of this emerging group. Further development is also and moral value of funding the collaboration between academics
needed in establishing the vicarious costs associated with being and police forces as they attempt to make the most efficient use
a victim of an IIOC offense. In this study, we estimate that there of limited public sector resources.
could be between 1,700 and 2,300 future IIOC offenses but it is With more detailed and focused cost estimates (as outlined
not possible, as yet, to attach a meaningful cost to these offenses above) we can start to produce a sketch of how the economic
in terms of victim harms. Evidence is beginning to emerge that and social burden of targeted and random approaches can be

calculated. This could take the form of quasi experiments where 2008). A further concern is that online pedophilic individuals
police forces who take targeted approaches Vis “business as will continue to gain social cohesion, positive reinforcement
usual” or random approaches can be compared. Alternatively the and validation for their identities whilst offending online
implementation of targeted approaches can be staggered so as to (Brennan and Hammond, 2011). Finally, children may be able
provide a baseline of “business as usual” or random approaches to access IIOC more easily (e.g., Koontz, 2004; Dombroski
for police forces in the period before targeted approaches are et al., 2007) which raises concerns for indirect victimization or
rolled out (Bandyopadhyay, 2017). Sufficient time will be needed desensitization to IIOC related themes.
to establish legal outcomes, that is, whether targeted approaches Our aim has been to estimate the potential scale of the
did, (a) identify individuals with more historical contact offenses, problem in the UK to help understand the need for continued
and related to this then (b) safeguard more historical victims targeted action. As our analysis stands we suspect that we
and lower potential rates of victimization and offending, and may have underestimated rather than overestimated the scale
potentially (c) contribute to higher rates of conviction with (d) of the problem. Whilst we acknowledge limitations with the
longer sentences (as per crime harm index calculations). It may assumptions underlying our estimates these figures are a starting
be the case that targeted approaches help lower offending rates point for future discussion. We encourage feedback from others
because of deterrence and improved apprehension may deter about how we might develop our calculations.
more would be offenders with consequent benefits for potential
victims. It is necessary to establish how much of the outcomes can DATA AVAILABILITY STATEMENT
be attributed to targeted approaches as concurrent interventions
as well as socio-economic and technological change within the The original contributions presented in the study are included
same time frame can all impact the outcomes we are trying to in the article/supplementary material, further inquiries can be
measure. The cost of incorrect identification of offenders would directed to the corresponding author/s.
also need be taken into account.
It is clear, given the exponential rise of IIOC online, that AUTHOR CONTRIBUTIONS
action must be taken now to offset the potential long term
problems we might witness by failing to address the problem SG undertook literature searches and reviews, she conducted the
before it gets worse. For example, the proliferation of IIOC on economic analysis and prepared the paper for publication. LA
the internet could stretch police forces beyond their investigative identified the need for the research and worked with policing
capabilities and could lead to desensitization to IIOC images. The colleagues to help establish the review questions and contributed
widespread availability of IIOC could promote trivialization of to various drafts of the paper. Both authors contributed to the
content and thereby encourage further offending (e.g., Quayle, article and approved the submitted version.
REFERENCES Bank of England Inflation Calculator. Available online at: https://www.

bankofengland.co.uk/monetary-policy/inflation/inflation-calculator (accessed
Abdullahi, H., Jafojo, R. O., and Udofia, O. (2015). Paraphilia among September 8, 2020).
undergraduates in a Nigerian university. Sex. Addict. Compulsivity J. Treat. Bartova, K., Androvicova, R., Krejcova, L., Weiss, P., and Klapilova, K. (2020).
Prev. 22, 249–257. doi: 10.1080/10720162.2015.1057662 The prevalence of paraphilic interests in the Czech population: preference,
Abel, G. G., Becker, J., Mittelman, M., Cunningham-Rather, J., Rouleau, J. L., arousal, the use of pornography, fantasy, and behavior. J. Sex Res. 58, 86–96.
and Murphy, W. D. (1987). Self-reported sex crimes of nonincarcerated doi: 10.1080/00224499.2019.1707468
paraphiliacs. J. Interpers. Violence 2, 3–25. doi: 10.1177/088626087002001001 Bolger, D. (2016). Gender violence costs: schools’ financial obligations under title
Abel, G. G., and Harlow, N. (2001). The Abel and Harlow Child IX. Yale Law J. 125, 2106–2130.
Molestation Prevention Study. Excerpted From the Stop Child Bourke, M. L., and Hernandez, A. E. (2009). The “butner study” redux: a report of
Molestation Book. Philadelphia, PA: Xlibris. Available online the incidence of hands-on child victimization by child pornography offender. J.
at: www.childmolestationprevention.org/pdfs/study.pdf (accessed September Fam. Violence 24, 183–191. doi: 10.1007/s10896-008-9219-y
1, 2020). Brand, S., and Price, R. (2005). The Economic and Social Costs of Crime. Home
Ahlers, C. J., Schaefer, G. A., Mundt, I. A., Roll, S., Englert, H., Willich, S. N., Office Research Study 217. London: Home Office Research, Development and
et al. (2011). How unusual are the contents of paraphilias? Paraphilia-associated Statistics Directorate.
sexual arousal patterns in a community-based sample of men. J. Sex. Med. 8, Brennan, M. L., and Hammond, S. (2011). Complete critical literature review. iCOP:
1362–1370. doi: 10.1111/j.1743-6109.2009.01597.x identifying and catching originators in P2P networks (Technical report).
American Psychiatric Association (2013). Diagnostic and Statistical Manual of Brennan, M. L., and Hammond, S. (2017). A methodology for profiling paraphilic
Mental Disorders, 5th Edn. Arlington, VA: American Psychiatric Publishing. interest in child sexual exploitation material users on peer-to-peer networks. J.
doi: 10.1176/appi.books.9780890425596 Sex. Aggression 1:90. doi: 10.1080/13552600.2016.1241308
Babchishin, K. M., Hanson, R. K., and VanZuylen, H. (2014). Online child Broome, L. J., Izura, C., and Lorenzo-Dus, N. (2018). A systematic review
pornography offenders are different: a meta-analysis of the characteristics of of fantasy driven vs. contact driven internet-initiated sexual offenses:
online and offline sex offenders against children. Arch. Sex. Behav. 44, 45–66. discrete or overlapping typologies? Child Abuse Neglect 79, 434–444.
doi: 10.1007/s10508-014-0270-x doi: 10.1016/j.chiabu.2018.02.021
Bandyopadhyay, S. (2017). KIRAT: A Note on Economic Evaluation. Report Cassell, P. G., Marsh, J. R., and Christiansen, J. M. (2013). The case for
written as part of Better Policing Collaborative work funded by HEFCE Police full restitution for child pornography victims. George Washington Law Rev.
Knowledge Fund. Available on request. 82, 61–110.

CEOP (2013). Threat Assessment Sexual Exploitation. Available online at: https:// Heeks, M., Reed, S., Tafsiri, M., and Prince, S. (2018). The economic and social
childhub.org/en/child-protection-online-library/threat-assessment-child- costs of crime second edition (Home Office Research Report 99). London:
sexual-exploitation-and-abuse-2013 (accessed January 25, 2021). Home Office.
Cohen, L., and Galynker, I. (2002). Clinical features of pedophilia HM Revenue and Customs Exchange Rates (2011). HM Revenue and
and implications for treatment. J. Psychiatr. Pract. 8, 276–289. Customs Exchange Rates. Available online at: https://www.gov.uk/government/
doi: 10.1097/00131746-200209000-00004 publications/exchange-rates-for-customs-and-vat-yearly (accessed September
Cohen, M. A., Rust, R. T., Steen, S., and TIDD, S. T. (2004). Willingness- 8, 2020).
to-pay for crime control programs. Criminology 42, 89–110. Home Office (2011). Revisions Made to the Multipliers and Unit Costs of Crime
doi: 10.1111/j.1745-9125.2004.tb00514.x Used in the Integrated Offender Management Value for Money Toolkit,
Committee on the Judiciary (2008). Sex Crimes and the Internet. US Government September 2011. Available online at: https://www.gov.uk/government/uploads/
Printing Office. Available online at: https://www.gpo.gov/fdsys/pkg/CHRG- system/uploads/attachment_data/file/97813/IOM-phase2-costs-multipliers.
110hhrg38335/html/CHRG-110hhrg38335.htm (accessed September 1, 2020). pdf (accessed August 22, 2020).
Daly, K., and Boughours, B. (2010). Rape and attrition in the legal process: a Howard, P., Barnett, G., and Mann, R. (2014). Specialization in and within
comparative analysis of five countries. Crime Justice Rev. Res. 39, 565–650. sexual offending in England and Wales. Sex. Abuse J. Res. Treat. 26, 225–251.
doi: 10.1086/653101 doi: 10.1177/1079063213486934
Dawson, S. J., Bannerman, B. A., and Lalumiere, M. L. (2014). Paraphilic interest: Hughes, D., Walkerdine, J., Coulson, G., and Gibson, S. (2006). Peer-to-peer: is
an examination of sex differences in a nonclinical sample. Sex. Abuse J. Res. deviant behavior the norm on P2P file-sharing networks? IEEE Distributed Syst.
Treat. 28, 20–45. doi: 10.1177/1079063214525645 Online 7, 1–11. doi: 10.1109/MDSO.2006.13
Dolan, P., Loomes, G., Peasgood, T., and Tsuchiya, A. (2005). Estimating Hunt, P., Anderson, J., and Saunders, J. (2017). The price of justice: new national
the intangible victim costs of violent crime. Br. J. Criminol. 45, 958–976. and state-level estimates of the judicial and legal costs of crime to taxpayers.
doi: 10.1093/bjc/azi029 Am. J. Crim. Justice 42, 231–254. doi: 10.1007/s12103-016-9362-6
Dombert, B., Schmidt, A., Banse, R., Briken, P., Hoyer, J., Neutze, J., et al. (2016). Jacques, R. W. (2011). Amy and Vicky’s cause: perils of the federal restitution
How common is men’s self reported sexual interest in prepubescent children? framework for child pornography victims. Georgia Law Rev. 45, 1167–1198.
J. Sex Res. 53, 214–223. doi: 10.1080/00224499.2015.1020108 Jonsson, L. S., Fredlund, C., Priebe, G., Wadsby, M., and Göran Svedin, C.
Dombroski, S., Gischlar, K., and Durst, T. (2007). Safeguarding young people (2019). Online sexual abuse of adolescents by a perpetrator met online:
from cyber pornography and cyber sexual predation: a major dilemma for the a cross-sectional study. Child Adolesc. Psychiatry Ment. Health 13, 1–10.
internet. Child Abuse Rev. 16, 153–170. doi: 10.1002/car.939 doi: 10.1186/s13034-019-0292-1
Dubourg, R., Hamed, J., and Thorns, J. (2005). The economic and social costs of Joyal, C. C. (2014). How anomalous are paraphilic interests? Arch. Sex. Behav. 43,
crime against individuals and households 2003/04. London: Home Office Online 1241–1243. doi: 10.1007/s10508-014-0325-z
Report, 30/05. Kingston, D., Fedoroff, P., Firestone, P., Curry, S., and Bradford, J. (2008).
Eke, A. W., Helmus, L. M., and Seto, M. C. (2019). A validation study of Pornography use of sexual aggression: the impact of frequency and type of
the child pornography offender risk tool (CPORT). Sexual Abuse 4:456. pornography use on recidivism amongst sexual offenders. Aggressive Behav. 34,
doi: 10.1177/1079063218762434 341–351. doi: 10.1002/ab.20250
Elkin (2020), Child Sexual Abuse in England and Wales, Year Ending March 2019. Koontz, L. D. (2004). File sharing program: users of peer-to-peer networks
Office for national statistics. Available online at: www.ons.gov.uk (accessed can readily access child pornography: GAO-04-757T (GAO Reports). U.S.
August 26, 2020). Government Accountability Office. Available online at: https://www.gao.gov/
Fang, X. M., Brown, D. S., Florence, C. S., and Mercy, J. A. (2012). The assets/120/110915.pdf (accessed January 25, 2021).
economic burden of child maltreatment in the United States and implications Leary, M. (2007). Self-produced child pornography: the appropriate societal
for prevention. Child Abuse Negl. 36, 156–165. doi: 10.1016/j.chiabu.2011. response to juvenile self-sexual exploitation. Virginia J. Soc. Policy Law 1, 1–50.
10.006 Letourneau, E. J., Brown, D. S., Fang, X. M., Hassan, A., and Mercy, J. A. (2018).
Gelles, R. J., and Perlman, S. (2012). Estimated Annual Cost of Child The economic burden of child sexual abuse in the United States. Child Abuse
Abuse and Neglect. Chicago IL: Prevent Child Abuse America. Available Negl. 79, 413–422. doi: 10.1016/j.chiabu.2018.02.020
online at: https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.691.820& Long, M., Alison, L., Tejeiro, R., Hendricks, E., and Giles, S. (2016). KIRAT: law
rep=rep1&type=pdf (accessed September 8, 2020) enforcement’s prioritization tool for investigation indecent image offenders.
Goller, A., Graf, M., Frei, A., and Dittmann, V. (2010). PW01-117—recidivism Psychol. Public Policy Law. 22, 12–21. doi: 10.1037/law0000069
of internet sex offenders—an epidemiological study on more than 4600 McCollister, K., French, M., and Fang, H. (2010). The cost of crime to society:
offenders in Switzerland. Eur. Psychiatry 25:1533. doi: 10.1016/S0924-9338(10) new crime-specific estimats for policy and program evaluation. Drug Alcohol
71516-9 Depend 108, 98–109. doi: 10.1016/j.drugalcdep.2009.12.002
Graf, M., and Dittmann, V. (2011). “Forensic-psychiatric treatment for internet Murray, J. (2000). Psychological profile of pedophiles and child molesters. J.
sex offenders,” in International Perspectives on the Assessment and Treatment Psychol. 134, 211–224. doi: 10.1080/00223980009600863
of Sexual Offenders: Theory, Practice, and Research, eds D. P. Boer, R. Eher, Neutze, J., Grundmann, D., Scherner, G., and Beier, K. M. (2012). Undetected
L. A. Craig, M. H. Miner, and F. Pfafflin (West Sussex: Wiley), 479–488. and detected child sexual abuse and child pornography offenders. Int. J. Law
doi: 10.1002/9781119990420.ch25 Psychiatry 35, 168–175. doi: 10.1016/j.ijlp.2012.02.004
Hall, R. C., and Hall, R. C. (2009). A profile of pedophilia: definition, characteristics New, M., and Berliner, L. (2000). Mental health service utilization by victims of
of offenders, recidivism, treatment outcomes, and forensic issues. FOCUS 7, crime. J. Trauma. Stress 13, 693–707. doi: 10.1023/A:1007818402277
522–537. doi: 10.1176/foc.7.4.foc522 Olavarria-Gambi, M. (2007). The economic cost of crime in Chile. Glob. Crime 8,
Hamilton-Giachritsis, C., Hanson, E., and Whittle Beech, A. (2017). “Everyone 287–310. doi: 10.1080/17440570701739686
deserves to be happy and safe” the impact of online and offline sexual abuse. Peterson, C., DeGue, S., Florence, C., and Lokey, C. N. (2017). Lifetime
NSPCC. Available online at: https://learning.nspcc.org.uk/media/1123/impact- economic burden of rape among U.S. adults. Am. J. Prev. Med. 52, 691–701.
online-offline-child-sexual-abuse.pdf (accessed September 11, 2020). doi: 10.1016/j.amepre.2016.11.014
Hammond, S., Quayle, E., Kirakowski, J., O’Halloran, E., and Wynne, F. (2009). An Peterson, C., Liu, Y., Kresnow, M.-J., Florence, C., Merrick, M. T., DeGue,
examination of problematic paraphilic use of peer to peer facilities. Available S., et al. (2018). Short-term lost productivity per victim: intimate partner
online at: http://antipaedo.lip6.fr/T24/TR/hebe.pdf. violence, sexual violence, or stalking. Am. J. Prev. Med. 55, 106–110.
Hanson, E. (2017). “The impact of online sexual abuse on children and young doi: 10.1016/j.amepre.2018.03.007
people,” in Online Risk to Children: Impact, Protection and Prevention, ed J. Post, L. A., Mezey, N. J., Maxwell, C., and Wibert, W. N. (2002). The rape tax
Brown (London: Wiley), 97–122. doi: 10.1002/9781118977545.ch6 - tangible and intangible costs of sexual violence. J. Interpers. Violence 17,
Heaton, R., and Tong, S. (2016). Evidence-based policing: from effectiveness to 773–782. doi: 10.1177/0886260502017007005
cost-effectiveness. Policing 10, 60–70. doi: 10.1093/police/pav030 Quayle, E. (2008). The COPINE project. Ir. Probat. J. 5, 65–83.

Saied-Tessier, A. (2014). Estimating the Costs of Child Sexual Abuse in the in Spain. Child Abuse Negl. 98:104178. doi: 10.1016/j.chiabu.2019.
UK. Available online at: https://letterfromsanta.nspcc.org.uk/globalassets/ 104178
documents/research-reports/estimating-costs-child-sexual-abuse-uk.pdf Steel, C. M. S. (2009). Child pornography in peer-to-peer networks. Child Abuse
(accessed January 25, 2021). Negl. 33, 560–568. doi: 10.1016/j.chiabu.2008.12.011
Santtila, P., Antfolk, J., Rafsa, A., Hartwig, M., Sariola, H., Sandnabba, Tennessee, A. M., Bradham, T. S., White, B. M., and Simpson, K. N. (2017). The
N., et al. (2015). Men’s sexual interest in children: one-year incidence monetary cost of sexual assault to privately insured US women in 2013. Am. J.
and correlates in a population based sample of finnish male Public Health 107, 983–988. doi: 10.2105/AJPH.2017.303742
twins. J. Child Sex Abuse 24, 115–134. doi: 10.1080/10538712.2015. Thornton, A., Sidebottom, A., Belur, J., Tompson, L., and Bowers, K.
997410 (2019). On the development and application of EMMIE: insights from
Santtila, P., Mokros, A., Hartwig, M., Varjonen, M., Jern, P., Witting, K., et al. the what works centre for crime reduction. Policing Soc. 29, 266–282.
(2010). Childhood sexual interactions with other children are associated with doi: 10.1080/10439463.2018.1539483
lower preferred age of sexual partners including sexual interest in children in von Weiler, J., Haardt-Becker, A., and Schulte, S. (2010). Care and treatment of
adulthood. Psychiatry Res. 1–2, 154–159. doi: 10.1016/j.psychres.2008.10.021 child victims of child pornographic exploitation (CPE) in Germany. J. Sex.
Sentencing Council (2012). Sexual Offenses Sentencing Data. Analysis and Aggression 16, 211–222. doi: 10.1080/13552601003759990
Research Bulletin. Available online at: www.sentencingcouncil.org.uk (accessed Wakeling, H., Howard, P., and Barnett, G. (2011). Comparing the validity of
September 1, 2020). the RM2000 scales and OGRS3 for predicting recidivism by Internet sexual
Seto, M. (2008). Pedophilia and Sexual Offending Against Children: Theory, offenders. Sex. Abuse J. Res. Treat. 23, 146–168. doi: 10.1177/1079063210375974
Assessment and Intervention. Washington, DC: American Psychological Walby, S. (2004). “The Cost of Domestic Violence,” September 2004, Woman and
Association. doi: 10.1037/11639-000 Equality Unit funded by the DTI. Available online at: https://paladinservice.
Seto, M. (2018) Pedophilia and Sexual Offending Against Children: Theory, co.uk/wp-content/uploads/2013/07/cost_of_dv_research_summary-Walby-
Assessment and Intervention, 2nd Edn. Washington, DC: American 2004.pdf (accessed September 8, 2020).
Psychological Association. doi: 10.1037/0000107-000 Wang, C., and Holton, J. K. (2007). Total Estimated Cost of Child Abuse and Neglect
Seto, M. C. (2009). Pedophilia. Annu. Rev. Clin. Psychol. 5, 391–407. in the United States. Chicago: Prevent Child Abuse America. Available online
doi: 10.1146/annurev.clinpsy.032408.153618 at: http://noexcuseforchildabuseinamerica.com/resources/Economic+Impact+
Seto, M. C. (2013). Internet Sex Offenders. Washington, DC: American Study+Cost+Of+Child+Abuse.pdf (accessed September 8, 2020)
Psychological Association. doi: 10.1037/14191-000 Williams, K. M., Cooper, B. S., Howell, T. M., Yuille, J. C., and Paulhus, D. L.
Seto, M. C., and Eke, A. W. (2015). Predicting recidivism among adult male child (2009). Inferring sexually deviant behavior from corresponding fantasies: the
pornography offenders: development of the child pornography offender risk role of personality and pornography consumption. Crim. Justice Behav. 36,
tool (CPORT). Law Hum. Behav. 39, 416–429. doi: 10.1037/lhb0000128 198–222. doi: 10.1177/0093854808327277
Seto, M. C., Hanson, R. K., and Babchishin, K. M. (2011). Contact sexual Wolak, J., Liberatore, M., and Levine, B. N. (2014). Measuring a year of child
offending by men with online sexual offenses. Sex. Abuse 23, 124–145. pornography trafficking by U.S. computers on a peer-to-peer network. Child
doi: 10.1177/1079063210369013 Abuse Negl. 38, 347–356. doi: 10.1016/j.chiabu.2013.10.018
Seto, M. C., Hermann, C. A., Kjellgren, C., Priebe, G., Svedin, C. G., and Wurtele, S. K., Simons, D. A., and Moreno, T. (2014). Sexual interest in children
Långström, N. (2015). Viewing child pornography: prevalence and correlates in among an online sample of men and women: prevalence and correlates. Sex.
a representative community sample of young Swedish men. Arch. Sex. Behav. Abuse J. Res. Treat. 26, 546–568. doi: 10.1177/1079063213503688
44, 67–69. doi: 10.1007/s10508-013-0244-4 Yang, J., Miller, T. R., Zhang, N., Lehew, B., and Peek-Asa, C. (2014). Incidence
Shanahan, M., and Donato, R. (2001). Counting the cost: estimating and cost of sexual violence in Iowa. Am. J. Prev. Med. 47, 198–202.
the economic benefit of pedophile treatment programs. doi: 10.1016/j.amepre.2014.04.005
Child Abuse Negl. 25, 541–555. doi: 10.1016/S0145-2134(01)0
0225-3 Conflict of Interest: The authors declare that the research was conducted in the
Sherman, L., Neyroud, P. W., and Neyroud, E. (2016). The Cambridge crime absence of any commercial or financial relationships that could be construed as a
harm index: measuring total harm from crime based on sentencing guidelines. potential conflict of interest.
Policing 3:171. doi: 10.1093/police/paw003
Smith, R., Jorna, P., Sweeney, J., and Fuller, G. (2014). Counting the Costs of Copyright © 2021 Giles and Alison. This is an open-access article distributed
Crime in Australia: A 2011 Estimate. Research and public Policy Series No. 129. under the terms of the Creative Commons Attribution License (CC BY). The use,
Canberra, ACT: Australian Institute of Criminology. https://www.aic.gov.au/ distribution or reproduction in other forums is permitted, provided the original
publications/rpp/rpp129 (accessed September 8, 2020). author(s) and the copyright owner(s) are credited and that the original publication
Soldino, V., Carbonell-Vayá, E. J., and Seigfried-Spellar, K. C. (2019). in this journal is cited, in accordance with accepted academic practice. No use,
Criminological differences between child pornography offenders arrested distribution or reproduction is permitted which does not comply with these terms.

ORIGINAL RESEARCH
published: 09 April 2021
doi: 10.3389/fpsyg.2021.623668
Characteristics and Behaviors of

Anonymous Users of Dark Web
Platforms Suspected of Child Sexual
Offenses
Jessica Woodhams 1* , Juliane A. Kloess 1 , Brendan Jose 2 and
Catherine E. Hamilton-Giachritsis 3
1
Centre for Applied Psychology, School of Psychology, University of Birmingham, Birmingham, United Kingdom, 2 WMG,
The University of Warwick, Warwick, United Kingdom, 3 Department of Psychology, University of Bath, Bath, United Kingdom
International law enforcement have noted a rise in the use of the Dark Web to facilitate
and commit sexual offenses against children, both prior to and since the start of the
COVID-19 pandemic. The study presented here therefore aimed to investigate the
characteristics and behaviors of anonymous users of Dark Web platforms who were
suspected of engaging in the sexual abuse of children. Naturally-occurring data on
53 anonymous suspects, who were active on the Dark Web and had come to police
attention in the United Kingdom (UK), were sampled. Analysis of the data yielded 462
Edited by:
Nadia Marie Wager, features that could be coded reliably. Analysis of these features provided novel insights
University of Huddersfield, into suspects’ characteristics, their motivations for using the Dark Web, the nature of the
United Kingdom
offending behavior they reported engaging in, their technical and security precautions,
Reviewed by:
Kathryn Sharratt,
sexual interests, and the content of their interactions with one another. Findings are
University of Huddersfield, discussed in relation to theoretical and practical implications, as well as directions for
United Kingdom
future research.
Jonas Krüppel,
FernUniversität Hagen, Germany Keywords: online (Internet) child sexual abuse, Internet sex offending, offender characteristics, TOR, dark web
*Correspondence:
Jessica Woodhams
J.Woodhams@bham.ac.uk INTRODUCTION
Specialty section: Internet technologies have been purported to serve various functions for individuals with a sexual
interest in children and/or those who engage in the sexual exploitation and abuse of children
(hereafter CSEA1 ). Online communities that are geared toward users with an interest in CSEA
Frontiers in Psychology provide access to: (i) material depicting child sexual abuse (Quayle and Taylor, 2001); (ii) an
opportunity to communicate with like-minded individuals (Durkin, 1997; Europol, 2014); (iii)
Received: 30 October 2020
Accepted: 12 March 2021
a sense of belonging and acceptance (Quayle and Taylor, 2002; Holt et al., 2010); and (iv)
Published: 09 April 2021 an environment in which sexual fantasies can be shared, verified and gratified (Davidson and
Citation:
Gottschalk, 2011). The latter is very reinforcing given that CSEA is self-/other-justified (Quayle
Woodhams J, Kloess JA, Jose B et al., 2000; Quayle and Taylor, 2003; Holt et al., 2010). Clearly, expressing such thoughts and
and Hamilton-Giachritsis CE (2021) behaviors would be highly stigmatizing in society outside of such communities.
Characteristics and Behaviors
of Anonymous Users of Dark Web
Platforms Suspected of Child Sexual 1
A range of terms are used to refer to the sexual exploitation and abuse of children. For the purpose of this paper, we
Offenses. Front. Psychol. 12:623668. will use the term ‘child sexual exploitation and abuse,’ in line with its usage by WeProtect (n.d.) and the United Kingdom
doi: 10.3389/fpsyg.2021.623668 Government (Home Office, 2019).

1 April 2021 | Volume 12 | Article 623668
Woodhams et al. Anonymous Users of Dark Web Platforms
Online communities of individuals with a sexual interest in nature of these platforms on the Dark Web, including TOR2 ,
children have existed on the Surface Web for some time and they are spaces of the World Wide Web that are not stumbled
several reviews of the literature regarding the characteristics of upon accidentally. For some forums and other restricted areas,
online child sex offenders have been conducted (e.g., Babchishin vetting of a potential member is required, which involves them
et al., 2011; Seto, 2013; Kloess et al., 2014; Ly et al., 2018). having to supply new CSEA material for entry and on an ongoing
For example, in a meta-analysis by Babchishin et al. (2011), basis in order to retain membership (Europol, 2014). As a
the authors compared studies on online sexual offenders (who result, one might expect that members of such forums, and/or
used the Internet to facilitate their offending) to those on offline other restricted areas, are highly motivated to offend, given the
offenders (who committed contact sexual offenses), and found effort expended to access them. Restricted areas are associated
that online offenders were more likely to be younger and of an with discussions and material that are of a more violent and
ethnic minority. They were also found to be less likely to have sadistic nature (Europol, 2014), and the Dark Web is home
a history of physical abuse, report fewer cognitive distortions, to bulletin boards for niche sexual interests, including CSEA
including less emotional congruence with children, and present that involves very young children and sadism (Europol, 2020).
with less socially desirable responding. Compared to offline Individuals who utilize the Dark Web for the various purposes
offenders, they scored higher on sexual deviance and victim related to CSEA may therefore present with more deviant sexual
empathy. When compared to the wider population, they were interests, for which they are less likely to find a suitable outlet on
more likely to report having experienced physical and sexual the Surface Web.
abuse, and being unemployed, as well as less likely to be, or Historically, individuals operating on the Dark Web would
have been, married. have required a certain level of technological expertise. However,
More recently, international law enforcement has recorded a Internet communication platforms on the Dark Web are
rise in the use of the Dark Web by individuals for the purpose of becoming increasingly more accessible (Europol, 2016), meaning
engaging with the topic of CSEA and related material (Europol, that less technologically-sophisticated users are now able to join
2014, 2016, 2020; National Crime Agency, 2018). The Dark Web the various online communities available. Balfe et al.’s (2015)
refers to “content on the World Wide Web that is not indexed review of the precautions employed by Internet child sexual
by standard search engines” (Weimann, 2016, p. 175). In 2016, offenders found that the security measures they employed varied
Europol reported that both the number of Dark Web forums depending on their age and level of experience, with those who
dedicated to pedophilia and CSEA, and the volume of material were younger and less experienced taking fewer precautions.
being exchanged thereon, had been increasing. This includes both Overall, across the sample, very few offenders adopted security
‘known’ and first-generation (i.e., new) CSEA material. Similarly, measures, although it should be noted that while the study
Owen and Savage (2015) collected data on the Dark Web over sampled 11 years of empirical research (from 2000 to 2011), this
a period of 6 months, and found that ‘abuse sites’ were by far would now be considered dated. In addition, the offenders in
the most popular (represented by 80% of the total number of the samples from the studies that were included in the review
requests). Since the outbreak of COVID-19, Europol (2020) have had all been apprehended, which is likely to have occurred as
seen a further rise in activity on Dark Web forums dedicated a result of their low level of technical sophistication, and the
to CSEA. Alongside the rise in frequency, the content depicted precautions they took (if any). Our paper, in part, responds
in CSEA material has become increasingly more extreme and to the call by Balfe et al. (2015) for further research on the
violent (Europol, 2014). precautions taken by online CSEA offenders, by studying a
Membership of Dark Web forums dedicated to CSEA is not recent sample of anonymous online CSEA offenders who used
insignificant. Web-Iq (2018) reported that within seven such Dark Web platforms, in situ, and who were undetected at the
forums they indexed, there were more than two million unique time of data collection (i.e., their offline identities had not
user IDs. While some users register on multiple forums, it was been resolved).
estimated that this number equates to between 300,000 and 1 As well as studying the security precautions taken by online
million users across the seven forums. In light of these findings, CSEA offenders, researchers have considered the types and nature
enhancing law enforcement’s understanding of the threat posed of interactions between members of these online communities in
by users on these forums has been identified as a key priority, the Surface Web. Holt et al. (2010) note that such communities
as has the prioritization of users suspected of engaging in CSEA are self-policing in that members police one another’s behavior
for the purpose of identifying them and disrupting offending to avoid security breaches. Security measures and avoiding
behavior (Europol, 2016). We cannot assume that behavior detection are therefore a common discussion topic in such
displayed by users on the Dark Web is the same as behavior communities on the Surface Web (see also Quayle and Taylor,
displayed by users on the Surface Web – further research into 2001; Davidson and Martellozzo, 2008; Europol, 2015). Members
this is therefore needed in order to ensure that law enforcement also offer advice, and share ‘tips’ with one another, on accessing
efforts are underpinned by a relevant evidence base. victims (Quayle and Taylor, 2001; Davidson and Martellozzo,
While CSEA offenders who utilize Internet communication 2008; Europol, 2014, 2015; Web-Iq, 2018).
platforms on the Surface Web as part of their offending behavior
have been studied for some time, there is limited knowledge 2
TOR (The Onion Router) is software which enables anonymous communication
around and understanding of the behavior of CSEA offenders via a free open network with multiple encryption layers (Huang and Bashir, 2016);
who make use of such platforms on the Dark Web. By the sheer use of TOR increases the difficulty of tracing internet usage by an individual.

O’Halloran and Quayle’s (2010) study of a support forum samples to inform the policing of offenders who have not
for ‘boy lovers’ focused on the justifications used and the yet been identified. This is because their characteristics and
excuses given by self-identified pedophiles who were members behavior may be related to the reason(s) for their apprehension,
of the forum (sometimes in response to members who were and as such differ in important ways to those who remain
anti-pedophilia). Justifications focused on denying harm or unapprehended (e.g., Bennell and Canter, 2002; Neutze et al.,
injury to the child as a result of the abuse, claiming that 2012). For example, apprehended child sexual offenders may
the abuse actually benefitted the child, and that the child show greater stability in their offending behavior and patterns
deserved or attracted sexual advances from the adult(s) (i.e., than those who are unapprehended; this would be similar to
denial of victimhood). While not the primary focus of their findings that unapprehended serial rapists have higher levels of
study, O’Halloran and Quayle (2010) also reported on the polymorphism than those observed in apprehended serial rapists
purpose of other communications between members. Like the (Lovell et al., 2017). In addition, research on unapprehended
studies above, these included: (i) policing the forum to prevent child sexual offenders has found that they differ from those
infiltration (e.g., by law enforcement), (ii) protecting one another who have been apprehended in terms of their level of sexual
(by giving advice), and (iii) setting ground rules regarding preoccupation, breadth of paraphilic interests, coping styles,
what were and what were not the purposes of the forum. sexual abuse victimization, education, as well as employment
Communications that acknowledged the causation of harm to a (Neutze et al., 2012). When law enforcement personnel are
child were very rare. interacting with suspects of CSEA online, whether assuming the
The range of forums available on the Dark Web include those identity of a child or an adult (Wortley and Smallbone, 2012;
that are similar in focus to the ones included in O’Halloran Martellozzo, 2015), it is important that they are aware of the
and Quayle’s (2010) study on the Surface Web. We can current knowledge base regarding this group of users, and for
therefore expect communities on the Dark Web to display this, as well as related research findings, to be incorporated into
similar behaviors. However, within the Dark Web, there are relevant training.
forums and bulletin boards that are specifically dedicated to
CSEA that is of a sadistic nature, which is very different given
its explicit focus on causing harm through both physical and MATERIALS AND METHODS
psychological means (Europol, 2014, 2020). Here, discussions
centered around sexual offenses committed against children Ethics Statement
would likely be characterized by entitlement, indifference to The research project was granted full ethical approval by
victimsŠ suffering, and a lack of remorse (Beech et al., 2005; the Science Technology Engineering and Mathematics Ethics
Mokros et al., 2011). Committee at the University of Birmingham [ERN_14-1435E],
In summary, while we can hypothesize what similarities and and the Psychology Research Ethics Committee at the University
differences there might be in terms of the characteristics and of Bath. Written informed consent from the participants was not
behaviors of individuals who use the Surface Web or the Dark required to participate in this study in accordance with national
Web (for the purpose of committing offenses relating to CSEA), legislation and the institutional requirements. All members of
to date, there has been no study that has examined in depth what the research team received security clearance to undertake
these look like on various platforms on the Dark Web3 . Our study research within a United Kingdom police force. The research
was therefore exploratory and aimed to document, in detail, the project also received scrutiny from the United Kingdom’s
characteristics and behaviors of these suspects. Such a study is Information Commissioner’s Office and the Office for the
needed because existing literature on suspects who operate on Surveillance Commissioner, now the Investigatory Powers
CSEA sites on the Dark Web neglect to report their methodology Commissioner’s Office.
(e.g., Web-Iq, 2018), or their findings are based on surveys of
law enforcement professionals’ perceptions of the threat (e.g., Sample/Participants
Europol, 2015). In addition, much of the previous research Our sample constituted 53 anonymous individuals who were
on Internet child sexual offenders has derived findings from active on the Dark Web between October 2014 and November
convicted, and therefore apprehended, offenders [see Seigfried- 2016, and by virtue of the nature of the platform they were using,
Spellar’s (2014) study of Internet child pornography consumers had come to the attention of a large United Kingdom police force.
for an exception]. Based on the information the individuals had disclosed to others
Users encountered by law enforcement on Dark Web in the Dark Web, they were suspected by the police of committing
platforms are anonymous. As such, their offline identities or having committed a contact sexual offense involving a child
are not (yet) known, and they have yet to be apprehended. and/or an offense relating to CSEA material (i.e., possession,
While sampling apprehended offenders makes sense when using distribution, and production). At the point of data collection,
research findings to inform risk assessment procedures and none of the suspects were subject to a formal police investigation.
psychological interventions, it is more problematic to use such Some became the subject of a formal police investigation at a later
stage, but this was not the case for all of them. The 53 suspects
3 were members of and contributed to one of four different TOR
This is not to suggest that suspects who use platforms on the Dark Web are a
special type of offender, and we acknowledge that such individuals are likely to use forums (geared toward CSEA) that were under surveillance by the
the Surface Web too (e.g., Kokolaki et al., 2020). United Kingdom police force at the time the research was being

conducted (n = 27; 51%), alternatively using Internet Relay Chat4 as a suspect. CSEA material (i.e., in the form of both still and
(IRC, n = 19; 73%), TOR chat (n = 6; 23%), or a peer-to-peer file moving images) was included in the data of 21 suspects (40%
sharing network5 (n = 1; 4%). of the sample), where it formed part of the interactions between
In terms of the nature of the four forums, seventeen suspects suspects, or was the subject of forum postings. However, this
were members of a forum which presented children as sexual material is not described further here, as it is not the focus of the
beings with whom to have a loving relationship, and condoned analysis presented in the present article.
material depicting the sexual abuse of babies and toddlers (i.e., All of the data available for each suspect were included in
Forum 1). Eight suspects were members of a second forum, the analysis; the data were the original contributions made by
which was the largest forum on the Dark Web at the time of suspects, as recorded on the relevant Internet communication
data collection (i.e., Forum 2). Members of Forum 2 varied platform. The number of pages of available data per suspect
in terms of sexual interest in age and type of activity, but varied and ranged from 1 to 12 transcripts (M = 4.63, SD = 3.66),
torture and/or sadistic acts of abuse were not tolerated. One and 1 to 411 pages (M = 25.82, SD = 73.76) respectively,
suspect was a member of a forum dedicated to the torture of resulting in a total of 1265 pages (single-spaced format, font
children and sadism (i.e., Forum 3). A fourth forum, of which size 12) of data.
one suspect was a member, had varied content but was more
focused on torture and sadism (including the killing of children) Data Analysis
(i.e., Forum 4). All of the forums were invite-only, including The transcripts of forum postings and private emails/messages
specific criteria that had to be met when a suspect wished to were formatted to enable their importing into MAXQDA12,
apply for membership. a professional software package, to facilitate the process of
During data collection, all suspects were anonymous (i.e., qualitative data analysis. The purpose of the coding was to
their real-world identities were unknown to the police), however, identify features that described the online activity and behavior
11 of the suspects’ identities were subsequently resolved and of the suspects. As features were identified, they were recorded in
the individuals behind the usernames were charged with having the coding scheme as a code using a descriptive label. The textual
committed sexual offenses against children. Eight of the 11 data was annotated with the relevant code by highlighting the
offenders (73%) were charged with possession of CSEA material, relevant area of text and linking it to the corresponding codes or
and three (27%) were charged with committing a contact sexual sub-codes in the coding framework. The coding and qualitative
offense against at least one child. analysis was completed by the second author on a line-by-line
basis. An inductive, content-driven approach was taken in order
Procedure to allow for the identification of any key information, trends,
Data were collected on a secure police site. At the time of themes or ideas apparent across the data (Guest et al., 2012).
data collection, the researchers were kept blind to the type of Additional codes were developed as new features were
investigation being conducted by the police, and the suspicions identified, while reappearing elements were assigned codes that
regarding the suspect, in order to avoid biasing the coding of already existed within the coding scheme. Codes and sub-codes
data. The data used in the research project consisted of forum6 were grouped in order to form superordinate codes relating to the
postings and private emails/messages, as well as digital material key concepts under investigation. A hierarchical grouping style
in the form of images and videos that came from the main facilitated the organization of codes and sub-codes according
forum used by each suspect. These data were derived by means to their similarity and relationship with each other within
of extracting existing forum postings from networks on the Dark these key concepts.
Web for the relevant usernames under investigation. Forum data Throughout the analysis, an iterative approach was employed
was limited to the posts of the suspects under investigation rather by revisiting and adjusting existing codes and sub-codes, as
than also including the posts of other forum members for reasons well as revising them, where appropriate. Additionally, textual
of privacy and proportionality. These data were not, therefore, data within codes and sub-codes were re-examined to ensure
interactive. For 26 of the 53 suspects (49%), data from IRC that codes described the information within them accurately.
was available; this was interactive and represented interactions Transcripts were re-read and segments of text re-coded, where
between the suspect and an undercover police officer, posing necessary. If new information or insight was gained, the
coding scheme was modified to further develop and refine
4
Internet Relay Chat: IRC was created in 1988 and enables real-time text it. The descriptions of codes and sub-codes were developed
messaging between internet-connected computers. It is mainly used for group throughout, as well as verified and revised through discussions
discussion in chat rooms but can also be used for private message exchange with the research team.
between two users (Radware: https://security.radware.com/ddos-knowledge-
In order to determine whether the data could be coded reliably
center/ddospedia/irc-internet-relay-chat/#:~{}:text=IRC%20(Internet%20Relay%
20Chat)%20is,side%20and%20client%2Dside%20commands). using the coding scheme, inter-rater reliability was assessed once
5
Peer-to-peer file sharing is the distribution and sharing of digital media using an initial coding scheme had been developed. This coding scheme
peer-to-peer (P2P) networking technology. P2P file sharing allows users to access contained 479 variables (codes). All coding was conducted by
media files, using a P2P software program that searches for other connected the second author, with a second researcher coding 10% of the
computers on a P2P network to locate the desired content.
6 data derived from the first five cases (i.e., 10% of the textual
An Internet forum, or message board, is an online discussion site used by users to
hold conversations in the form of posted messages. Some forums have an option data available for each of the first five cases). This researcher was
for exchanging private messages. employed on the research project for the purpose of analyzing

TABLE 1 | Inter-rater reliability values across cases. characteristics displayed by suspects as part of their online
communication/interactions, a range of features were coded.
Case 1 Case 2 Case 3 Case 4 Case 5
These related to (i) the suspects’ demographics (e.g., gender,
% agreement 0.96 0.99 0.98 0.99 0.99 relationship status, and employment); (ii) their self-reported
Kappa 0.77 0.62 0.79 0.73 0.69 motivations for and use of various Internet communication
platforms (e.g., accessing CSEA material, chatting, forums,
and P2P file sharing network); (iii) security measures and
CSEA material (which is the focus of another article, and will precautions suspects reported taking (e.g., not using a
therefore not be presented here). Both coders were experienced webcam/not showing their face); (iv) their self-reported
in conducting qualitative data analysis. For the purpose of sexual interests and likes; (v) their behaviors on the relevant
assessing inter-rater reliability, each coder recorded for each Internet communication platform (e.g., discussing/sharing
variable whether it was present or absent in the material. The modus operandi, encouraging other users to contribute to
level of agreement between the two coders was evaluated using the forum and share material, requesting information and/or
percentage agreement and Cohen’s Kappa. Percentage agreement material); as well as (vi) the various topics contained within
is a simple indicator of reliability; in contrast, Kappa is a posts on the forum (ranging from advice about how to approach
conservative measure of inter-rater reliability, used to assess the and interact with children to how to avoid detection, to
reliability between two coders with categorical or ordinal data, various different materials related to CSEA). Key findings
and with the added benefit that it corrects for chance agreement. are reported below under common themes. Unless otherwise
There can be a Kappa Paradox, where the percent agreement stated, percentages cited are a proportion of the overall
is very high, but the Kappa is very low; this tends to occur sample of 53 suspects.
when almost all cases fall into one category. These measures of
inter-rater reliability were calculated on a case-by-case, as well
as variable-by-variable, basis. Table 1 presents the κ-values for
Suspects’ Demographic Characteristics
the case-wise analysis. All statistics were in the acceptable range, Based on suspects’ self-reported gender, all except one were male.
namely ≥80% for percentage agreement (Hartmann, 1977), and The exception refers to what appeared to be a female suspect –
≥0.61 for Kappa (Landis and Koch, 1977). this assessment was based on their username, and the type of
When reliability was assessed on a variable-by-variable basis, offending behavior and descriptions of abuse reported by them.
60 of the 479 variables achieved a kappa of 1.00, and percentage As defined on the forum, 13% (n = 7) of suspects self-reported
agreement of 100%. 377 variables achieved 100% percentage holding a senior position on the main forum (from which their
non-occurrence agreement (i.e., both coders agreed that the data was derived) (i.e., owner, administrator or moderator). 28%
variable was not present in any of the five cases), and 31 (n = 15) of the suspects reported having biological children of
variables achieved 80% percentage non-occurrence agreement. their own. Two of these suspects (13%) also reported having
Seven variables achieved a Kappa of less than 0.60 (i.e., 0.55), access to other children, and a further three suspects (6%, who
but did achieve an acceptable level of percentage agreement did not disclose having biological children) reported generally
(i.e., 80%). Only four remaining variables achieved both an having access to children (e.g., extended family members). In
unacceptable Kappa value, and a percentage agreement of less total, therefore, 34% (n = 18) of our sample reported having
than 80%. Inspection of the occasions of coding where there was regular access to children.
disagreement revealed no clear pattern of disagreement; instead,
they were genuine coding errors on the part of the second coder. Technical and Precautionary Behavior
The initial coding completed by the primary coder (the second As part of their forum posts, or during conversations/interactions
author) for these five cases was therefore retained. with others, suspects explained why they were using the platforms
Having established inter-rater reliability, the remaining case they did. For some, it was accessing CSEA material (32%; n = 17),
material was coded by the second author. As is to be expected and for others, it was conversing with like-minded users (23%;
with data of this nature, the coding scheme continued to n = 12). Eight suspects (14%) self-reported producing CSEA
be refined, and expanded upon, as new information became material, and 11 suspects (21%) self-reported being in possession
apparent, and new features were identified. This resulted in a final of CSEA material.
coding scheme of 462 variables (codes). The development of the The suspects took a range of precautions in order to
coding scheme, and the coding of the 53 cases, took the duration avoid detection, and to maintain their anonymity. These
of the research project (i.e., 18 months). The length of time taken included: (i) cautioning one another about their behavior (23%;
per case was highly variable, and depended on the amount of data n = 12); (ii) keeping personal information concealed (e.g.,
available per case. face, identifiers, location) (9%; n = 5); (iii) informing one
another about legislation (6%; n = 3); (iv) informing one
another about major legal cases; (v) using encryption software
RESULTS and TOR (9%; n = 5); and (vi) not distributing material
(11%; n = 6). The majority of suspects’ usernames (89%,
A wealth of information was identified in the textual n = 47) were not related to a sexual interest in children/an
data. While our main focus lay with the behaviors and interest in CSEA.

Reported Engagement in Child Sexual interested in a topic was clear from the statements and/or
Abuse postings they made (e.g., “I find X a real turn-on”). Of the
53 suspects, 19 (36%) self-disclosed a liking/interest in one or
The 53 suspects in our sample self-reported engaging in a range
more different types of sexual interest. Sadism (26%; n = 14)
of sexual behaviors against children. For 76% (n = 40) of the
and incest (17%; n = 9) were the sexual interests self-reported
suspects, there was evidence of the possession and collection
by most suspects. The other sexual interests self-reported by
of CSEA material. For 21% (n = 11), there was evidence that
suspects in order of frequency were: bestiality (8%; n = 4),
they had or were inciting another person to sexually abuse a
urination (8%; n = 4), nappies/diapers (6%; n = 3), dressing-
child. Nine percent (n = 5) of the suspects self-reported sexually
up (6%; n = 3), child dolls (6%; n = 3), and feet (4%; n = 2).
abusing a child via Internet technologies (e.g., by means of
All references to a sexual interest in sadism and urination were
Internet communication platforms, including a webcam). For
specifically related to children (as opposed to adults). References
64% (n = 34), there was evidence that they had or were engaging
to incest were clearly in relation to children for four suspects,
in the sexual abuse of a child in person (i.e., committing a contact
however, this could not be conclusively verified for the remaining
sexual offense). While none of the offending behavior engaged
five suspects (although it very likely related to pre-pubescent
in by the sample of 53 suspects resulted in the death of a child,
children due to the nature of the forum in which the material
at least for some, their described actions reportedly involved the
was posted). In terms of the association between the remaining
physical abuse of a child (15%; n = 8), or clearly fulfilled the
sexual interests and children (as opposed to adults, or this
function of a sadistic sexual interest (13%; n = 7).
being unspecified), the representation was as follows: dressing
In terms of cross-over of offending behaviors, for 8% (n = 4)
up (4%; n = 2), nappies (4%; n = 2), feet (2%; n = 1), and
of the 53 suspects, there was evidence that they were sexually
bestiality (2%; n = 1).
exploiting and abusing children both online and in the physical
Where disclosures of sexual interests were made by suspects
world. For 43% (n = 23) of suspects, there was evidence that they
(n = 19), the number of different types self-reported across
were in possession of and/or collected CSEA material, as well as
suspects ranged from one to six: one (47%; n = 9), two (26%;
committed sexual offenses against children in the physical world.
n = 5), three (2%; n = 1), four (4%; n = 2) and six (4%; n = 2).
While it would have been interesting to examine in more detail
Topics Posted any associations between a suspect’s type of sexual interest, as
We analyzed the nature of posts made by suspects on the relevant well as the gender and age (child and/or adult and/or stage of
forums to which they belonged. This is presented as a separate sexual development) of the focus of their sexual interest, there
subsection from our other results because posting to a forum were insufficient cases to facilitate this.
is less interactive than conversing with other offenders, with For 36 suspects, we were able to record their sexual interests in
some suspects solely engaging in posting behavior rather than terms of gender, age, and/or stage of sexual development. Clearly,
interacting with other users. Thus, it was important to capture all 53 suspects were operating on Internet communication
content across all the suspects. platforms that were geared toward users with an interest in CSEA,
A broad range of topics were identified, some of which were and as such it is unsurprising that 35 of the 36 suspects (97%)
merely mentioned by a few suspects (i.e., 2–10%), while others specifically disclosed a sexual interest in children. Of these 35,
were more common. These were predominantly related to sexual four (11%) also self-reported having a sexual interest in adults.
acts (40%; n = 21) (including various forms of penetration, such A further one suspect (2%) specifically disclosed a sexual interest
as anal, oral, and digital), and deviant sexual interests (38%; in adults only. For the remaining 17 suspects (32%), there was no
n = 20), such as bestiality, breast-feeding, incest, female genital explicit disclosure of sexual interest relating to age.
mutilation, feces and sadism, with the latter being one of the In terms of sexual interests around children’s gender, seven
most prevalent (i.e., 30%, n = 16). Some of the posts were about of the 35 suspects (19%) disclosed a sexual interest in children
a particular age group of victims (e.g., babies [15%; n = 8]). of both genders, seven (19%) disclosed a sexual interest in
Issues related to security measures and precautions, with a view female children, and six (17%) disclosed a sexual interest in
to avoiding detection, were also frequently posted, including male children. The remaining 15 out of the 35 suspects (43%)
updates on law enforcement activity (23%; n = 12). Posts also who disclosed a sexual interest in children did not specify any
related to discussions that contrasted individuals with sexual particular interest in terms of the gender of a child. Twenty-
interests in children of different ages (11%; n = 6), as well as nine suspects expressed a sexual interest related to the stage of
predilections for inflicting pain vs. sexual abuse (6%; n = 3). sexual development of a child. Ten suspects (34%) expressed a
Finally, some posts related to the rules of the group and expected sexual interest in children under 5 years of age. The reference
conduct, including topics that were out-of-bounds (21%; n = 11). to ‘under 5 years’ is not a category created by the researchers,
but is used by the suspects themselves to specifically refer to
Sexual Interests children under the age of 5 years. Thirteen (45%) expressed a
Suspects were recorded as presenting with a sexual interest sexual interest in pre-pubescent children, nine (31%) in early-
in a particular topic or subject matter if they self-disclosed pubescent children, and four (14%) in pubescent children. ‘Pre-
it as part of forum postings, conversations or interactions, or pubescent’ refers to children whose body shows no sign of any
where they (re-) posted content/material in a manner that was development of secondary sex characteristics; ‘early-pubescent’
endorsing rather than critical of it. Whether they were sexually refers to children whose body shows initial development of

secondary sex characteristics; ‘pubescent’ refers to children whose (ii) sharing material (26%; n = 14), (iii) responding positively to
body shows development of secondary sex characteristics. material (74%; n = 39), and (iv) re-posting posts (13%; n = 7).
Regarding polymorphism (i.e., cross-over in sexual interest), In our sample of 53 suspects, five (9%) suggested arranging
four of the 36 suspects’ (11%) self-disclosures implied this for a physical meeting with another suspect, and two (4%) offered
age (i.e., child/adult), seven for gender of child (19%), and five their contact details to other users. As might be expected,
suspects self-disclosed a sexual interest in children spanning there was a relatively high occurrence of statements made by
different developmental stages (14%), predominantly the two suspects that sought to (i) normalize and minimize the physical
youngest categories (i.e., under 5 and pre-pubescent). In the and psychological harm suffered by children in the context of
under-5 years and pre-pubescent categories, there was a mix CSEA (34%; n = 18); (ii) support and advocate for child sexual
of suspects sexually interested in females, males and also both abuse (17%; n = 9); and (iii) objectify children (i.e., referring to
genders (where gender was specified by the suspect). children as existing to serve the sexual needs of others, and being
‘deserving’ of abuse) (30%; n = 16).
Suspect-Suspect Interaction Behaviors

The way in which the suspects interacted with one another was DISCUSSION
varied. Given the nature and focus of the forums (from where
most data were derived), it is unsurprising that the topic of CSEA Given the increasing use of the Dark Web by individuals to
featured in most of their interactions. This went beyond merely engage with CSEA material, and interact with like-minded users
reporting engagement in CSEA; for example, suspects discussed on forums that are dedicated to CSEA, it is important to
their experiences of committing sexual offenses against children establish an evidence base on how such individuals operate
(15%; n = 8), their own victimization as children (6%; n = 3), on Dark Web platforms. Our study is novel in terms of
and the sharing of their biological children for sexual abuse by its in-depth examination of the characteristics and behaviors
others (4%; n = 2). They exchanged CSEA material (6%; n = 3), of 53 suspects of CSEA who used a number of different
and engaged in promoting and selling it (2%; n = 1). Dark Web platforms for various purposes related to CSEA.
Communication between suspects, however, was not solely Furthermore, by studying unapprehended suspects and their
related to CSEA, but included general chat and conversation- behaviors in situ, through the use of naturally-occurring data, the
making, as well as sexually explicit chat. The occurrence of present article makes an original contribution to the literature
general chat was more frequent in Dark Web communities on Internet child sexual offending in general, and responds to
than on IRC. Forensic linguistic research recognizes the process calls for more research of unapprehended suspects in particular
of the creation of communities of practice, including in Dark (O’Halloran and Quayle, 2010).
Web forums (Grant and MacLeod, 2018; Chiang, 2019). Within The dataset comprised 53 cases identified from various
our dataset, group identity was apparent through the explicit Internet communication platforms on the Dark Web, which
welcoming of new members (n = 11). In addition, suspects were active at the time of the study, providing conversational
supported each other by offering advice and suggestions around data in the form of forum postings, chat and private messages
(i) ‘finding’ victims and particular material; (ii) modus operandi; that were derived from interactions between suspects and with
(iii) security measures and precautions (e.g., safeness of payment undercover police officers. For those suspects who were members
services and software programs) in order to avoid detection; of forums, they held positions of varying authority (i.e., some
and (iv) how to interact and behave both online and offline were owners, and some were administrators and moderators).
if presenting with a sexual interest in children. As far as we They also varied in how much they contributed to the forum of
could determine, there appeared to be different motivations for which they were a member.
such demonstrations of expertise, with some suspects engaging Regarding our findings of self-disclosed demographic
in posturing and attempting to show superiority. For others, characteristics, like most studies of perpetrators of online CSEA,
it appeared to be shared interests, and the potential for whether apprehended or unapprehended (Beier et al., 2009;
creating new material. Aslan and Edelmann, 2014), the vast majority of our suspects
Acts of dominance were also exhibited, whereby some suspects were male. 34% reported having access to children who were
challenged the behavior and contributions of others. This their own/biological children or extended family members.
included challenging the correctness of information posted (8%; This compares to Long et al.’s (2016) figures of 42%, and 46%
n = 4), the quality of material shared (with regard to the material’s respectively, Beier et al.’s (2009) figures of 28–37%, and Shelton
realness, newness and content in terms of the age of the child et al.’s (2016) figure of 27%, in terms of offenders having access
depicted) (26%; n = 14), the feasibility of the abuse claimed to to and sexually abusing their own children.
have been perpetrated (11%; n = 6), as well as making accusations Many studies of online child sex offenders have focused
that the material had been stolen (4%; n = 2). In addition, on individuals convicted or charged with possession and
suspects expressed disapproval (26%; n = 14), disappointment distribution of CSEA material (i.e., indecent images of children,
(15%; n = 8), and dislike of one another (4%; n = 2). On being sometimes known as “child pornography”), with a smaller
challenged, some suspects responded by apologizing (13%; n = 7). number including within their sample offenders who engaged
Positive endorsements of one another’s behavior were also seen in in the sexual grooming of children (Babchishin et al., 2011).
the form of (i) thanking others for positive comments (8%; n = 4), However, online child sex offending behavior is broader than this

(Clevenger et al., 2016), and this breadth of behavior was evident As noted, a small minority of suspects were explicit about
in our Dark Web sample. The majority of our sample were in having a sexual interest in adults, in addition to a sexual interest
possession and/or collected CSEA material. In addition, most in children. Where it was stated, the number of suspects with a
suspects reported that they were (or had previously) engaged in sexual interest in one gender of children, versus a sexual interest
the sexual abuse of a child in the physical world (i.e., a contact in both genders, was comparable. This contrasts with Neutze
sexual offense). With the advent of digital devices and computers, et al. (2012), where a larger proportion of the sample showed a
it is also possible for offenders to commit very serious sexual gender preference (i.e., 94–96%); however, it should be noted that
offenses against a child via Internet technologies (including a self-disclosures about gender-related sexual interests regarding
webcam and recording facilities; Whittle et al., 2013a; Kloess children were not made by more than half of the sample and, as
et al., 2015), which have been found to be just as harmful as a such, it is unknown whether they had any.
traditional, offline contact sexual offense with regard to mental Within our sample, some suspects self-disclosed specific
health outcomes and subsequent suicidal behavior (Hamilton- sexual interests in and preference for types of CSEA material in
Giachritsis et al., 2017, 2020). As such, it was important to their discussions and postings (e.g., “I don’t mind boys or girls,
capture forms of CSEA that have previously been neglected by but girls under 5 really do it for me”). Most of these suspects
historic papers, or (inadvertently) trivialized through the use of merely self-disclosed one sexual interest, while other suspects
dichotomies of ‘contact-driven’ and ‘fantasy-driven’ to describe self-disclosed a sexual interest in a wide range of types (e.g.,
offenders (Briggs et al., 2011). In our sample, a small number of sadism, incest, bestiality, defecation, and urination). The types
suspects self-reported sexually abusing a child via webcam, and of CSEA material most frequently mentioned by suspects were
one-fifth were inciting others to sexually abuse a child. sadism and incest. This echoes reports from Europol (2014, 2020)
The reasons suspects revealed for using the Dark Web echo of the Dark Web as an enabler for those with sadistic and other
the assertions by Durkin (1997), and those reported in studies of niche sexual interests. In line with Lanning (2012), there was a
other child sexual offenders (e.g., Shelton et al., 2016), with regard subset of suspects who presented with diverse sexual interests,
to using the Internet for purposes related to CSEA: (i) to access whether this was in terms of the nature of the material or the
CSEA material, (ii) to converse with like-minded individuals, type of victim they expressed an interest in. Lanning (2010)
and (iii) to share self-generated material of CSEA with others. also reported a relationship between developmental stage/age and
The efforts required from the suspects in our sample in terms of gender of victim, whereby the older and more sexually developed
accessing the relevant forums, or in obtaining material from other the child, the more likely it was that the offender had a strong
suspects, gives an indication of their motivation to engage in this preference for gender7 . While our data suggest that those who
type of offending behavior. disclose no preference in terms of gender do express a sexual
Risk assessment and risk management by suspects is critical interest in younger, pre-pubescent children, some of the suspects
to their ‘safe’ use of the Internet to facilitate their engagement did not fit this pattern.
in CSEA, whether that be the ‘grooming’ of children for CSEA Suspect-to-suspect behaviors displayed as part of their online
(Whittle et al., 2013b; Kloess et al., 2015; Kloess et al., 2019), or conversations and interactions, as well as topic postings,
maintaining the security of an online community (Holt et al., gave a clear impression of a community of individuals. New
2010; O’Halloran and Quayle, 2010). Risk management and members were welcomed, and there was general conversation-
security measures/precautions were also a notable focus in the making alongside discussions of CSEA. There were examples of
suspects’ communications we sampled. Advice and discussions cooperative behaviors, such as providing assistance and expert
centered around maintaining anonymity by concealing personal advice, as well as problem solving. Much like in other online
identifiers (e.g., face and name), rules of conduct within the communities, posts were ‘liked,’ and assistance or sharing was
online community, what to say to victims to avoid detection acknowledged through thanking. There also appeared to be
by others, advice on technical means of avoiding detection examples of dominance being enacted through speech acts such
(e.g., use of encryption software or TOR), not distributing as directives (Leech, 1983), which varied in their directness from
CSEA material, and updating one another on legislation or new suggestions around the next stage of abuse against a specific child
major police/legal cases. These topics relate to precautionary to giving orders to others. The use of a similar range of directives
features of models of ‘grooming,’ particularly overcoming has been seen within interactions of groups who engage in sexual
external inhibitors (Finkelhor, 1984), grooming the environment offending in the physical world (Woodhams et al., 2012). Other
and others (Craven et al., 2006), and preventing discovery acts of dominance included challenging and/or disapproving of
(Sullivan, 2009). other others and their actions.
Despite existing studies citing the employment of usernames Within the textual data, there was evidence of suspects
associated with CSEA by online child sexual offenders making pro-child sexual abuse statements, normalizing and
(O’Connell, 2003), the majority of suspects’ usernames in minimizing the harm caused to children as a result of sexual
our sample were not related to a sexual interest in children/CSEA abuse experiences, as well as referring to children as sexual
more broadly. Whether these differences between studies stem
7
from variations in sample composition (e.g., heightened security Lanning’s (2010) publication is based on a database of cases that he has collated
during his 27 years in practice. The exact size of the database is unclear with him
awareness in our sample given their use of the Dark Web), or
referring to it as containing “thousands” of cases (p. 3). While he comments on
a change over time in choices regarding usernames (possibly in a relationship between gender preference and age of child, no statistical output is
response to law enforcement activity), is unclear. provided.

beings. These assertions about the sexual abuse of children are CONCLUSION
similar to those observed by O’Halloran and Quayle (2010)
in their study of interactions among users on a ‘boy-lover’ The use of the Dark Web for purposes related to CSEA
website on the Surface Web. In addition, there were almost is growing; the study presented here is therefore timely
as many suspects who referred to children as objects whose in its investigation of the characteristics and behaviors of
purpose was to serve the sexual needs of others, and, in a subset of individuals who operate on platforms on the
some cases, as being ‘deserving’ of the abuse. Such statements Dark Web that are geared toward CSEA. Our findings
are not surprising given the sadistic nature of some of the are relevant to law enforcement efforts in combating such
content presented within the forums, as well as the sexual offending behavior (e.g., by informing their training and
interests of some of the suspects in our sample, who are likely methods of investigation), providing the first description of
characterized by entitlement, and a lack of empathy and remorse a subset of individuals, more of whom will be entering
(Beech et al., 2005; Mokros et al., 2011). These assertions and correctional and rehabilitative services, and coming onto
justifications will be discussed in more detail as part of a caseloads of forensic psychologists and other practitioners in
separate article. the future.
Limitations
While this paper makes an original contribution to the literature DATA AVAILABILITY STATEMENT
on Internet child sexual offending, and the less well-known
features of those suspects who operate on platforms on the Dark The datasets generated for this study are not readily
Web, it has a number of limitations that need to be highlighted available because the data are owned by the police
in order to inform the design of future research studies. While force who collaborated on the study. Any access to the
an advantage of our data is that they were naturally-occurring, data would require their explicit permission. Requests
one limitation is that we merely have coded information for to access the datasets should be directed to JW,
suspects/cases where the relevant features were present. There j.woodhams@bham.ac.uk.
is therefore no way of obtaining a complete dataset for all
suspects/cases across the entire coding scheme. Analyses of
some features (e.g., sexual interests) were therefore based on a ETHICS STATEMENT
small number of suspects/cases, and should be interpreted with
caution. By sheer nature of using naturally-occurring data, our The research project was granted full ethical approval by the
data are less likely to be distorted by impression management Science Technology Engineering and Mathematics Ethics
that is often observed in qualitative interviews with apprehended Committee at the University of Birmingham (ERN_14-
offenders, whose future outcomes are likely affected by the way 1435E) and the Psychology Research Ethics Committee
they present themselves. However, it is also important to note at the University of Bath. Written informed consent
that impression management is very relevant within Dark Web from the participants was not required to participate in
communities, where it would influence one’s standing within this study in accordance with national legislation and the
the group, and the likelihood of receiving new material from institutional requirements.
other members. Our dataset comprised of a small sample of 53
individuals who had come to the attention of law enforcement
due to intelligence that suggested that they were engaging AUTHOR CONTRIBUTIONS
in the sexual abuse of children. It is therefore unlikely that
our sample is representative of all suspects and/or users who JW and CH-G designed the original study. JW, CH-G, and
present with a sexual interest in children, and are operating on BJ sought and secured multiple funding streams for the
platforms on the Dark Web. This selection criterion may well project. JW and CH-G sought and obtained ethical approval
have introduced bias into our findings, and would therefore for the study. They supervised the researchers employed on
benefit from further exploration as part of future empirical the project (including JK). JK was the primary developer
research. It is important for such research studies to assess the of the coding scheme which forms the basis of the results.
generalizability of our findings with other samples, including The development of the coding scheme was also informed
those derived from platforms with a different focus. Finally, by and refined following contributions from CH-G and JW.
inter-rater reliability was assessed using 10% from each of the JK conducted the coding of the 53 cases. BJ conducted
first five cases included in the dataset, once the coding of research to provide outcome data for the study. JW assessed
these had been completed. This was to ensure that the coding the inter-rater reliability of the coding, and conducted the
scheme could be used reliably before coding was continued. Inter- statistical analyses of the data reported in this article. JK
rater reliability was therefore assessed early on in the analytic produced the Method section. JW is the primary author of
process, rather than once all cases had been analyzed, with the this article. JK and CH-G also contributed to the writing of
coding scheme evolving and being further developed by the the present article. BJ reviewed and provided feedback on the
second author as they progressed through the analysis of the article. All authors contributed to the article and approved the
remaining cases. submitted version.

FUNDING funded by an award from the Police Knowledge

Fund.
This research project was funded by an ESRC Impact
Acceleration award, a Higher Education Innovation
award, and an EPSRC Knowledge Transfer Scheme ACKNOWLEDGMENTS
award from the University of Birmingham, an EPSRC
Impact Acceleration Award from the University of Bath, The authors would like to thank the unnamed police force who
and with funding from the unnamed United Kingdom collaborated with them in the production and delivery of this
police force. Later stages of the research project were research project.
REFERENCES Grant, T. D., and MacLeod, N. (2018). Resources and constraints in linguistic
identity performance: A theory of authorship. Lang. Law 5, 80–96.
Aslan, D., and Edelmann, R. (2014). Demographic and offence characteristics: Guest, G., MacQueen, K. M., and Namey, E. E. (2012). Applied thematic analysis.
A comparison of sex offenders convicted of possessing indecent images of Thousand Oaks, CA: Sage Publications, Inc.
children, committing contact sex offences or both offences. J. Forens. Psychiatr. Hamilton-Giachritsis, C. E., Hanson, E., Whittle, H., Alves-Costa, F., and Beech,
Psychol. 25, 121–134. doi: 10.1080/14789949.2014.884618 A. R. (2020). Technology-assisted child sexual abuse in the UK: Young
Babchishin, K. M., Hanson, R. K., and Hermann, C. A. (2011). The characteristics people’s views on the impact of online sexual abuse. Children Youth Serv. Rev.
of online sex offenders: A meta-analysis. Sexual Abuse J. Res. Treat. 23, 92–123. 119:105451. doi: 10.1016/j.childyouth.2020.105451
doi: 10.1177/1079063210370708 Hamilton-Giachritsis, C. E., Hanson, E., Whittle, H., and Beech, A. R. (2017).
Balfe, M., Gallagher, B., Masson, H., Balfe, S., Brugha, R., and Hackett, S. (2015). Impact of online and offline child sexual abuse: ‘Everyone deserves to be happy
Internet child sex offenders’ concerns about online security and their use of and safe.’. London: NSPCC.
identity protection technologies: a review. Child Abuse Rev. 24, 427–439. doi: Hartmann, D. P. (1977). Considerations in the choice of interobserver reliability
10.1002/car.2308 estimates. J. Appl. Behav. Anal. 10, 103–116. doi: 10.1901/jaba.1977.10-103
Beech, A., Fisher, D., and Ward, T. (2005). Sexual murderers’ implicit theories. Holt, T. J., Blevins, K. R., and Burkert, N. (2010). Considering the pedophile
J. Interpers. Violen. 20, 1366–1389. doi: 10.1177/0886260505278712 subculture online. Sexual Abuse J. Res. Treat. 22, 3–24. doi: 10.1177/
Beier, K. M., Neutze, J., Mundt, I. A., Ahlers, C. J., Goecker, D., Konrad, 1079063209344979
A., et al. (2009). Encouraging self-identified pedophiles and hebephiles to Home Office, (2019). Factsheet: Child Sexual Exploitation and Abuse. London:
seek professional help: First results of the Prevention Project Dunkelfeld Home Office.
(PPD). Child Abuse Neglect 33, 545–549. doi: 10.1016/j.chiabu.2009. Huang, H. Y., and Bashir, M. (2016). The Onion Router: Understanding a privacy
04.002 enhancing technology community. Proc. Assoc. Informat. Sci. Technol. 53, 1–10.
Bennell, C., and Canter, D. (2002). Linking commercial burglaries by modus doi: 10.1002/pra2.2016.14505301034
operandi: Tests using regression and ROC analysis. Sci. Justice 42, 153–164. Kloess, J. A., Beech, A. R., and Harkins, L. (2014). Online child sexual exploitation:
doi: 10.1016/S1355-0306(02)71820-0 Prevalence, process and offender characteristics. Trauma Violen. Abuse 15,
Briggs, P., Simon, W. T., and Simonsen, S. (2011). An exploratory study of Internet- 126–139. doi: 10.1177/1524838013511543
initiated sexual offenses and the chat room sex offender: Has the Internet Kloess, J. A., Hamilton-Giachritsis, C. E., and Beech, A. R. (2019). Offense
enabled a new typology of sex offender? Sexual Abuse J. Res. Treat. 23, 72–91. processes of online sexual grooming and abuse of children via Internet
doi: 10.1177/1079063210384275 communication platforms. Sexual Abuse J. Res. Treat. 31, 73–96. doi: 10.1177/
Chiang, E. (2019). Rhetorical moves and identity performance in online child sexual 1079063217720927
abuse interactions. Ph. D. thesis, Birmingham: Aston University. Kloess, J. A., Seymour-Smith, S., Hamilton-Giachritsis, C. E., Long, M. L., Shipley,
Clevenger, S. L., Navarro, J. N., and Jasinski, J. L. (2016). A matter of low self- D., and Beech, A. R. (2015). A qualitative analysis of offenders’ modus operandi
control? Exploring differences between child pornography possessors and child in sexually exploitative interactions with children online. Sexual Abuse J. Res.
pornography producers/distributers using Self-Control Theory. Sexual Abuse J. Treat. 29, 563–591. doi: 10.1177/1079063215612442
Res. Treat. 28, 555–571. doi: 10.1177/1079063214557173 Kokolaki, E., Daskalaki, E., Psaroudaki, K., Christodoulaki, M., and Fragopoulou,
Craven, S., Brown, S., and Gilchrist, E. (2006). Sexual grooming of children: Review P. (2020). Investigating the dynamics of illegal online activity: The power of
of literature and theoretical considerations. J. Sexual Aggress. 12, 287–299. doi: reporting, dark web, and related legislation. Comp. Law Secur. Rev. 38:105440.
10.1080/13552600601069414 doi: 10.1016/j.clsr.2020.105440
Davidson, J., and Gottschalk, P. (2011). Characteristics of the Internet for criminal Landis, J. R., and Koch, G. G. (1977). The measurement of observer agreement for
child sexual abuse by online groomers. Crimin. Justice Stud. Crit. J. Crime Law categorical data. Biometrics 33, 159–174. doi: 10.2307/2529310
Soc. 24, 23–36. doi: 10.1080/1478601X.2011.544188 Lanning, K. V. (2010). Child molesters: A behavioral analysis. For professionals
Davidson, J., and Martellozzo, E. (2008). “Protecting children online: Towards a investigating the sexual exploitative of children. Virginia: NCMEC.
safer Internet,” in Sex as crime?, eds G. Letherby, K. Williams, P. Birch, and M. Lanning, K. V. (2012). “Cyber ‘pedophiles’: A behavioral perspective,” in Serial
Cain, (Cullompton: Willan Publishing), 338–355. offenders: Theory and practice, eds K. Borgeson, and K. Kuehnle, (Sudbury, MA:
Durkin, K. F. (1997). Misuse of the Internet by pedophiles: Implications for law Jones & Bartlett Learning), 71–87.
enforcement and probation practice. Feder. Probat. 61, 14–18. doi: 10.4135/ Leech, G. N. (1983). Principles of pragmatics. London: Longman.
9781452229454 Long, M. L., Alison, L. A., and McManus, M. (2012). Child pornography
Europol, (2014). Internet Organised Crime Threat Assessment (iOCTA) 2014. The and likelihood of contact abuse: a comparison between contact child sexual
Hague: Europol. offenders and noncontact offenders. Sex. Abuse 25, 370–395. doi: 10.1177/
Europol, (2015). Virtual Global Taskforce Child sexual exploitation environment 1079063212464398
scan, 2015. The Hague: Europol. Lovell, R., Luminais, M., Flannery, D. J., Overman, L., Huang, D., Walker, T., et al.
Europol, (2016). Internet Organised Crime Threat Assessment (iOCTA). The Hague: (2017). Offending patterns for serial sex offenders identified via the DNA testing
Europol. of previously unsubmitted sexual assault kits. J. Crimin. Justice 52, 68–78. doi:
Europol, (2020). Exploiting isolation: Offenders and victims of online child sexual 10.1016/j.jcrimjus.2017.08.002
abuse during the Covid-19 pandemic. The Hague: Europol. Ly, T., Dwyer, R. G., and Fedoroff, J. P. (2018). Characteristics and treatment
Finkelhor, D. (1984). Child sexual abuse: New theory and research. New York, NY: of internet child pornography offenders. Behav. Sci. Law 36, 216–234. doi:
Free Press. 10.1002/bsl.2340

Martellozzo, E. (2015). Policing online child sexual abuse. The British experience. offending behaviour. Aggress. Violent Behav. 30, 15–23. doi: 10.1016/j.avb.2016.
Eur. J. Policing Stud. 3, 32–52. 07.002
Mokros, A., Osterheider, M., Hucker, S. J., and Nitschke, J. (2011). Psychopathy and Sullivan, J. (2009). Professionals who sexually abuse the children with whom they
sexual sadism. Law Hum. Behav. 35, 188–199. doi: 10.1007/s10979-010-9221-9 work. Ph. D. thesis, Birmingham: University of Birmingham.
National Crime Agency, (2018). National Strategic Assessment of Serious and Web-Iq, (2018). Web-IQ newsletter. Amsterdam: Web-IQ.
Organised Crime 2018. London: National Crime Agency. Weimann, G. (2016). Going dark: Terrorism on the Dark Web. Stud. Conflict
Neutze, J., Grundmann, D., Scherner, G., and Beier, K. M. (2012). Undetected Terror. 39, 195–206. doi: 10.1080/1057610X.2015.1119546
and detected child sexual abuse and child pornography offenders. Int. J. Law WeProtect, (n.d.). Preventing and Tackling Child Sexual Exploitation and Abuse
Psychiatr. 35, 168–175. doi: 10.1016/j.ijlp.2012.02.004 (CSEA):A Model National Response. Abu Dhabi: WeProtect. Available online at:
O’Connell, R. (2003). A typology of child cybersexploitation and online grooming https://www.weprotect.org/wp-content/uploads/WePROTECT-2015-Model-
practices. Bristol: jisc. National-Response-1-page-1.pdf
O’Halloran, E., and Quayle, E. (2010). A content analysis of a “boy love” support Whittle, H. C., Hamilton-Giachritsis, C., and Beech, A. (2013a). Victims’ voices:
forum: Revisiting Durkin and Byrant. J. Sexual Aggress. 16, 71–85. doi: 10.1080/ The impact of online grooming and sexual abuse. Univers. J. Psychol. 1, 59–71.
13552600903395319 doi: 10.13189/ujp.2013.010206
Owen, G., and Savage, N. (2015). The Tor Dark Net. Waterloo: Centre for Whittle, H., Hamilton-Giachritsis, C., Beech, A., and Collings, G. (2013b). A review
International Governance Innovation. of online grooming: Characteristics and concerns. Aggress. Violent Behav. 18,
Quayle, E., and Taylor, M. (2001). Child seduction and self-representation on the 62–70. doi: 10.1016/j.avb.2012.09.003
Internet. CyberPsychol. Behav. 4, 597–608. doi: 10.1089/109493101753235197 Woodhams, J., Cooke, C., Harkins, L., and Da Silva, T. (2012). Leadership in
Quayle, E., and Taylor, M. (2002). Paedophiles, pornography and the Internet: multiple perpetrator stranger rape. J. Interpers. Violence 27, 728–752. doi: 10.
Assessment issues. Br. J. Soc. Work 32, 863–875. doi: 10.1093/bjsw/32.7.863 1177/0886260511423244
Quayle, E., and Taylor, M. (2003). Model of problematic Internet use in people Wortley, R., and Smallbone, S. (2012). Internet child pornography: Causes,
with a sexual interest in children. CyberPsychol. Behav. 6, 93–106. doi: 10.1089/ investigation and prevention. Oxford: Praeger.
109493103321168009
Quayle, E., Holland, G., Linehan, C., and Taylor, M. (2000). The Internet and Conflict of Interest: The authors declare that the research was conducted in the
offending behaviour: A case study. J. Sexual Aggress. 6, 78–96. doi: 10.1080/ absence of any commercial or financial relationships that could be construed as a
13552600008413311 potential conflict of interest.
Seigfried-Spellar, K. C. (2014). Distinguishing the viewers, downloaders,
and exchangers of Internet child pornography by individual differences: Copyright © 2021 Woodhams, Kloess, Jose and Hamilton-Giachritsis. This is an
Preliminary findings. Digit. Investig. 11, 252–260. doi: 10.1016/j.diin.2014. open-access article distributed under the terms of the Creative Commons Attribution
07.003 License (CC BY). The use, distribution or reproduction in other forums is permitted,
Seto, M. C. (2013). Internet sex offenders. Washington, D.C: American provided the original author(s) and the copyright owner(s) are credited and that the
Psychological Association. original publication in this journal is cited, in accordance with accepted academic
Shelton, J., Eakin, J., Hoffer, T., Muirhead, Y., and Owens, J. (2016). Online child practice. No use, distribution or reproduction is permitted which does not comply
sexual exploitation: An investigative analysis of offender characteristics and with these terms.

ORIGINAL RESEARCH
published: 12 April 2021
doi: 10.3389/fpsyg.2021.619036
Obstacles in the Process of Dealing

With Child Sexual Abuse–Reports
From Survivors Interviewed by the
Independent Inquiry Into Child Sexual
Abuse in Germany
Wiebke Schoon* and Peer Briken
Institute for Sex Research, Sexual Medicine and Forensic Psychiatry, University Medical Center Hamburg-Eppendorf (UKE),
Hamburg, Germany
Obstacles in dealing with child sexual abuse (CSA) can hinder survivors in the process of
coming to terms with their experiences. The present study aims to identify and analyze
factors that may pose obstacles in the long-term process of dealing with CSA. It is part
of a larger research consortium “Auf-Wirkung,” funded by the German Federal Ministry
Edited by:
Noora Ellonen,
of Education and Research, and was conducted in cooperation with the Independent
University of Tampere, Finland Inquiry into Child Sexual Abuse in Germany (IICSAG). The IICSAG was appointed by
Reviewed by: the Independent Commissioner for Child Sexual Abuse Issues and the German Federal
Ulrich Schweiger,
Ministry for Family Affairs, Senior Citizens, Women, and Youth in 2016. To determine
Helios Hanseklinikum, Germany
Kirsi Peltonen, responsibilities, recognize injustice, and further acknowledge the survivors of CSA in
University of Turku, Finland the Federal Republic of Germany (FRG) and the German Democratic Republic (GDR),
*Correspondence: the Independent Inquiry has held 1,303 private sessions with survivors of CSA by Oct.
Wiebke Schoon
w.schoon@uke.de
17th, 2020. The present study focuses on exploring reoccurring problematic experiences
reported by survivors in private sessions regarding the long-term process of dealing with
Specialty section: experiences of CSA. A total of 30 transcripts of private sessions, conducted by members
and appointees of the IICSAG between September 2016 and June 2019, were analyzed
a section of the journal using qualitative content analysis. Attendants of private sessions described a variety of
obstacles, including negative social reactions to disclosure, institutions’ unwillingness to
Received: 20 October 2020 elucidate occurrences of CSA within their midst, as well as general financial difficulties,
Accepted: 10 March 2021
Published: 12 April 2021 and those linked to redress claims. Manipulative grooming by perpetrators and limited
Citation: access to adequate psychotherapy were perceived as obstructive by survivors dealing
Schoon W and Briken P (2021) with CSA. In the context of criminal proceedings, survivors reported long durations of
Obstacles in the Process of Dealing
With Child Sexual Abuse–Reports
court proceedings and negative experiences in connection to credibility assessment.
From Survivors Interviewed by the Results will be discussed to better support survivors of CSA in the process of dealing
Independent Inquiry Into Child Sexual with their experiences in the future.
Abuse in Germany.
Front. Psychol. 12:619036. Keywords: child sexual abuse, dealing with abuse, public inquiries, Independent Inquiry into Child Sexual Abuse,
doi: 10.3389/fpsyg.2021.619036 qualitative research and analysis

Schoon and Briken Obstacles in Dealing With CSA
DEALING WITH EXPERIENCES OF CSA OBSTACLES IN DEALING WITH

EXPERIENCES OF CSA
Dealing with experiences of CSA is a continuous and lengthy
process for survivors that often accompanies them throughout Obstacles in dealing with experiences of CSA are often related
their lives and during which many face a variety of challenges to factors or experiences which aggravate the symptoms and
and obstacles. Researchers have proposed a variety of empirical effects of the traumatic experiences, which, according to
and theoretical models to describe and conceptualize the process the Traumagenic Dynamics Model of Child Sexual Abuse
of dealing with experiences of sexualized violence, with many are traumatic sexualization, betrayal, stigmatization, and
postulating 3 or 5 different phases in which coping can powerlessness (Finkelhor and Browne, 1985). Past research
occur (Horowitz, 1986; Kleber and Brom, 1992; Roth and was able to identify some evidence for obstacles in the process
Newman, 1993; Figley, 2013; Fischer and Riedesser, 2016). Many of dealing with CSA. For example, the way the social peer
models see the integration of an incomprehensible experience group or environment reacts to the disclosure of experiences of
into one’s understanding of self and the world as the most sexualized violence is important to the process of dealing with
meaningful part in the process of dealing with CSA. According CSA. Research on the topic suggests that disclosure might be
to Figley (2013), however, “adaptation” (the last stage) can the first step in the healing process even for some individuals to
only be achieved if adequate resources (personal, social, and regain control and to pave the way for meaning and integration
financial) are available. Herman (1994) agrees and claims that of the abuse experience (Birck, 2001; Chouliara et al., 2014).
restoring a sense of security is of fundamental importance Active or passive inhibition of disclosure, therefore, poses a
to survivors, including pragmatic aspects, like ensuring basic potential obstacle in dealing with CSA. Research consistently
needs such as financial security. For Fischer and Riedesser suggests that negative reactions to disclose attempts have a
(2016) the social dimension of dealing with CSA is in the lasting adverse impact on the coping process of survivors
foreground since the authors assume that traumatic experiences (Birck, 2001; Filipas and Ullman, 2001; Ullman et al., 2007;
cannot be dealt with by an individual alone. Many aspects Ullman and Peter-Hagene, 2014). Thus, negative social reactions
of the described models were also picked up and further from a trusted or formal source are associated with greater
developed by Gahleitner (2003). Like Fischer and Riedesser PTSD symptoms, the relationship is mediated by maladaptive
(2016) the author emphasizes the processual nature of dealing coping (Ullman and Peter-Hagene, 2014), and it reduces the
with CSA in a three-phase model. Similar to Figley (2013) likelihood that survivors will seek support (Birck, 2001). Wyatt
she concludes that the integration of the traumatic experiences and Mickey (1987) indicate, that a negative immediate and
into the self-concept may only succeed with a minimum long-term reaction to disclosure of CSA may mediate adverse
of relative security and support of at least one sustainable effects of the initial abuse in survivors. This may be since a
relationship. This conclusion is further supported by Draucker person or group not believing them can be seen as a betrayal
et al. (2011). Their model includes four stages (grappling with and increase the feeling of powerlessness associated with the
the meaning, figuring out the meaning of CSA, tackling the abuse itself (Finkelhor, 1987). Consequently, negative reactions
effects of the CSA, laying claim to one’s life), five domains to disclosure undermine survivors’ attempts to regain control
of functioning (life patterns, parenting, disclosure of CSA, and act in a self-effective way. Furthermore, social support, in
spirituality, altruism) and six enabling factors to progress general, seems important – short-term and long-term – as it may
from one stage to the next. These six were contextual factors mitigate negative outcomes and lower symptomology (Murthi
like receiving affirmative messages, having ongoing support, and Espelage, 2005).
and experiencing a critical life event and personal factors Manipulative perpetrator behavior also called grooming, may
like personal agency, personal resolves, and commitment to similarly hinder the process of dealing with CSA. Wolf and Pruitt
transcend from CSA. (2019) examined the effects of grooming, namely verbal coercion,
In summary, giving meaning, regaining control, and being grooming that used drugs/alcohol, as well as threatening/violent
able to integrate the abuse is most relevant for survivors of grooming, and found that in a linear regression model grooming
sexualized violence (Horowitz, 1986; Kleber and Brom, 1992; categories predicted trauma symptom severity, with threatening
Frazier et al., 2004; Walsh et al., 2010; Draucker et al., 2011). or violent tactics having the most severe effect on survivors
Conversely, the absence of enabling factors as well as any psychological well-being. The authors conclude that the isolation
form of social and societal forms of support may hinder and normalization of sexually abusive interactions may aggravate
these relevant goals and therefore the process of dealing the trauma and therefore hinder the healing process. These
with CSA (Birck, 2001; Gahleitner, 2003; Draucker et al., findings were supported by Chouliara et al. (2014). The authors
2011). interviewed 22 adult survivors of CSA using qualitative analysis
and identified four relevant themes, amongst others “Factors
hindering Recovery” and “Hurdles of Recovery.” Survivors
reported that they felt insufficiently supported by family
Abbreviations: IICSAG, Independent Inquiry into Child Sexual Abuse in
Germany; CSA, Child Sexual Abuse; FRG, Federal German Republic; GDR, members, who oftentimes refused to address the abuse further
German Democratic Republic; OEG, Opferentschädigungsgesetz (victims’ when confronted. Furthermore, survivors feared stigmatization,
compensation law). especially regarding issues of mental health. Social inviolability

of the perpetrator due to his or her position enabled longer abuse form of a private session or a written report. Private sessions,
and prevented exposure or disclosure. Survivors reported their with a total of 1,303 by Oct. 17th, 2020, were held in 12 different
credibility being called into question as well as grooming and major cities by members of IICSAG and qualified representatives
other manipulative perpetrator behavior. of the Inquiry, providing a safe space for survivors to share
Adequate psychotherapy may reduce psychological distress their experiences. Interviewers used a set of semi-structured
caused by experiences of CSA and improve the overall guidelines, containing 10 topics related to the subject, including
functioning and well-being of survivors (Price et al., 2001; immediate and long-term consequences of CSA, disclosure,
Sánchez-Meca et al., 2011). Unfortunately, many survivors seeking support, and experiences with religious and government
report a lack of access to appropriate psychological treatment institutions (see Supplementary Material). The interview
options in Germany with the average waiting time for a guidelines served primarily as a narrative-generating aid for the
place in a psychotherapy program being three months in interviewers, intending to encourage and enable survivors to
larger German cities and six months in more rural areas shape the pace and direction of the conversation and report as
(Pawils et al., 2017; Unabhängige Kommission zur Aufarbeitung freely as possible on their experiences.
sexuellen Kindesmissbrauchs, 2019). Additionally, Sommer This project is funded by the German Federal Ministry
(2016) illustrated the situation for individuals with complex of Education and Research and is part of the larger research
PTSD in Germany, showing that programs and therapies often consortium “Auf-Wirkung,” which aims to accomplish a
fail to successfully stabilize patients, mostly due to a lack of comprehensive and extensive investigation of structural
expertise in professionals treating survivors with these specific conditions in connection with sexualized violence against
needs. The lack of access to adequate and efficient psychotherapy children and adolescents. The current study presents the
can therefore be seen as a factor hindering the process of dealing first results of sub-project three, which examines recurring
with CSA. experiences of survivors, which may hinder the process of
A factor whose influence on processing experiences of CSA dealing with experiences of CSA.
has not been investigated more thoroughly are obstacles in the
context of criminal investigations (Walsh et al., 2010; Görgen Data Selection
et al., 2012; Volbert, 2012). Some research indicates, that long- A set of 100 summaries of transcripts, selected by the Inquiry
duration of court proceedings, as well as the anticipation to according to general thematic overlap, were made available to the
make a statement in front of a court (in presence of the accused project’s research assistants. Each sub-project of “Auf-Wirkung”
perpetrator), may lead to a short-term negative influence on well- received 25 summaries which were then sorted according to
being and slower recovery of the damage caused by the crime descriptive as well as key data points and organized using excel,
(Runyan et al., 1988; Volbert, 2012). Furthermore, doubts about providing a basis for further selection. From this, 30 transcripts
testimony or the severity of the abuse impact can be challenging were chosen according to thematic overlap with the specific
for survivors, especially if there is no other “objective” evidence research questions at hand. The selected material, therefore,
(Volbert, 2012). Stressful as court proceedings may be, going represents a nonprobability “convenience sampling,” meaning
through with it may be associated with increased self-efficacy and data was included due to availability (Robinson, 2014). Before
restoration of control long-term (Volbert, 2012). data analysis, the research consortium “Auf-Wirkung” and each
The present study aims to identify and analyze different sub-project were approved by the Ethics Committee of the
obstacles in the process of dealing with CSA. It explores these Department of Education at Goethe University Frankfurt.
reoccurring problematic experiences reported by adult survivors
in the context of private sessions held by the IICSAG using openly Data Analysis
designed interviews to encourage spontaneous statements. In A qualitative content analysis of 30 transcripts was conducted
this study, we focus on the analysis of contextual obstacles that following Kuckartz (2018) using the qualitative data analysis
survivors report in interaction with others (persons, state actors, software MAXQDA version 18.2.4 (www.maxqda.de/). Included
institutions). The goal is to apply the findings to better support private sessions were held by members of the IICSAG between
survivors in dealing with experiences of CSA in the future. September 2016 and August 2019, transcribed verbatim by a
collaborating company, and made available by the Inquiry’s
MATERIALS AND METHODS corresponding office.
The analytic procedure of the content structuring
Study Design content analysis (Kuckartz, 2018) compromises a seven-
This study was conducted in cooperation with the Independent step deductive/inductive process. Firstly, the entire material
Inquiry into Child Sexual Abuse in Germany (https://www. was read by two researchers separately, significant text passages
aufarbeitungskommission.de/english/). The Inquiry was were marked, and thoughts were recorded in memos. Data
appointed by the Independent Commissioner for Child Sexual was then structured by developing thematic main categories,
Abuse Issues and the German Federal Ministry and consists of based on research questions on the one hand (deductive) and
seven volunteer members of different subject-related fields and on topics that emerged directly from the material (inductive)
professions. Starting in January 2016 the Independent Inquiry on the other (Kuckartz, 2018, p. 101–102). Using these main
called on survivors and contemporary witnesses of CSA in the categories, the material was then coded sequentially by the
GDR or FRG to contact the Inquiry staff to tell their story in two researchers independently, whereby multiple coding of

individual text passages was possible (Kuckartz, 2018). In the TABLE 1 | Descriptive data: demographics and contexts of abuse (N = 30).
next step, sub-categories were determined using inductive P
n/M %/SD = 67
category development (Kuckartz, 2018, p. 72–86). For this, all
text passages coded with the same main category were compiled Demographics
and arranged systematically to identify recurring aspects and Male (n, %) 8 26,7
themes relevant to survivors in the process of dealing with Female (n, %) 22 73,3
CSA which had not been considered previously. Additionally, Age in years
definitions for all main and sub-categories were formulated M (SD) 54,6 (11,0)
for better comprehensibility in the further course of analysis Missing (n,%) 1 0,03
(Kuckartz, 2018). Using the resulting category system, the Abuse experience
entire material was once again coded. After 30% of the material
Repeated/Multiple (n, %) 26 86,7
was coded this way, the categories were re-evaluated, revised,
Once (n, %) 4 13,3
and definitions concretized, if necessary until the categories
Germany
adequately reflected the data. The remaining material was then
FRG (n, %) 2 6,7
coded. Finally, the coded material was analyzed along with the
GDR (n, %) 10 33,3
main categories, giving an overview of all themes that emerged
Both (n, %) 1 3,3
during the conversations regarding potential barriers in dealing
Missing (n, %) 17 56,7
with CSA (Kuckartz, 2018, p. 118). Furthermore, we explored
Contexts of abuse*:
relationships of subcategories within the main category as well
Family 17
as relationships between the main categories by focusing on the
Religious: 13
proximity of certain subcategories and themes within and across
Catholic 8
main categories (Kuckartz, 2018, p. 118–119). By exploring
Protestant 4
sub-categories that were often mentioned simultaneously or
in proximity, we aimed to identify more complex associations Jehovah’s Witnesses 1
between relevant themes. Federal Institutions 21

Recreational Activities 3
n.a., not applicable; M, mean value; SD, standard; *multiple entries possible.
RESULTS
Sample and Data TABLE 2 | Main- and subcategories of contextual obstacles in dealing with CSA.
The present study included 30 transcripts of private sessions held
by the Independent Inquiry into Child Sexual Abuse in Germany Main categories Subcategories
with survivors, who report having experienced CSA in the FRG
1. Disclosure • Negative social reactions to disclosure
or the GDR between the 1950s and early 2000s. Table 1 displays • Negative consequences of disclosure
socio-demographic and abuse context-related information. Most • Factors enabling disclosure
participants were female, 54 years on average, and half of them • Factors hindering disclosure
had children. At the time of the first abuse experience, survivors 2. How others deal with CSA • How the family of origin deals with CSA
were between 7 and 13 years old on average, with one person • How the institution deals with CSA
• How society deals with CSA
reporting sexual abuse before age of one.
All 30 survivors reported long-term effects on their physical 3. Grooming • Grooming the victim
• Grooming the environment
or mental health in some way or form. Pathological somatic
4. Obstacles in state legal and • Obstacles in the mental health care
phenomena were described most frequently including sleep mental health care system system
disorders, pain disorders (e.g., migraine), diseases of the • Obstacles in criminal and civil law
gastrointestinal tract, and obesity. Reports of mental health issues • Financial and organizational hurdles
ranged from precise diagnoses, ascribed by survivors themselves
or by clinical professionals, to implicit descriptions of general
psychological discontentment. Substance abuse, anxiety, PTSD,
and depression were reported frequently. Eating disorders and grooming, and obstacles in state legal and health care systems (see
self-harming behavior were reported as direct or long-term Table 2).
consequences of the abuse. Suicidal thoughts or suicide attempts
Disclosure
were reported by 19 out of 30 survivors and arose either while the
Survivors described disclosing their experiences as a turning
abuse was still ongoing or as a long-term consequence within the
point in the process of dealing with CSA, either positively or
coping process.
negatively. Many reported a variety of immediate negative social
reactions to their disclosure attempts, which then influenced
Main Categories how they continued to deal with their abuse experiences. Thus,
Contextual obstacles in dealing with experiences of CSA have inadequate responses to disclosure affected further disclosure
been organized around disclosure, how others deal with CSA, attempts and therefore survivors’ long-term coping process in

general. Survivors especially suffered when a trusted person did Survivors reported in-depth how institutions dealt with
not believe them. CSA. From their perspective, several insufficiencies regarding
actions not taken by institutions occurred. In some cases,
“I then confronted her [. . . ], and asked whether she knew about neither any form of recognition nor adequate compensation was
it. Until the end she said that I imagined it and that I only ever pursued by the institution concerned. Admissions of guilt were
caused her trouble, anyway.” seldomly made. Survivors reported a general lack of transparent
communication about misconduct on the part of the institution,
Besides immediate inadequate social reactions some survivors as well as perspectives on how future protection can be secured.
experienced continuing and longer-lasting negative (social) Moreover, survivors frequently described that they did not know
consequences of disclosure for example resentment and hostility whom they could have turned to for help after the experience of
from their family or social peers. Some even connected abuse because corresponding contact points either did not exist
this continued distress to an aggravation of psychological or were not sufficiently declared. Lastly, some survivors criticized
symptoms. a general lack of debate around the issue of CSA in institutions.
From the reports of the participating survivors, some factors Survivors reported that some religious and state institutions
enabling disclosure emerged, including a change in family installed structures that allowed for internal handling of
dynamics, seeking support from a medical or psychological allegations of abuse. It is reported how these structures were
professional, networking with other survivors, increasing media authorized by the institutions and in some catholic church
coverage of the topic, or the death of the perpetrator. The institutions even replaced state or secular criminal prosecution.
reports furthermore contained descriptions of factors hindering Within these structures, specially defined values and legal
disclosure, including an existing difficulty or inability to put concepts prevailed and determined the respective handling of the
the experience into words, fear of destabilizing the family or abuse allegations, which often resulted in the protection of the
social environment. perpetrator and the moral conviction of the survivors. According
Or as one participant put it: to survivors one goal behind this procedure was to protect the
reputation of the institution at any cost.
“About the family climate, I would like to tell you something Furthermore, survivors reported no access to appropriate sex
briefly. In our family, the language of violence prevailed. All that education in childhood or adolescence and that this hindered
was left was. . . silence.” them from disclosing their experiences. This lack of knowledge
made it more difficult for survivors to recognize the abuse as an
How Others Deal With CSA injustice at the time, which in turn reduced effort or ability to turn
The second theme survivors addressed was how others, more to other adults in search of help.
specifically the family of origin, state or church institutions, and A special mention should be made about religious institutions.
society, deal with experiences of CSA in their midst. Survivors Some survivors denounced the taboo approach toward sexuality
reported various negative ways in which the family of origin and sex education of catholic institutions. Survivors suspected
dealt with CSA. For example, family members ignored specific that the taboos around sexuality contributed to a further
changes in the behavior of survivors, which the victims attributed internalization of shame and guilt regarding the experience of
to the abuse. Repressions, reinterpretations of events, or even CSA in religious institutions. Part of this was a demonization of
punishments were also described. Some survivors described sexuality (outside the institutionalized framework of “marriage”),
being blamed for the abuse by family relatives. One participant punishment systems, and corresponding induction of guilt
described this incident: were described.
Survivors described various barriers regarding how society
“And then my sister-in-law came in and saw him abusing me, [. . . ] deals with CSA. Due to a general lack of visibility of the topic
And then [. . . ] she didn’t offer help, but instead she threw him out within the public eye, survivors of CSA reported feeling that
of the apartment and accused me. . . I seduced him, so to speak.” their experiences and interests are insignificant and irrelevant
to other members of society. This lack of recognition and the
These descriptions often were connected to reports of family feeling of being overlooked is amplified by concerns about
members trying to protect the perpetrator and his/her reputation, social stigmatization.
legacy, or career by instructing the victim not to talk about the
abuse in front of other family members or friends. “I always had the feeling that I had a stamp on my forehead
that said: sorted out. And it took me a very long time to
“And I was instilled, ‘You don’t talk about it, it’s something very, somehow develop a certain self-confidence where I said, ‘This is
very bad, something unbelievable that God doesn’t want.’ And interesting. . . this is nobody’s business’.”
you don’t destroy the future for the ∗ (perpetrator) either. And you
just don’t tell your grandparents.’” Survivors described tangible situations in which they felt
discriminated against and recounted prejudiced encounters as
As seen in the quote above, survivors reported being shamed by well as generally insensitive behavior toward them by employees
relatives – sometimes by using god as a means to impose guilt of government service agencies. Another aspect of this experience
on survivors. was that survivors often had their credibility questioned – either

in an official or social context, which consequently led survivors In the aftermath of the abuse as well as during, some survivors
to form the impression that nobody believes them in general. reported that perpetrators made them take an oath of secrecy to
prevent disclosure or exposure. Besides this, the use of blackmail,
Grooming threats, and other forms of violent manipulation was reported.
Another theme that emerged during the analysis of the reports
was manipulative perpetrator behavior–also known as grooming. “He said that something would happen to me if I talked. And I
Passages were coded only when survivors themselves described believed him.”
the perpetrator’s behavior (subsequently) as manipulative. This
included all behavior that was initiated by perpetrators to enable Bribery through gifts as “rewards” were also used in some
and facilitate abuse as well as to prevent exposure and was aimed instances. There are also reports of targeted induction of shame
at the child concerned or its environment. The subcategories, and guilt, which gave survivors the feeling that they are partially
therefore, are “Grooming the Victim” and “Grooming the or fully responsible for the abuse because they did not signal their
Environment,” a distinction also used by Craven et al. (2006). unwillingness earlier. Some perpetrators insisted that survivors
In grooming the victim several survivors reported experiences were the initiators of the abuse in the first place.
of manipulative perpetrator behavior before, during, or in the As well as this, the threat of punishment and consequences
aftermath of the abuse. Survivors described violent and non- for the perpetrators in some cases led to the cover-up of the
violent types of grooming that appear to belong to a set of abuse, especially in case of partly positive emotional attachment.
reoccurring strategies that perpetrators used in different phases Additionally, perpetrators justified their behavior in religious
of the abuse and with various goals. abuse contexts by shaming victims and making them believe that
Before the abuse, some survivors described being “selected” by perpetrators were only acting on or fulfilling “Gods will”
a perpetrator, for example by exploiting their vulnerabilities and
emotional needs. “And he always justified the whole thing by saying that God sees
everything but can’t punish everything, and that he was sort of
“In retrospect, [. . . ] I think he, the ∗ (perpetrator), was always made to do it, and that he just had to do it so that we wouldn’t
selecting, which boys were suitable for later sexual abuse. I’ll say become evil people.”
now, that he probably noticed from the start: ‘Oh, the ∗ (survivor),
he is probably predestined or suitable.’”
One survivor was told that she was chosen by a higher power.
Furthermore, perpetrators were described to deliberately form
“Like, yeah, that I am chosen and will be liberated through
an interpersonal bond with children accompanying the abuse in
suffering. And. . . through the pain I become free from sin and
some cases. Strategies described to strengthen this relationship stuff, right? And they contribute. . . They contribute to that.
were “payment” in form of money, gifts, or affection, treating Right.”
the victims like “adults” and making alcohol, drugs, and
pornographic material available. Another means used to enhance
Apart from grooming the victim, survivors described instances of
the relationship between perpetrator and survivor was the
perpetrators grooming the environment by integrating themselves
gradual isolation of the survivor from other close relatives
into the social environment of the survivor or deploying their
or peers. This targeted attention led some survivors to feel a
societal status (e.g., chaplain, teacher, city mayor) purposefully.
calculated increase in self-worth.
This sometimes led to family members protecting the perpetrator
willingly or not believing the victim as illustrated in the cross-
“So he kind of made sure that I was the predestined boy. And that
category analysis.
also made me feel like I was special. So it was actually pretty. . .
pretty ingenious, that system.”
Obstacles in the State Legal System and Mental
Health Care
During periods of abuse, some survivors reported being
Finally, topics that emerged from the transcripts were hurdles
intoxicated by perpetrators to facilitate the abuse or establish
and obstacles regarding the state legal system and mental health
compliance. Furthermore, some perpetrators normalized the
care. Many survivors reported obstacles in the mental health care
abusive behavior to offer patterns of explanation that trivialized
system, especially regarding psychotherapy or other counseling
the abuse or made it seem without alternatives. Violent
services like self-help groups, and the youth welfare office.
manipulation was also reported. Some used emotional blackmail
Most commonly, survivors described experiencing a lack of
to put pressure on survivors.
expertise and/or sensitivity by professionals working in these
areas or institutions.
“[. . . ] he always told me, and that was the worst thing for me,
Many survivors reported a lack of access to adequate
that he pressured me so much that I had to have sex with him.
I understood what sex was because he loves me so much. And if psychotherapy within Germany. Furthermore, if they managed
I don’t do that, then he will take his own life, and then my sister to secure a spot, some mentioned being confronted with a lack of
will be left alone with the three kids, and I can’t be responsible for specific qualifications of the therapist concerned, which in some
that and. . . . I think that was one of the worst things for me, that I cases led to a misdiagnosis or therapy that didn’t fit a survivor’s
felt responsible and then I gave myself to it.” specific needs.

In addition, many participants described financial and Furthermore, survivors reported being questioned and
organizational hurdles, e.g., struggles with basic income, since assessed repeatedly (police, court, witness cred.) which was
many survivors were unable to work due to long-term experienced as extremely stressful, wearing, and led to feelings
consequences of CSA and therefore needed to seek long-term of re-traumatization in some cases.
financial support. In some cases, this then led to poverty, which, Additionally, survivors reported gaps and discontinuities in
due to insufficient pension payments, continued into old age and their memory being regarded as inconsistencies during witness
meant an additional psychological burden to these survivors. credibility assessment.
Many survivors sought financial support in the form of
victim compensation, covered by the victim’s compensation law “The problem is credibility. [. . . ] There are always gaps in the. . .
(Opferentschädigungsgesetz, OEG). Almost all survivors who story, right? [. . . ] There can’t be a complete picture. But that’s
applied for support from the OEG reported structural problems exactly why it’s not a lie, that’s exactly why. You just didn’t have
or organizational hurdles in the application process of the the happiest childhood.”
procedure. Amongst other things, survivors had to undergo a
variety of assessments and evaluations as part of the process. Aspects criticized associated with the person conducting
Survivors reported not being able to prove beyond a reasonable the procedure included lack of information regarding
doubt that their suffering is the result of experiences of CSA or the procedure, insinuations of therapy-induced memories,
even that the abuse took place (Plausibility Check). However, this stigmatization, and in some cases, evaluators appeared untrained
is a requirement to get support from the OEG. or behaved insensitively.
Other obstacles in dealing with CSA included administrative
resistance, bureaucratic hurdles, and a lack of information and “[. . . ] Then she said, we have to conduct an expert evaluation. And
I didn’t even realize at that moment that this was going to be a
support from governmental institutions and offices. Above all,
credibility assessment, right?”
this included reports on overcomplicated forms to be filled
out, impenetrable bureaucratic processes, lack of networking
Additionally, negative experiences during the credibility
of various official apparatus, missing information for survivors
assessment in some cases led to a general skepticism toward
on official websites, and a lack of support in finding the
various official procedures, which then affected other areas in the
appropriate information.
process of dealing with CSA.
Survivors reported costs that were not covered by health
An important part was also that survivors felt they did not
insurance and therefore had to be paid for out of pocket. This led
have any control over the process:
to high financial burdens in some cases and the discontinuation
of treatment, which then, in turn, hindered survivors in their
“And then, though, that loss of control, that’s just that again, that
further healing process. loss of control, knowing exactly, I can’t do anything, I’m standing
Obstacles in criminal and civil law included the duration there again, I can’t do anything. I have to surrender to it.”
of court proceedings, the statute of limitations, and witness
credibility assessment. Survivors reported lengthy criminal Overall obstacles in the state legal system and mental health care
proceedings, which were described as particularly exhausting and were reported to negatively impact the process of dealing with
demotivating. In some cases, due to the statute of limitations experiences of CSA for survivors.
of the offense(s), no report could be made, ongoing criminal
prosecutions were interrupted, or affected persons refrained from Cross-Category Analysis
filing a complaint at all. In addition to analyzing the main themes, we also examined
In terms of expert evaluations and assessments, survivors the relationships between the different categories to provide a
reported negative experiences in assessments conducted in the more contextualized and comprehensive picture of experiences
process of the application for the Victim Compensation Fund and factors posing as contextual obstacles in the process of
(Plausibility check) as well as with regards to witness credibility dealing with experiences of CSA. Doing this it became clear
assessments in criminal court cases. that most of the categories were interconnected and interacted
Regarding the latter, survivors’ criticism can be divided into in certain ways. For example, how the family dealt with CSA
two different categories: In the first, survivors reported negative was highly related to disclosure and specifically, factors hindering
experiences that they attributed to the procedure itself. In the disclosure and consequences of disclosure. Thus, many of the
second, those that they associated primarily with the person described negative behaviors by family members happened in
conducting the procedure, i.e., the assessor. Some experiences can the context of disclosure attempts, for example, the instructions
be classified into both categories. to either never talk about the abuse or stop talking about it in
Aspects criticized regarding the procedure itself include the the future.
lack of clarity and consequently misunderstanding of the use
of the term “null hypothesis” in credibility assessments, which “Then at some point she said, ‘Stop stirring up old stories.’ And,
describes the assumption, that the witness’s report is not based on ‘You need to lay these stories to rest now. . . ’”
genuine experiences. This hypothesis shall then be refuted in the
process. This gave survivors the impression that their testimonies There is also an overlap between how the family of origin deals
were not believed from the beginning. with CSA and grooming the environment since survivors on some

occasions reported family members being close to and friendly In their CSA healing model Draucker et al. (2011) postulate,
with the perpetrator or perpetrators blatantly and successfully that receiving at least one “affirmative message” from a trusted
discrediting the survivors’ credibility in front of family person or formal source can function as an important enabling
members. factor and therefore help survivors to progress in the process
of dealing with their experiences. These findings are supported
“And I wanted to clarify things, or at least understand them, and I by Ullman and Peter-Hagene (2014), who found that negative
caused a lot of turmoil in the family. And I was often told to stop social reactions to disclosure of abuse were related to greater
now and [. . . ] that I was the crazy one and that I should go to my PTSD symptoms, with survivors perceiving less control over
therapy and finally stop digging around in the old story.” their dealing process. Furthermore, Sivagurunathan et al. (2019)
identified feelings of shame, guilt, and self-blame, a higher
This also occurred in connection with how institutions deal with social standing of the perpetrator as well as negative social
CSA. Through a high social status or institutionalized authority reactions as obstacles affecting disclosure in male survivors
and the social prestige that goes along with it, perpetrators were of CSA. Our findings are also supported by Chouliara et al.
in some cases protected from any consequences. Oftentimes (2014) who recognized disclosure as a major factor in enabling
perpetrators were more likely to be believed than the survivor the healing process for survivors of CSA with negative social
and consequently acquitted of any guilt by the institution. reactions posing as an obstacle. Apart from being confronted
Furthermore, to protect the reputation of the institution, criminal with continuing resentment and disbelief, survivors reported
prosecution was sometimes waived. being shamed or humiliated by family relatives, who sometimes
Grooming the victim was also related to factors hindering supported the offender rather than the survivor. This implies
disclosure. By inducing shame and guilt, using violent threats, that survivors also relived stigmatization, betrayal, and also
or acquiring an oath of secrecy from survivors, perpetrators powerlessness in these situations (Finkelhor and Browne, 1985).
tried to prevent disclosure and exposure in some cases. The Another enabling factor identified by Draucker et al. (2011) is
internalization of feelings of shame and complicity in some cases “ongoing support,” which according to the authors goes beyond a
created an illusion of consensuality that perpetrators exploited. one-time affirmative message but describes the feeling of trusted
In summary, factors, and experiences that may hinder the process persons being there for the survivor regardless. Our results
of dealing with CSA are located on many levels and interact with are supported by a study by Schönbucher et al. (2014) which
each other in a complex way. indicates that adolescents who experienced CSA wish for more
support from their parents.
DISCUSSION Our data further indicate that state and religious institutions
frequently ignored survivors and didn’t support them sufficiently.
Using qualitative methods, the present study aimed to identify They seldomly took adequate action to elucidate cases of abuse
and analyze factors and experiences that may hinder the process in their midst or compensate survivors, which made survivors
of dealing with experiences of CSA. Four main themes which had feel powerless and betrayed (Finkelhor and Browne, 1985).
12 subthemes were analyzed in this study. Our analysis showed This is most likely connected to the power imbalance that often
that the categories are mostly interconnected and interact with prevailed and still in some cases prevails in some institutions,
each other on various levels. Survivors of CSA reported being summarized under the term clericalism by Dreßing et al. (2018)
confronted with a variety of obstacles, throughout the process for catholic church institutions. This describes a hierarchical-
of dealing with their experiences with some obstacles occurring authoritarian system that enabled priests in superior positions
simultaneously, repeatedly or in a cumulative way. to dominate unconsecrated persons and promoted secrecy,
To better understand how the experiences described by self- cover-ups, and unsuitable reactions. Comparable structures and
identified survivors may negatively influence the process of their connection to sexual abuse, as well as the use of additional
dealing with CSA it is useful to take a closer look at the repressive measures to ensure the silence of survivors, can also
Traumagenic Dynamics Model of CSA by Finkelhor and Browne be found in children’s homes and other state institutions in the
(1985) as well as the CSA healing model by Draucker et al. (2011). FRG and GDR (Wazlawik et al., 2014; Hackenschmied et al.,
The first provides a particularly suitable framework as to why the 2018; Sachse et al., 2018). Since many of the events reported by
described experiences are perceived as problematic by survivors, survivors took place a long time ago, it can be assumed that the
while the second might explain how the identified factors may circumstances in some institutions have changed considerably,
hinder the process of dealing with CSA. while other institutions (e.g., those in the GDR) no longer
Our results suggest that disclosure, or rather negative social exist. Nevertheless, Nagel et al. (2021) point out, that important
reactions to disclosure attempts, pose an obstacle in dealing insights can be drawn from reports of survivors which should
with CSA. This is mostly because survivors get discouraged be taken into account in the development of institutional
to share their story with anyone else out of shame, after being prevention programs today.
confronted with resentment and disbelief by loved ones or Manipulative perpetrator behavior may also pose as an
formal sources (Birck, 2001). This seems to be because the obstacle in the process of dealing with CSA, mainly by delaying
internalized stigma and the utmost feeling of betrayal connected or hindering disclosure. Our findings implicate that perpetrators
to the traumatic experience are relived in moments of these used a variety of violent and non-violent strategies to facilitate
unsuccessful disclosure attempts (Finkelhor and Browne, 1985). abuse and prevent exposure. Besides, perpetrators induced shame

and guilt by creating an illusion of consensuality, instilling in (witness credibility assessments, plausibility check) (Draucker
victims the feeling of being actively and willingly involved in et al., 2011). Survivors frequently described these experiences as
the abuse. This shame and guilt oftentimes kept survivors from reinforcing the belief that they are not believed, accompanied
disclosing and regaining self-worth, self-efficacy, and restoration by a deep-rooting feeling of loss of control. They feared being
of self-confidence, thus aggravating the feeling of powerlessness unable to influence the alleged “outcome” of the process, which
and stigma (Finkelhor and Browne, 1985). These findings are in combination with missing information about the procedure
supported by Schröder et al. (2020b) who found that perpetrators and partial insensitivity by experts led to severe stress for these
used similar strategies to prevent exposure in the context of survivors (see also Schröder et al., 2020a). Furthermore, the
organized and ritual CSA. In addition, Wolf and Pruitt (2019) procedure of witness credibility assessment in Germany has
found that especially violent grooming in form of threats predicts been criticized over the years, some authors suggest evaluating
higher trauma symptoms, anxiety, depression, sleep problems, the procedure for a diverse group of survivors and specific
and dissociative issues in survivors of CSA. Draucker et al. circumstances (Schoon and Briken, 2019).
(2011) describe “personal agency” as another factor enabling Our cross-category analysis showed that many of the
survivors to eliminate shame and guilt by acknowledging that identified factors interact with each other in a layered and
what happened to them was wrong and that they are not at fault. complex way. Most notably many themes interact with factors
A feeling of personal agency therefore could help survivors to hindering disclosure. This result is supported by Alaggia
regain controllability and to fight the feeling of powerlessness et al. (2019) who analyzed 33 studies between 2000 and 2016
(Finkelhor and Browne, 1985; Frazier et al., 2004; Draucker et al., examining factors influencing disclosure of CSA and found that
2009). amongst other things a close, family-like relationship to the
Lastly, survivors reported several obstacles regarding the state perpetrator, shame, self-blame, fear of negative consequences,
legal system and mental health care in Germany. The most and stigma, dysfunctional family communication and a general
pressing issues being a lack of access to adequate (psycho)therapy, lack of discussion about sexuality within society are obstacles
financial burdens, duration of court proceedings, and witness for disclosing experiences of CSA at any given point. Similar
credibility assessment as well as plausibility checks. These results have been shown by Birck (2001) who interviewed
findings are consistent with research by Görgen et al. (2012) who 22 women who had experienced “sexualized violence in the
showed, that apart from the better documented psychological context of a relationship of trust” in their childhood and who
and physical long-term effects of CSA financial burdens and had completed psychotherapy as adults. She could show that
other long-term economic consequences may play a substantial disclosure was often followed by denial responses (disbelief,
role in the process of dealing with CSA but haven’t been defensiveness, blaming the victim) by the family of origin
examined in detail yet. In line with this Herman (1994) as well especially. Nevertheless, disclosing their experiences was
as Figley (2013) stated that financial, just as much as social perceived as liberating and empowering by survivors, especially
and personal resources, are necessary for survivors to rebuild a in the long-term perspective. Birck (2001) concludes, that
safe environment for themselves and help them integrate their disclosure is an important milestone in the process of dealing
experiences in self and world. Our data illustrate how financial with CSA. Especially since the reactions to disclosure, positive or
burdens acted as a direct obstacle, as they prevented survivors negative, influence the further process in a significant way. Thus,
from regaining their sense of security (Herman, 1994; Gahleitner, those who received an affirmative message were more likely
2003; Figley, 2013), and as an indirect obstacle, as they e.g., to seek therapeutic help while those who experienced negative
hindered survivors from accessing suitable psychotherapy. reactions often decided never to talk about it again. This matches
Despite several measures taken to better support survivors the enabling factors “affirmative message” and “ongoing support”
in recent years, negative experiences in court proceedings were identified by Draucker et al. (2011).
reported by survivors. This finding is supported by the results All obstacles identified in this study contain aspects of the
of a study by Dreßing et al. (2018) in which survivors from a traumatic experience (traumatic sexualization, stigmatization,
wider age range also reported experiencing stress and discomfort powerlessness, and betrayal) and thus a reliving of it (Finkelhor
as well as long-term dissatisfaction with the court proceedings and Browne, 1985). They may thus prevent survivors from
and judicial outcome in some cases. This may be due to integrating the traumatic experiences into their self-concept and
measures not being implemented properly in practice, measures moving forward in the process of dealing with experiences of
not being communicated by victim representatives, or victim- CSA (Gahleitner, 2003; Draucker et al., 2011). All this could be
oriented discussion awakening expectations in survivors that an explanation for why survivors of CSA often report that the
cannot be fulfilled (Volbert, 2012). The author concludes that experiences following the abuse were at least as bad as the abuse
many obstacles and hurdles often reported by survivors of itself (Birck, 2001).
CSA describe factors that are inherent to the legal process and
therefore cannot be eradicated easily.
Nevertheless, it seems important to point out, that not LIMITATIONS
being able to control the legal process may enhance the feeling
of powerlessness in survivors (Finkelhor and Browne, 1985). The material used in this study represents a cross-sectional
Regaining control as part of personal agency therefore might also sample and was not primarily collected for scientific purposes.
play an important role in assessment or evaluation situations This poses a methodological challenge. Nevertheless, it enables

an impartial approach to experiences of sexualized violence from It should be noted that in recent years there have been several
the perspective of survivors in a unique way. studies examining specific structures enabling abuse in individual
Additionally, the current sample “suffers” from selection institutions also formulating recommendations for the further
bias on different levels. Firstly, survivors who responded to handling of reprocessing (Keupp et al., 2017; Dreßing et al.,
the Commission’s call to tell their story to the Independent 2018; Rau et al., 2019). Our results suggest that it could further
Inquiry in a private session might differ from those who did be useful to provide information about grooming strategies
not. Whether a survivor decided to share their experiences using and their long-term effects on survivors as part of prevention
one of the available formats might be linked to numerous and and rehabilitation processes as well as include these in training
heterogeneous motives we can neither identify nor anticipate programs of clinicians. Service providers and governmental
in hindsight. agencies also play a role in supporting and guiding survivors in
Furthermore, the sample might be prone to gender bias as the process of dealing with experiences of CSA. Therefore we
well. Up to this point, 83% of the survivors who shared their story agree with Sivagurunathan et al. (2019) and believe it useful to
with the IICSAG were female and 16% male. However, this is not educate personnel accordingly. Concerning witness credibility
representative. We have twice as many female survivors as male assessment, our findings suggest that educating evaluators about
survivors in our sample. We did not address potential gender the aspects of the evaluation process survivors describe as
differences regarding obstacles. Additionally, participants in our obstacles and needs of survivors could be helpful. Potentially, a
sample were 54 years old on average. This means that in most uniform certification of credibility assessors in Germany could
cases experiences of CSA go back up to 45 years and therefore also be discussed after further researching this specific topic.
relate to structures and circumstances that no longer exist today.
Nevertheless, our results implicate that many obstacles and
hurdles survivors reported still exist today and are encountered CONCLUSION
by survivors of diverse contexts in a reoccurring pattern in the
process of dealing with experiences of CSA. Our findings are consistent with prior research on the topic and
Besides, the study included only a small sample out of 1,303 first and foremost support the assumption that survivors of CSA
private sessions due to the scope of the framework of the research in Germany encounter a variety of obstacles in the process of
consortium “Auf-Wirkung.” Furthermore, the sample was not dealing with their experiences. These obstacles may actively or
picked randomly, since only a fraction of the private sessions has passively hinder survivors to integrate their experiences in their
been transcribed verbatim. The data therefore cannot be regarded concept of self and world and reestablish controllability over
as representative and generalized conclusions are not possible. their lives. Regaining control is a counterpart to experiences and
Furthermore, due to the sensitivity of the subject, we were able feelings of persistent powerlessness and therefore significant in
to get permission to quote only from a few participants. the recovery process of survivors of CSA. While some of the
Among other reasons the open question format and the findings can be supported by existing research, others have not
unstructured interview style, have ensured that only a superficial been investigated thoroughly, yet. Among these are financial
exploration of what survivors may themselves interpret as obstacles, lack of access to adequate psychotherapy, aspects
hindering has taken place. On the other hand, the detailed regarding the (German) legal state system (e.g., duration of
description of a variety of obstacles implies a high relevance of judicial court proceedings, witness credibility assessment). All
these issues for the survivors interviewed in this setting. the factors and recurring experiences identified in this study
can be viewed individually, but their influence on the process
of dealing with CSA becomes clear when they are considered in
PRACTICAL IMPLICATIONS their complexity. Future research may focus on survivors with
Our findings suggest that creating safe spaces and opportunities a wide range of backgrounds in contexts, using a structured
for disclosure could be helpful in more than one way for interview potentially asking specifically about obstacles in dealing
survivors in the process of dealing with experiences of CSA with experiences of CSA.
(Chouliara et al., 2014). In the case of state and religious
institutions this could be achieved for example by these
institutions ensuring that, as part of prevention programs,
suitable contact persons are available for potential victims of The dataset consists of transcripts with survivors of child sexual
sexualized violence. Furthermore, given the replicated reports abuse. Only the project researchers are allowed to view and
of lack of access to therapy a potential expansion of the analyze the data. The Data set can not be made available
availability of suitable therapy places for survivors of CSA in to anyone outside the research team. Questions regarding the
Germany could be discussed (Görgen et al., 2012). Focusing datasets should be directed to w.schoon@uke.de.
on low-threshold offers and specific expertise could be helpful
(Pawils et al., 2017; Alaggia et al., 2019). Insights and research
regarding obstacles in the process of dealing with CSA and their ETHICS STATEMENT
meaning for survivors could be part of targeted training for
clinicians. Institutions willing to process cases of child sexual The studies involving human participants were reviewed and
abuse in the past are recommended to ensure the participation approved by Ethics Committee of the Department of Education
of survivors as a part of recognition (Kavemann et al., 2019). at Goethe University Frankfurt. The patients/participants

provided their written informed consent to participate in conducted in cooperation with the Independent Inquiry into
this study. Child Sexual Abuse in Germany (IICSAG), which was appointed
by the Independent Commissioner for Child Sexual Abuse Issues
AUTHOR CONTRIBUTIONS and the German Federal Ministry for Family Affairs, Senior
Citizens, Women, and Youth in 2016.
PB designed the study. Data was collected by PB and the IICSAG.
WS and PB analyzed and interpreted the data. WS wrote the ACKNOWLEDGMENTS
initial draft of the manuscript in constant consultation with PB.
PB and WS had full access to all the data in the study and take The authors want to thank all survivors, who shared their
responsibility for the integrity of the data and the accuracy of data story with the Independent Inquiry into Child Sexual Abuse in
analysis. All authors have contributed to, read, and approved the Germany (IICSAG) and made the transcripts of their private
final version of the manuscript. sessions available for research. We also want to thank Ricarda
Bauch and Melina Westermann for their support.
FUNDING
SUPPLEMENTARY MATERIAL
The present study was funded by the Federal Ministry of
Education and Research in Germany, as part of the second line The Supplementary Material for this article can be found
of funding Research on sexualized violence against children and online at: https://www.frontiersin.org/articles/10.3389/fpsyg.
youth in pedagogical contexts (grant number: 01SR1709C). It was 2021.619036/full#supplementary-material
REFERENCES Gahleitner, S. (2003). Sexuelle Gewalterfahrung und ihre Bewältigung bei Frauen
und Männern: eine explorative Untersuchung aus salutogenetischer Perspektive.
Alaggia, R., Collin-Vézina, D., and Lateef, R. (2019). Facilitators and barriers to Dissertation. Berlin: Freie Universität Berlin.
child sexual abuse (CSA) disclosures: a research update (2000–2016). Trauma Görgen, T., Rauchert, K., and Fisch, S. (2012). Langfristige Folgen sexuellen
Violence Abuse 20, 260–283. doi: 10.1177/1524838017697312 Missbrauchs Minderjähriger. Forensische Psychiatr. Psychol. Kriminol. 6, 3–16.
Birck, A. (2001). Die Verarbeitung einer sexuellen Missbrauchserfahrung in der doi: 10.1007/s11757-011-0129-0
Kindheit: bei Frauen in der Psychotherapie [Doktorarbeit]. Köln: Universität Hackenschmied, G., Mosser, P., and Keupp, H. (2018). Individuelle und
zu Köln. kollektive Aufarbeitung sexualisierter Gewalt am Beispiel der Odenwaldschule
Chouliara, Z., Karatzias, T., and Gullone, A. (2014). Recovering from childhood – ein Werkstattbericht. Forensische Psychiatr. Psychol. Kriminol. 13:814.
sexual abuse: a theoretical framework for practice and research. J. Psychiatr. doi: 10.1007/s11757-018-0508-x
Ment. Health Nurs. 21, 69–78. doi: 10.1111/jpm.12048 Herman, J. L. (1994). Die Narben der Gewalt: Traumatische Erfahrungen verstehen
Craven, S., Brown, S., and Gilchrist, E. (2006). Sexual grooming of children: und überwinden (V.Koch and R. Weitbrecht, Übers; Original erschienen 1992:
review of literature and theoretical considerations. J. Sexual Aggr. 12, 287–299. Trauma and Recovery). Seattle, WA: Kindler.
doi: 10.1080/13552600601069414 Horowitz, M. J. (1986). Stress-response syndromes: a review of posttraumatic
Draucker, C. B., Martsolf, D. S., Roller, C., Knapik, G., Ross, R., and Stidham, A. and adjustment disorders. Hosp. Commun. Psych. 37, 241–249.
W. (2011). Healing from childhood sexual abuse: a theoretical model. J. Child doi: 10.1176/ps.37.3.241
Sex. Abus. 20, 435–466. doi: 10.1080/10538712.2011.588188 Kavemann, B., Nagel, B., Doll, D., and Helfferich, C. (2019). Erwartungen
Draucker, C. B., Martsolf, D. S., Ross, R., Cook, C. B., Stidham, A. W., Betroffener sexuellen Kindesmissbrauchs an die gesellschaftliche Aufarbeitung.
and Mweemba, P. (2009). The essence of healing from sexual violence: a Berlin: Independent Commission to Investigate Child Sexual Abuse.
qualitative metasynthesis. Res. Nurs. Health 32, 366–378. doi: 10.1002/nur. Keupp, H., Straus, F., Mosser, P., Gmür, W., and Hackenschmied, G. (2017).
20333 Schweigen – Aufdeckung – Aufarbeitung. Wiesbaden: Springer Fachmedien.
Dreßing, H., Salize, H. J., Dölling, D., Hermann, D., Kruse, A., Schmitt, doi: 10.1007/978-3-658-14654-2
E., et al. (2018). Sexueller Missbrauch an Minderjährigen durch katholische Kleber, R. J., and Brom, D. (1992). Coping with Trauma: Theory, Prevention, and
Priester, Diakone und männliche Ordensangehörige im Bereich der Deutschen Treatment. Lisse: Swets and Zeitlinger.
Bischofskonferenz. Mannheim, Heidelberg, Gießen. Kuckartz, U. (2018). Qualitative Inhaltsanalyse. Methoden, Praxis,
Figley, C. R. (2013). Trauma And Its Wake. New York, NY: Routledge. Computerunterstützung (4. Auflage). Grundlagentexte Methoden. Beltz
doi: 10.4324/9780203776209 Juventa. Available online at: http://ebooks.ciando.com/book/index.cfm?bok_
Filipas, H. H., and Ullman, S. E. (2001). Social reactions to sexual assault victims id/2513416 (accessed February 20, 2021).
from various support sources. Violence Vict. 16, 673–692. Murthi, M., and Espelage, D. L. (2005). Childhood sexual abuse, social support, and
Finkelhor, D. (1987). The trauma of child sexual abuse: two models. J. Interpers. psychological outcomes: a loss framework. Child Abuse Negl. 29, 1215–1231.
Violence 2, 348–366. doi: 10.1177/088626058700200402 doi: 10.1016/j.chiabu.2005.03.008
Finkelhor, D., and Browne, A. (1985). The traumatic impact of child Nagel, B., Kavemann, B., Pham, S., and Helfferich, C. (2021). Räume und
sexual abuse: a conceptualization. Am. J. Orthopsychiatry 55, 530–541. Organisationskultur in stationären pädagogischen Einrichtungen. Trauma
doi: 10.1111/j.1939-0025.1985.tb02703.x Gewalt 15, 64–77. doi: 10.21706/tg-15-1-64
Fischer, G., and Riedesser, P. (2016). Lehrbuch der Psychotraumatologie (4., Pawils, S., Nick, S., Metzner, F., Lotzin, A., and Schäfer, I. (2017).
aktualisierte und erweiterte Auflage). UTB Medizin, Psychologie: Vol. 8165. Versorgungssituation von Kindern, Jugendlichen und Erwachsenen
München: Ernst Reinhardt Verlag. Available online at: http://www.utb-studi- mit sexuellen Gewalterfahrungen in Deutschland: Ein kritischer
e-book.de/9783838581651 (accessed February 20, 2021). Überblick [Healthcare situation for children, adolescent and
Frazier, P., Tashiro, T., Berman, M., Steger, M., and Long, J. (2004). Correlates of adult victims of sexual violence in Germany: a critical overview].
levels and patterns of positive life changes following sexual assault. J. Consult. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 60,
Clin. Psychol. 72, 19–30. doi: 10.1037/0022-006X.72.1.19 1046–1054. doi: 10.1007/s00103-017-2601-2

Price, J. L., Hilsenroth, M. J., Petretic-Jackson, P. A., and Bonge, D. (2001). A review abuse: the service providers’ perspective. Child Abuse Neglect 88, 455–465.
of individual psychotherapy outcomes for adult survivors of childhood sexual doi: 10.1016/j.chiabu.2018.08.015
abuse. Clin. Psychol. Rev. 21, 1095–1121. doi: 10.1016/S0272-7358(00)00086-6 Sommer, J. (2016). Die psychotherapeutische Versorgungsrealität komplex
Rau, M., Breiling, L., and Rettenberger, M. (2019). Regensburger traumatisierter Menschen in Deutschland - Ergenisse einer Studie. Göttingen:
Aufarbeitungsstudie. Wiesbaden: KrimZ. Initiative Phoenix Bundesnetzwerk für angemessene Psychotherapie e.V.
Robinson, O. C. (2014). Sampling in interview-based qualitative research: doi: 10.21706/tg-10-4-308
a theoretical and practical guide. Qual. Res. Psychol. 11, 25–41. Ullman, S. E., and Peter-Hagene, L. (2014). Social reactions to sexual assault
doi: 10.1080/14780887.2013.801543 disclosure, coping, perceived control and PTSD symptoms in sexual assault
Roth, S., and Newman, E. (1993). The process of coping with incest for adult victims. J. Commun. Psychol. 42, 495–508. doi: 10.1002/jcop.21624
survivors. J. Interpers. Violence 8, 363–377. doi: 10.1177/088626093008003005 Ullman, S. E., Townsend, S. M., Filipas, H. H., and Starzynski, L. L. (2007).
Runyan, D. K., Everson, M. D., Edelsohn, G. A., Hunter, W. M., and Coulter, M. L. Structural models of the relations of assault severity, social support, avoidance
(1988). Impact of legal intervention on sexually abused children. J. Pediatr. 113, coping, self-blame, and PTSD among sexual assault survivors. Psychol. Women
647–653. doi: 10.1016/S0022-3476(88)80373-1 Q. 31, 23–37. doi: 10.1111/j.1471-6402.2007.00328.x
Sachse, C., Knorr, S., and Baumgart, B. (2018). Sexueller Missbrauch in der Unabhängige Kommission zur Aufarbeitung sexuellen Kindesmissbrauchs (ed.).
DDR: Historische, Rechtliche und Psychologische Hintergründe des Sexuellen (2019). Geschichten die zählen: Bilanzbericht 2019. Göttingen: Initiative Phoenix
Missbrauchs an Kindern und Jugendlichen in der DDR. Sexuelle Gewalt Bundesnetzwerk für angemessene Psychotherapie e.V.
in Kindheit und Jugend. Vieweg. Available online at: https://ebookcentral. Volbert, R. (2012). “Geschädigte im Strafverfahren: Positive Effekte oder
proquest.com/lib/gbv/detail.action?docID=5261350 (accessed February 20, sekundäre Viktimisierung?,” in Ambivalenzen der Opferzuwendung des
2021). Strafrechts: Zwischenbilanz nach einem Vierteljahrhundert opferorientierter
Sánchez-Meca, J., Rosa-Alcázar, A. I., and López-Soler, C. (2011). The Strafrechtspolitik in Deutschland. 1st Edn, eds S. Barton and R. Kölbel
psychological treatment of sexual abuse in children and adolescents: a meta- (Baden-Baden: Nomos Verlagsgesellschaft mbH and Co. KG), 195–213.
analysis. Int. J. Clin. Health Psychol. 11, 67–93. Available online at: https://www. doi: 10.5771/9783845239781-195
redalyc.org/pdf/337/33715423005.pdf Walsh, K., Fortier, M. A., and DiLillo, D. (2010). Adult coping with childhood
Schönbucher, V., Maier, T., Mohler-Kuo, M., Schnyder, U., and Landolt, sexual abuse: a theoretical and empirical review. Aggress. Violent Behav. 15,
M. A. (2014). Adolescent perspectives on social support received in the 1–13. doi: 10.1016/j.avb.2009.06.009
aftermath of sexual abuse: A qualitative study. Arch. Sex. Behav. 43, 571–586. Wazlawik, M., Dekker, A., Henningsen, A., and Retkowski, A. (2014). Sexuelle
doi: 10.1007/s10508-013-0230-x Gewalt gegen Kinder und Jugendliche in pädagogischen Kontexten. Soziale
Schoon, W., and Briken, P. (2019). Zur Anwendbarkeit der Passagen 6, 167–176. doi: 10.1007/s12592-014-0158-6
Glaubhaftigkeitsbegutachtung unter spezifischen Begutachtungsumständen – Wolf, M. R., and Pruitt, D. K. (2019). Grooming hurts too: the effects of types of
Eine narrative Übersichtsarbeit. Forensische Psychiatr. Psychol. Kriminol. 13, perpetrator grooming on trauma symptoms in adult survivors of child sexual
125–135. doi: 10.1007/s11757-019-00530-x abuse. J. Child Sex. Abus. 28, 345–359. doi: 10.1080/10538712.2019.1579292
Schröder, J., Behrendt, P., Nick, S., and Briken, P. (2020a). Hintergründe Wyatt, G. E., and Mickey, M. R. (1987). Ameliorating the effects of child sexual
und psychische Folgen organisierter und ritueller Gewalt – Berichte an die abuse. J. Interpers. Violence 2, 403–414. doi: 10.1177/088626058700200406
Unabhängige Kommission zur Aufarbeitung sexuellen Kindesmissbrauchs
[Background and psychological impact of organised and ritual child sexual Conflict of Interest: PB is a member of the IICSAG.
abuse: reports to the Independent Inquiry into Child Sexual Abuse in
Germany]. Fortschr. Neurol. Psychiatr. 88, 374–378. doi: 10.1055/a-11 The remaining author declares that the research was conducted in the absence of
49-2103 any commercial or financial relationships that could be construed as a potential
Schröder, J., Behrendt, P., Nick, S., and Briken, P. (2020b). Was erschwert conflict of interest.
die Aufdeckung organisierter und ritueller Gewaltstrukturen? [Which
factors impede the uncovering of organized and ritual child sexual Copyright © 2021 Schoon and Briken. This is an open-access article distributed
abuse? - a qualitative content analysis of reports of victims and under the terms of the Creative Commons Attribution License (CC BY). The use,
contemporary witnesses]. Psychiatr. Praxis 47, 249–259. doi: 10.1055/a-1123- distribution or reproduction in other forums is permitted, provided the original
3064 author(s) and the copyright owner(s) are credited and that the original publication
Sivagurunathan, M., Orchard, T., MacDermid, J. C., and Evans, M. (2019). Barriers in this journal is cited, in accordance with accepted academic practice. No use,
and facilitators affecting self-disclosure among male survivors of child sexual distribution or reproduction is permitted which does not comply with these terms.

ORIGINAL RESEARCH
published: 11 May 2021
doi: 10.3389/fsoc.2021.623661
Child Sexual Abuse as Lifespan

Trauma Within the Context of Intimate
Partner Violence: Experiences of
Caribbean Women
Adele D. Jones *
The None in Three Centre for the Prevention of Gender-Based Violence, Department of Behavioural and Social Sciences, The
University of Huddersfield, Huddersfield, United Kingdom
Background: There is a dearth of research which explores sexual abuse from

perspectives of Caribbean women, despite its high prevalence in the region. While
sexual violence is universal, tackling it requires a deep understanding of the contextual
specificities in which it arises and of the intersections of gender with other sources of
oppression and marginalisation. It also calls for the recognition that intimate partner
violence against women is not separate from, but linked to violence against girls, not
only because both are forms of gender-based violence but because together they speak to
its historical, persistent and accumulative effects.
Edited by:
Robert Horselenberg, Methods: In-depth intensive interviews were carried out with 35 women from Barbados
Maastricht University, Netherlands
and Grenada, aged 18–60 years who had experienced intimate partner violence under one
Reviewed by:
Carmen Martínez, of the following circumstances: during pregnancy (n 15), as a woman a with disability (n
University of Murcia, Spain 8), as a woman living with HIV (n 12). Interviews were digitally recorded, transcribed and
Tracie Rogers,
University of the West Indies, Jamaica
thematically analyzed.
Karene Nathaniel,
Results: The participants experienced multiple forms of violence within their relationships,
The University of the West Indies St.
Augustine, Trinidad and Tobago often concurrently. Twenty-one of the women had been subject to sexual violence and of
*Correspondence: these, 19 had experienced sexual abuse as children; these experiences were viewed as
Adele D. Jones interconnected and bolstered by the high level of violence-acceptance reported within
a.d.jones@hud.ac.uk
communities. Women were subject to different forms of control by their partners
Specialty section: depending upon prevailing discourses related to their circumstances (as pregnant,
This article was submitted to disabled, or HIV positive); being ‘vulnerable’ was synonymous with having one’s agency
Gender, Sex and Sexualities,
as an independent, autonomous person constrained and little external help was available.
Frontiers in Sociology
Conclusion: The study identified a clear chain of sexual behaviors, each of which fuel
Received: 30 October 2020
different layers of the problem: the prevalence of early sexualization of children is
Accepted: 22 April 2021
Published: 11 May 2021 associated with the prevalence of child sexual abuse; child sexual abuse is pervasive
Citation: in large part, because of the normalisation and social acceptance of violence against
Jones AD (2021) Child Sexual Abuse women and girls; “cultural” normalcy, in turn, fuels attitudes which contribute to sexual
as Lifespan Trauma Within the Context
of Intimate Partner Violence: violence against women and women in especially vulnerable circumstances face additional
Experiences of Caribbean Women. risks. Integrated policy, which tackles these as interconnected issues is called for.
Front. Sociol. 6:623661.
doi: 10.3389/fsoc.2021.623661 Keywords: child sexual abuse, Caribbean, pregnant, living with HIV, disability, woman, IPV
Frontiers in Sociology | www.frontiersin.org 106

1 May 2021 | Volume 6 | Article 623661
Jones Child Sexual Abuse as Lifespan Trauma
INTRODUCTION IPV and the meanings they attribute to associations with child
sexual abuse.
The sexual abuse of children (CSA) is high in the Caribbean; in a
survey of 15,695 students aged 10–18 years from nine countries,
47.6% of girls and 31.9% of boys reported having been subjected WOMEN IN ESPECIALLY “VULNERABLE”
to sexual abuse (Halcón et al., 2003). Further, the initiation of CIRCUMSTANCES–A CONSTRUCTIVIST
children into sexual activity is said to occur earlier in the
Caribbean than anywhere else in the world, excepting
LENS
countries that practice child marriage. In surveys carried out This research uses an intersectional lens to investigate the
by UNICEF, approximately 15% of children aged 11–12 years convergence of gender, sites of “vulnerability” and difference
and, 35% of young people, 14–15 years old reported having had among Caribbean women who have experienced sexual violence.
sexual experience (UNICEF, 2006; UNICEF, 2008; UNICEF, Intersectionality originated as a conceptual tool to analyze the
2012). Early sexual activity is highly correlated with sexual effects of race, gender and other forms of discrimination on
abuse and coercion, both as a causal link and also in women of color (Crenshaw, 1989) and is now widely used for the
increasing vulnerability to further victimisation (see for analysis of multi-layered social and institutional structures, their
instance, Barrow and Ince, 2008). In the study by Halcon inter-connectedness and the ways in which they intersect with
et al. (2003), over half of boys and about a quarter of girls women’s gendered identities (Sumi et al., 2013; Carastathis, 2016;
who were sexually active stated that the age of first intercourse Hill Collins, 2019). Women in many Caribbean countries have
was 10 years or younger and almost two-thirds had intercourse made extraordinary strides in challenging inequalities, and
before the age of 13, often as a result of force or coercion. human rights indicators point to the growth of women’s
Retrospective studies with adult survivors support these progress and achievements. However, despite their higher
findings. In Barbados, for example, 30% of female respondents education attainment, women are less likely than men to hold
had been sexually abused as children. This study, which included positions of power, have higher levels of unemployment and are
several countries, concluded that forced sexual initiation and more likely to be subjected to gender-based violence, including
early childhood abuse were not uncommon in the Caribbean sexual violence (UNFPA Caribbean, 2017). Furthermore,
(World Health Organization, 2013). Jones and Trotman Jemmott ingrained systems of patriarchy, intersecting with other
explored the perceptions, attitudes and opinions of adults in structures (economic, social, and cultural traditions etc.) that
regard to CSA within six Caribbean countries and highlighted together make up identities mean that many Caribbean women
societal acceptance of violence, patriarchal values which minimise continue to be constrained and controlled (Jones et al., 2014).
the rights of women and children and, norms which associate Patriarchy, viewed as symbiotic with processes of domination,
masculinities with domination and sexual entitlement, as key provides the context in which the phrase “women in especially
factors (2009). Violence within the home, including CSA, can vulnerable circumstances” is used in this article. Un-
contribute to the intergenerational transmission of intimate problematized, the terms connotes victimhood rather than
partner violence (IPV) including, sexual violence (Heise and agency, but situated within the critical realist standpoint
Garcia-Moreno, 2002; Capaldi et al., 2012; Jones et al., 2014); adopted for the study, “vulnerability” is viewed not as
indeed, survivors of one form of violence are more likely to be objective descriptor, but as a fluid state constructed at the
victims of other forms and, girls who have been sexually abused nexus through which normative ideology (the social meanings
are more likely to experience sexual re-victimization and be a attributed to women’s health status, in this instance as pregnant,
victim of IPV in adulthood (Milletich et al., 2010; Franklin and disabled, or living with HIV) meets structural reality.
Kercher, 2012). Focusing on women in especially vulnerable circumstances
There is a dearth of qualitative research which explores the can unearth what might be known about sexual violence among
problem from the perspectives of Caribbean women, despite UN women from diverse and marginalised backgrounds, but a critical
reports that every one of the Caribbean islands has a sexual realist perspective takes this further and insists on examining the
violence rate higher than the world average (United Nation, 2007; symbolic and material structures that limit women’s freedoms.
Guedes, 2012). Also missing, are studies which examine Critical realism calls also, for the recognition of human agency
differential experiences of sexual violence among women (Humphries, 2008). Acknowledging agentic power is important,
whose circumstances may place them at increased risk of but oftentimes analysis offers little more than a rhetorical nod in
victimisation. While sexual abuse is universal, tackling it the direction of women’s empowerment. It is not enough to
requires attention to the local social, cultural and contextual idealise agentic womanhood in constrained environments as
specificities in which it arises and also, to the intersections of heroic; we need to understand more deeply, the contexts and
gender-based violence with other sources of oppression. This liminality of women’s conscious action (Hill Collins, 2000). This
article makes a contribution to bridging this gap. study thus also makes use of Einspahr (2010) structural theory of
The study draws on the experiences of three groups of freedom which regards freedom as non-domination and calls for
Caribbean women in “especially vulnerable circumstances” the dismantling of the constraining effects of patriarchy as it
who were victims of IPV: women who were pregnant, impacts both the material and, symbolic circumstances of
disabled women and, women living with HIV. The study women. This is crucial in tackling sexual abuse in the
explores their experiences of sexual violence as a feature of Caribbean since here, hegemonic versions of masculinity not

only dictate what it means to be gendered, but also, what it means adults) and reported similar concerns. From a total of 9086
to be sexualised, as explored by Kempadoo and Taitt (2006) in females who completed the survey in 2010, women with a
their review of 150 journal articles, programs and “grey” disability were significantly more likely to experience rape
literature. In respect of the entry points of analysis for this than other women (1.7 compared to 0.4), other forms of
article, it also dictates what it means to be a woman who is sexual violence (4.5 compared to 1.8), physical abuse (7.1
pregnant, disabled or living with HIV. compared to 3.3), stalking (21.0 compared to 12.2), and
IPV is not demographic specific and there is no typical victim. psychological abuse and control of sexual health (2.4
There are myriad ways in which women might be rendered compared to 1.4) (Breiding and Armour, 2015). In the
vulnerable to IPV: homelessness; drug and alcohol abuse; United Kingdom, research commissioned by Women’s Aid
mental ill-health; trafficking; environmental disasters; ill-health (Hague et al., 2011) involving a survey of domestic violence
and more. There was, of course, simply no way of including all and, disability organisations and, interviews with 30 disabled
these groups in the study and through consultation with women’s women revealed that 50% of women with disabilities had
organisations in the participating countries, a consensus experienced IPV compared with 25% of women without
approach was used to determine the final selection. disabilities; they were twice as likely to be assaulted or raped
and were likely to have to endure IPV for longer because of the
Pregnancy and Intimate Partner Violence lack of appropriate support. Although there are some studies on
Accurate data on IPV during pregnancy is scarce in the English- the victimization of women with disabilities in Latin America
speaking Caribbean yet estimates for other Latin America and and the Caribbean, there are no reliable prevalence data. As the
Caribbean countries are among the highest in the world (Han and region has one of the highest rates of violence against women in
Stewart, 2014) signalling this as a significant public health the world, it can be assumed that Caribbean women with
problem across the region (Campbell et al., 2004). IPV during disabilities are disproportionately impacted by IPV because
pregnancy is a global problem too. In a Canadian study, 10.5% of of limited access to services and the position of power and
women of the 23,766 respondents, reported being physically and/ control abusers may have over them. Disabled women are also
or sexually abused during pregnancy (Taillieu et al., 2015); in likely to be more dependent on their abuser for care and support
India, 18% of 2199 women experienced violence during their last than other groups of women and have reduced access to
pregnancy (Ahmed et al., 2006) and a study in Uganda revealed economic autonomy; these factors exacerbate isolation and
that of 612 Ugandan women screened in their second trimester, vulnerability.
27.7% had been subjected to IPV. Burch and Gallup Jr (2004)
conducted research with 258 men convicted of spousal abuse and
found that the severity and frequency of violence they inflicted Living With HIV and Intimate Partner
actually doubled during their partner’s pregnancy. A meta- Violence
analysis of studies from 15 countries, suggests a prevalence Violence against women and HIV are conjoined public health
rate of between one and 28%, with up to half of all victims problems with profound implications for health, wellbeing and
receiving direct blows to the abdomen during pregnancy (Ellsberg social development. The Caribbean has the highest incidence of
et al., 2008). Though pregnancy is not of itself a source of HIV-AIDS outside of Africa and it is estimated that 53% of
vulnerability, it can be a catalyst for violence. (Audi et al., persons living with HIV in the region are women (Allen, 2011).
2008) A Caribbean study noted that for some women, violence Although women are biologically more susceptible to contracting
began when they got pregnant, while for others, men used the HIV infection, this does not fully explain how and why HIV has
pregnancy situation to exert more violence and control than they become so heavily feminized. Global studies reveal that this is
had previously (Jones et al., 2017). Pregnant women are rendered primarily due to gender inequalities, sexual cultures, violence
vulnerable because of IPV, in that violence places them at extreme against women and girls and women’s lack of control over their
risk of adverse maternal and infant outcomes (Silverman et al., sexual rights. Forced sex and rape, place women and girls at
2006; McMahon et al., 2011). particular risk, as these types of sex cause injuries in the vaginal
and anal tissue, allowing the virus to be more easily introduced
Disability and Intimate Partner Violence into the body (Luciano, 2013). HIV transmission in the
Both men and women with disabilities are more likely to Caribbean is increased by complex sexual cultures which
experience violence within their personal relationships than include: the early sexual initiation of girls; sexual exploitation;
non-disabled persons (UNAIDS/WHO, 2007) however, the forced, non-consensual sex; norms about male sexual
cumulative effect of gender inequalities and disablism exposes entitlement; multiple partnering and, sexual-economic
women and girls to a high risk of IPV, coercion and, sexual exchange relations. This construction of complex sexual
violence (Hughes et al., 2012). A survey carried out in European cultures is supported by qualitative research with sex workers
Union member states reported that almost 80% of women with (Kempadoo, 2004), with adolescent girls 16–18 years (Barrow
disabilities involved in the study had been victims of violence, and 2008) and also, through a review of literature conducted by
they were four times more likely to be subjected to sexual violence Bombereau and Allen (2008). At the ideological level, HIV
(la Riviére Zijdel, 2004). Breiding and Armour (2015) drew on and AIDS continue to symbolize deviancy and contagion and,
data from the United States National Intimate Partner and Sexual in the Caribbean, women living with HIV face stigmatization,
Violence Survey (NISVS) (an ongoing, telephone survey of social alienation and victim-blaming. These factors force women

into silence and often cause them to have to mask violence within Procedure
relationships. Violence against women therefore has both a causal This study used convenience, purposive, non-probability
and consequential relationship with HIV. There is a high level of sampling techniques in order to identify participants whose
awareness of these intersections between HIV and IPV among circumstances were appropriate for the research objective. A
Caribbean governments, and international organizations (UN data base of agencies and government departments providing
Women, PAHO/WHO, 2007) have initiated programs designed services to the relevant groups of women in the two countries was
to catalyse integrated responses, nevertheless, policy and created. Agencies that agreed to support the study distributed
programming tend to be fragmented and the mechanisms for information about the research to their service users. Women
effective implementation are often missing. notified the agency of their wish to be included in the study and
initial contacts by email were followed up with confirmatory
The Present Study telephone calls. To ensure participant safety, no direct contact
The empirical data presented in this article was collected in was made with the women and interviews were arranged through
2016 as part of a larger research project on the experiences of the organisation by three research assistants recruited and trained
IPV among women in Barbados and Grenada. The study in conducting IPV research. This approach was necessary given
utilized a cross-sectional qualitative design–data collected at the sensitivity of the issue explored and the potential risks to
one time-point through guided reflective interviews. The women of more representative and open recruitment methods.
objectives of the study were derived from a preliminary The recordings were transcribed verbatim and subject to iterative,
literature review and sought to 1) explore the ways in which in-depth, thematic analysis (Braun and Clarke, 2013). This
women from especially “vulnerable” groups who experience involved three stages: first, multiple, purposeful readings of the
IPV may be exposed to heightened risks as a consequence of transcripts and repeated listening to the recordings without
their partners’ exploitation their situations and, 2) to see what coding to gain an overall impression of meanings of
commonalities or differences exist between the situations of expression in the context in which they were used; second,
the women. detailed coding to identify significant themes on sexual
violence as a feature of IPV among the three groups of
women (pregnant women, women with disabilities and,
METHODS AND MATERIALS women living with HIV) and third, the merging of ideas to
produce thematic maps (Braun and Clarke 2013) and to check for
Participants were recruited through professional networks, patterns across groups of participants and across both countries.
agencies and snowballing. Interviews took between one-two Once the themes were identified, quotes were selected to best
hours and were digitally recorded. Interview prompts were illustrate the findings.
designed to facilitate extended narratives, to elicit information
about patterns of abuse and sexual violence experienced and to Ethical Statement
evoke memories of childhood. These prompts were: 1) How do Ethical approval for the study was granted by the ethics
you define domestic violence? 2) Do you face particular risks of committee at the United Kingdom university leading the
violence because of your health or situation? 3) What do you research. The participants’ identities were anonymised to
think the reasons are for the increase in risk? 4) How does being minimise risk to them and counselling was made available
pregnant/disabled/living with HIV affect you in getting support in the event of re-traumatisation. Participants were provided
in dealing with or escaping a violent relationship? 5) What helps with an information sheet detailing the nature of the study and
you to cope? informed that they could withdraw at any stage of the process.
Informed consent was obtained on this basis. Data collection
Participants and data management were subject to stringent ethical
Thirty-five women participated in the study: 15 of the women protocols to safeguard women’s privacy rights and all
were pregnant, eight were disabled and 12 were living with HIV transcripts and digital recordings were secured through
(four of these women were also pregnant). The ages of the password protection.
participants at the time of the interviews ranged from 18 to
60 years. Twenty of the women were single, separated or divorced
and 15 women were married or in a long-term relationship. Many RESULTS
women were able to participate in the study only because they had
been able to escape the relationship in which they were abused The results are organised around the three situational
and this accounts for the large number of single women in the groupings reflected in the study: 1) Being pregnant, 2)
study. Twenty-eight women had children and 13 women had four Women with disabilities and 3) Women living with HIV.
children or more. Thirty-one of the participants described The three situational groupings, though initially
themselves as poor or low income (a limitation of the predetermined based on the research objectives, were
sampling approach, which recruited participants through considered important to retain because these categories
agencies that primarily provide services to poorer women), contextualised and differentiated women’s experiences in
three women were of average income and one woman ways that mattered to them. In seeking to preserve
reported above average means. difference while at the same time exploring commonality,

severity of violence (e.g., reminding the victim and

perpetrator of the woman’s transgressions). The third theme
‘freedom’ was theoretically derived to explore the concept of
freedom as non-domination or being free from the constraining
effects of patriarchy (Einspahr, 2010). Freedom meant different
things for different women depending upon their situations but
for all women, was centered around how to escape the violence.
This in turn meant how women tried to free themselves from
patriarchal domination, to challenge the structural
disadvantages they faced, and to confront disempowering
discourses and the lack of services. In the absence of being
able to escape, it also meant the strategies women used when
forced to adapt to living within these constraints which included
making incremental improvements to their circumstances
(liminal agency). Liminal agency as described in this study,
can therefore be understood as survivors enacting agency in
ways which improved their day to day lives but bounded within
FIGURE 1 | Liminal survivor agency. external structural barriers, were unable to disrupt them. This is
represented in Figure 1.
In addition to the separate groups of women and the common
within these groupings, three overarching cross-cutting themes across them, two other themes relating explicitly to sexual
themes are discussed: Intimate Partner Violence, Sexual violence are explored: sexual abuse as lifespan trauma and the
Violence and Freedom. The first two of these are a product normalisation of sexual abuse. These themes dominated the
of the analysis. IPV is used to refer collectively to the many narratives of 19 of the 35 women and relate to childhood
forms of violence that were inflicted within the context of the experiences of sexual violence.
women’s intimate partner relationships; sexual violence refers
specifically to violence of a sexual nature. Of the 35 women in Being Pregnant
the study who had experienced IPV, 21 reported that these Intimate Partner Violence
experiences had included sexual violence and 19 referred to Motherhood was an important role for the majority of the
childhood experiences of sexual abuse although in two cases, women. Twenty-eight of the 35 participants had children and
the women did not view this as abuse (these two women had 13 women had four children or more. Half of the mothers (n 14)
been raped as children). Other terms used within the analysis had been subjected to IPV during their pregnancies and 12
include physical violence, which refers to the use of physical women reported that the violence either started or increased
force and, psychological violence which includes verbal and during pregnancy and then persisted throughout the relationship.
emotional abuse and threats of harm. It is important to stress Emma, a 39 year old married woman with two children,
that the revelations of child sexual abuse (CSA) were an experienced abuse from the third month of her second
emergent phenomena from this broader study of IPV and pregnancy:
that in the discussion of results on CSA, only data from the 19
participants who reported this is explored. “When I get home from work tired, my husband would
It is important to acknowledge that these distinctions are a start cursing and abusing me. I cry a lot and sometimes I
theoretical contrivance, which though seeking to provide a think of committing suicide. I blame myself a lot. I think I
comparative picture of women’s experiences, runs the risk of should have been able to see that it was not love. When
underplaying the embeddedness of different types of violence, someone loves you, they don’t hit you. I was not happy
one within the other. In mitigating this risk it is important to and I became very timid”.
stress therefore, that all the participants experienced multiple
forms of violence within their relationships, often concurrently. Most of the abuse reported was physical and emotional:
The women did not describe these as separate forms of violence
but rather, as violence with many “heads and shapes”. So, for “He dragged me in the grass and punched me. My baby
example, physical violence exerted in order to carry out sexual came at seven months, because of the constant verbal
violence was most likely accompanied by verbal abuse. These abuse and one incident of physical abuse.” (Vera,
were not experienced as three types of violence but as single 49 years)
violent acts containing these different and overlapping
elements. Verbal battering by itself seems to have been an Women’s abdomens and unborn babies were often targeted to
ever-present feature of most women’s experiences but when inflict maximum harm:
linked to other violent actions, it served specific functions: a
precursor to an assault, part of the build-up; a form of “He grab me by my throat . . ..and push me up against
‘justification’ for the perpetrator’s benefit; or, fuelling the the wall . . . then he cuff [punched] me in my stomach. He

said you don’t deserve no child and I am killing you Women With Disabilities
tonight”. (Joan). Intimate Partner Violence
“You cannot fight back because you have to study like, to The experiences of women with disabilities revealed that social and
hide my belly to prevent it being hit. It is hard to fight structural isolation compounded the violence they were exposed to.
back; I would study to run but I could not get nowhere Intimate partner violence often entailed the exploitation of specific
because I had a belly”. (Betty) impairments as a means of inflicting psychological harm. These
women did not feel limited by their impairments but their partners
“One day he gave me a blow in my temple and I fell to the
would make use of their impairments to constrain them and this
ground and my cervix tore. I began to bleed and had to be
was limiting. Aya, was blind and when her husband wanted to
rushed to the hospital. My cervix had to be stitched up to
punish her, he would move all of her markers of space and place in
hold the baby for the balance of the pregnancy.” (Peace).
the home so that she would be lost, stumble or require his
assistance. When Cathy, aged forty-one and a mother of five
Sexual Violence
children had a stroke, her husband became abusive. In addition
Sexual violence was also commonly reported among pregnant
to physical and sexual abuse, he told people she had had a “nervous
women. Two women had become pregnant as teenagers as a
breakdown” and was mentally unstable. He was so plausible that
result of rape while other women said they faced sexual assaults
friends began to shun her and even the courts believed him.
right through and immediately after their pregnancies. Jenny
recounted:
“Many times, he forced me to have sex. When I told him
that I would expose him and tell everyone that he is
“. . .he wanted sex. I told him the baby is too young and I
abusing me he told me that he would tell everyone I am
can’t go through with that. He took a knife and he cut off
mad. This was the beginning of the campaign to make
my clothing.”
everyone believe I was mad since the stroke”.
“Once he [describes violent vaginal sexual assault] and
tell me that he will rip the child from inside me.” Michelle had been badly burned in a fire and was visibly
(Maureen) scarred; she was left with physical impairments. A survivor of
physical and sexual violence, she had actually been burned saving
Freedom (Constraints to Escaping Violence) the life of another female victim of violence. Despite the IPV she
Mothers were reluctant to leave violent partners because it was was experiencing in her current relationship, this act was recalled
important to them that their children should not grow up without with pride and seemed symbolic of her resilience–she had no
a father. This was especially the case for women whose pregnancy doubt that she would leave as soon as she “saved a little money”.
had seemed to be the catalyst for violence. These women
convinced themselves that once the child was born the “A man wanted to kill her. I saved her, I put her on my
violence would stop; they were also encouraged by the family back and brought her out. I was 22 years when this
to remain in the relationship for the “sake of the baby”. happened. . .. They were fighting and the man came
Participants who had endured abuse over time recalled feeling home and light the fire in the night. I was in the
it was a “phase” that would pass but were often able to pinpoint newspaper when I got burn up in the fire.”
the time when they realised it would “never stop”. This was most
often when they became aware that their children were being Sexual Violence
directly affected. The participants believed it was important for Women with disabilities seemed to be at particular risk of sexual
women to get out of abusive relationships as soon as the abuse violence, especially if their partners had a role in their personal
starts, regardless of whether one was pregnant or not and yet they care. This violence was sometimes “opportunistic”, inflicted when
also recognised that physical and social constraints made this helping the woman with her personal care; women reported that
extremely difficult especially when there were other children. In it felt like their bodies did not belong to them but were simply
instances where women left their partners, they often felt there for the sexual gratification of their partners. Other women
compelled to return because there was nowhere to get help said their partners expected them to show gratitude, including
and without a partner, they were victim to social disapproval, sexual gratitude, in return for being provided with somewhere to
as Vera explains: live and food. Aya, explained that because she was completely
reliant upon her husband and had nowhere else to go, she
“People are judgmental so it will be hard for persons to believed she had no choice but to endure the abuse. She also
get help when pregnant”. described a longstanding pattern of sexual exploitation both
inside and outside the home.
Andrea, a 39 year old woman living with HIV, was abused
during her three pregnancies. She said she that she could not go “I’ve had of people pushing themselves at me sexually
anywhere because she was “big pregnant” and all she wanted to because I was dependent on them for money”.
do was to protect her baby. She gained temporary respite from the
beatings when the neighbors hid her, but she always had to Cathy, like some other women with disabilities, did not feel she
return. could ever deny her husband sex although he made her feel that

her disability rendered her sexually unattractive. Paradoxically, he had not been tested). She said that he continually mocked her and
continually accused her of “looking for sex” whenever she regularly threatened to tell their friends.
ventured out of the home without him. Cathy was routinely
sexually violated by her husband as a means of surveillance. “. . .he started insulting me in public. Like, he would
make a big fuss about not using the same glass or fork...
“Every time I go out, he would want to know where I Everybody would hear him and you could see what they
went and [describes being sexually assaulted] to see if I were thinking. I asked him, please don’t say that . . . , they
was having sex with someone else. I could not go will think I have some terrible disease. And he would say
anywhere without him thinking I was seeing someone”. well you DO have a terrible disease”.
Freedom (Constraints to Escaping Violence) Sexual Violence

Women with disabilities, like the other women in the study spoke For some women, sexual violence had been the pathway through
of having to depend on their abuser for economic and other forms which they had acquired the virus. Sula, a married woman aged
of assistance. Material dependency was compounded for some 59 years had been living with HIV for many years. She said she
disabled women by the psychological dependency borne out of had first been abused at the age of 9 years and this led to a pattern
chronic abuse and the lack of services to facilitate their of engaging in high risk sexual behavior during adolescence and
independence. Reduced personal autonomy resulted in high as a young adult. Although she had been a child, Sula did not view
levels of control and domination; this created both physical her early sexual encounters as coercive. She referred to being
and psychological barriers to leaving. raped as a young girl as “having sex” and described herself as
“simply going along with men who would come on to me, you
“. . .how can I tell anyone about him when I need him to know”.
help me get up, to bathe, to go to the toilet; how can I Annie did not know she was HIV positive until she attended
leave if I cannot live on my own?” (Sasha). the maternity clinic when she was pregnant. Her husband had
been working away and during this time, she had a sexual
Peace was locked in the house while her partner went to work. relationship with someone she met at college. She was
There were times when he forgot to take the keys but as she was devastated by her diagnosis and believed that she deserved to
not independently mobile, she was unable to leave and in any suffer for her infidelity.
event was so fearful of what he would do upon finding her that she
did not consider this a viable option. Having grown up with an “My husband never looked at me the same again.
acute awareness of IPV from witnessing it in her own family, Although I think he always loved me. But he could
Michelle struggled to escape it. As a disabled woman, “things only never see me the same again. Then he started having
got worse”. She was finally able to leave when someone called the sex with me again but only anally. Very hard. Very
police when she was being held captive. rough. Like an animal. Like a punishment”. (Annie).
“He would tie me up to the bed and have sex with me Andrea, a 37 year old mother of three children, had
when I was tied up. All the women that he has he beats experienced sexual abuse as a child and as an adolescent. She
them and tie them up and rapes them”. also lost a child to abuse; her two-year old daughter died after
being molested by an uncle. In her last relationship, she was
Tracy, a 20 year old single woman with a physical disability “pimped” by her boyfriend; he refused to work and sent her out
had been totally reliant on her father since being an infant. She prostituting to pay the bills and buy food and drugs, she was
was not permitted to use a wheelchair and in the absence of beaten often.
external help, had no means of escaping the sexual violence he
subjected her to. “I sold my body to pay for drugs, I was assaulted by many
customers; it was a very bad time”.
Living With HIV
Intimate Partner Violence Andrea acquired HIV from one of the male customers.
As with other participants, women living with HIV suffered all
forms of IPV, in addition however, they also lived with the Freedom (Constraints to Escaping Violence)
constant fear that their partner would disclose their HIV Women living with HIV often experienced high levels of stigma
status. In a context in which HIV-AIDS stigma and and alienation from families and their communities. This
discrimination can adversely affect all aspects of social life compounded IPV in specific ways and as their partners also
(employment, friendships, leisure activities) not only for the exploited their circumstances, increased their vulnerability to
woman, but for her children, this was a powerful ongoing abuse. Some women chose to put up with the abuse,
psychological weapon which inflicted a great deal of harm and rather than facing the social trauma of entering into a new
fueled feelings of low self-worth. Agnes, a 56 year-old woman relationship and having to disclose their HIV status. Several
living with HIV believed she acquired the virus from her husband participants said that their partners threatened to disclose their
as he had a lot of sexual partners (although, to her knowledge, he status if they left and they considered that living with IPV was

better than facing the unknown consequences of HIV disclosure. she wondered whether this contributed to the poor relationship
Some women had lost jobs because someone had disclosed their choices she said she had made.
status, and the fear of not being able to provide for themselves and Some women recalled witnessing their mothers being sexually
their children meant that survival required they develop the abused when they were children and later, as adults, being
personal resilience to remain in the relationship. For these subjected to abuse in front of their own children.
women, the home, though a place of violence, paradoxically
provided a ‘safety-net’ against the risk of being homeless and “He started abusing me, physically he would beat and hit.
unemployed as a woman living with HIV who has children. For Then sexually where he would come home at night and
Andrea, it was not fear of others learning of her HIV status that rape me in front of our daughter”.
forced her to stay in an abusive relationship, but her addiction to
drugs. She had started using drugs as a way of dealing with the Interviewer: How old was she?
abuse and death of her two-old daughter:
“She was three and she would be up. It was like seeing a
“I smoked and smoked and smoked and smoked and cycle of my mother when I was a child and that really
blocked out everything’, all of the pain, the abuse”. hurt me”.
Andrea’s dependency upon crack cocaine was sustained by the Normalization of Sexual Abuse
fact that her partner was also using drugs. Andrea was one of the Sami and the other women who had experienced abuse in
few women in the study who had been able to access appropriate childhood questioned whether the environments in which they
support. She was sent to a drug rehabilitation facility (where she had grown up in had laid the groundwork for their victimisation.
was living at the time of the interview) and was working to create Debbie, sexually abused by an uncle, later found out her
an independent abuse-free life for herself. mother had also been abused by an uncle.
“Right now, I am a real survivor for real. Because after “It was a generational thing because his father did it to
16 years of abuse in every shape and form I am currently my mother and he continued on that trait”.
doing two courses, one in jewellery making and the other
one in auxiliary nursing”. Several of the women described living in households in
which male sexual entitlement was an accepted social norm. In
Sexual Abuse as Lifespan Trauma the absence of oppositional voices, the oftentimes complicity of
This section of the article draws only on the data of the 19 women families and communities and, the failure of child protective
who revealed that as well as facing IPV as adults, they had been services to recognize and act on abuse, these norms became
subjected to child sexual abuse. embedded within everyday life. Tracy, a 20 year old disabled
As in Andrea’s circumstances (described in the previous woman said she was in primary school when she was first
section), women who reported sexual violence as adults, in sexually abused by her father. With a severe mobility
almost all cases, also referred to historic child sexual abuse impairment, Tracy was carried everywhere by her father;
and for some, sexual violence had continued across the span she did not have a wheelchair. From her being very young,
of their lives. Cynthia, a 39 year old married, but separated her father assumed responsibility for her personal care needs.
woman was sexually abused by an adolescent boy who was a After the first incident of sexual abuse, this became a part of the
close friend of the family when she was 11. Debbie, a twenty-six- daily routine of “care” he provided. Tracy said she loved her
year-old single mother with two children also said that her father and did not see a problem with what he was doing. She
abusive experiences began in childhood; she was abused by also revealed that she had become pregnant with her father’s
her uncle when she was 10 years old. As with Cynthia and child and had been forced to have an abortion. Her response to
Debbie, several women questioned whether their later the interviewer on how she felt about this was “I am only
experiences of sexual violence had been in part “scripted” grateful. Can you imagine?” Physical and economic
because of the failure of adults to believe them when they dependency intersecting with patriarchal domination and an
were children and the ways in which the abuse was “brushed assumption that Tracy has no rights to her own body or
over”. This was the case for Sami, a 38 year old single mother of sexuality, had created a life of which, she could not imagine
eight. She recalled being molested by her uncle when she was existing outside. When a family member suggested Tracy
about six and then she was raped (vaginally and anally) at around should consider emigrating so that she could go to college
10 years of age by her sister’s boyfriend. When she was 14, Sami and get support to live independently. She retorted: “Home is
ran away with 20 year-old a man to whom she became pregnant. fine for me”.
Having experienced sexual molestation, rape and sexual Sula too, had grown up thinking that it was normal for men to
exploitation at different stages during her childhood, she had have sex with young girls.
gone on to experience sexual violence in every one of her adult
relationships. Sami believed that her early sexualisation and abuse “. . .we have a lot of grown men having sex with young
had led to a lack of self-worth and a view of herself as having no girls. It’s normal. I don’t know any young girls who did
value outside of being available to service men’s sexual desires; not have sex with men. It was normal”.

Her initiation into sexual submissiveness was to have a DISCUSSION

profound effect on her sexual behavior as an adult, leading to
her acquiring HIV. This article has explored the intersections of patriarchal
Most women who had witnessed abuse as children or who had domination, structural disadvantage and interpersonal violence
been victimized themselves, recognized the behavior as abusive among three groups of women in especially vulnerable
but in the face of societal acceptance, did not think they had the circumstances in Barbados and Grenada. The experience of
power to do anything about it. violence associated with being pregnant, disabled or, living
with HIV was unique to each individual woman however, the
“I realized it was abuse as soon as it started, but because I axes of differentiation that structured the women’s lives were
grew up in an abusive home, I thought that this is how it connected to each other because of systemic inequalities. In
should be. My mom went through it so I thought that I applying an intersectional lens, a common theme to emerge
had to do the same”. (Emma) from the findings was that women’s material circumstances
and the extent to which they were subject to control and
Some women recalled being oppositional and defiant and constraint was related to the discourses which surrounded
several reported abuse to others even though this sometimes their social status. Weaponizing of difference and disadvantage
resulted in them being disbelieved or being blamed. was a central feature of many women’s experiences and
perpetrators of violence often exploited a woman’s impairment
“. . . I remember plainly my first words were “mommy or situation in order to humiliate and exert control. Being
uncle M . . . interfere with me”. I could remember being “especially vulnerable” was synonymous with having one’s
called a [offensive ] and a [offensive ] and being told that agency as an independent, autonomous person constrained
I want killing, at two years old. So, if you at two years and suppressed. These constraints were impacted by physical
hear that, ok, things die down for a while then he started and social restrictions and through the interplay of patriarchal
back. If you tell somebody in the beginning and nobody power with institutional failings. Two agentic factors emerged
believed you who will you tell again?” from the findings: women’s economic dependency and, the
psychological dependency generated by ideologies of female
Aya lost her sight at the age of 11 and soon after, her father submission. At the time of the interviews, some of the
started sexually abusing her; she refused to keep quiet about it: “I participants had been able to generate a modest income
spoke to my grandmother, I spoke to my mother and, I spoke to through crafts or selling, however most had been financially
my teacher”. After two years, the abuse stopped, although Aya reliant upon their partners when they experienced violence
was made to feel responsible for what had happened. and this severely limited agency and opportunity for escape.
All of the women in the study bemoaned the lack of services
“I don’t know how to explain how it affected them to support victims of violence in acquiring the means to live
because they never even stood up for me. All of them independently. Women needed help to live; this, they considered,
were against me”. was the best way of helping them to leave.
The normalization of sexual abuse meant that there were

rarely any negative consequences for the men who perpetrated Patriarchy and Constraints on Leaving
it. Michelle had become pregnant as a result of coercive sex when Violent Relationships
she was 15 and railed against the fact that while her abuser faced In identifying the plurality of power relations in the experiences of
no consequences, she was ostracized and given no support. women in especially vulnerable circumstances, it was clear that cycles
of continuous and coterminous oppression and violence are very
“. . .you know as a teenage mother . . . you are so difficult to exit without external support. Implicit in patriarchal role
stigmatized and discriminated against and people tend expectations that were a feature of the lives of the participants, was the
to judge you when they do not know the circumstances of requirement for women to adopt subject positions that embody
how you became pregnant, so I faced all of that and my acquiescence and sexual availability. Normative discourses around
life started going downhill from then”. being pregnant, disabled or living with HIV and the symbolic
meanings they generated, created additional social barriers to
Michelle was also angry that she had been ignored when she first leaving violent relationships. So, for example, expectations
reported CSA as a young girl and believed that had help been available, concerning maternal roles were found to be predicated on
she may have been able to avoid the series of abusive relationships she essentializing views of motherhood (Jeremiah, 2006). If a woman
later encountered. She was particularly distressed that at 15, she had was pregnant and the father of her baby had not abandoned her, then
been expelled from school because she was pregnant and was not able she simply could not leave, regardless of abuse or any other threats she
to complete her education. She had nevertheless demonstrated faced; the role is considered immutable and so are its social
considerable resilience and now, as a mother, she was determined requirements (Jeremiah, 2006). The experiences of women
that though her children had been exposed to violence, she would do with disabilities, revealed that disability is often openly
all she could to make sure they would not be at risk of violence in their parodied in ways that served to diminish them. These
own lives. women described being the target of fun, pity, and

discrimination not only in their personal lives, but in their Schröttle and Glammeier (2013) suggest that the risk of IPV is
interactions with public agencies (the police, the courts, welfare increased two/three times when women have experienced violence
officials). This aligns with other research on the topic (e.g., in childhood since these early experiences can lead to low self-esteem
Brownridge, 2006; Breiding and Armour, 2015). Prevailing and undermine the ability to set boundaries in relationships. The
paternalistic attitudes undermined disabled women’s rights to relationship choices women in especially vulnerable circumstances
autonomy; treated as childlike and lacking intellectual capacity, make in such circumstances cannot be deemed wholly agentic. This
their accounts of violence were often regarded as non-credible is because 1) these choices are predicated on and situated within
and they were systematically denied access to the means of self- wider ideological norms which promote patriarchal values of
determination. Disabled women could not leave violent dominance, male sexual entitlement and female acquiescence
relationships because they were assumed not to be able to (Harris et al., 2005; Ahrens et al., 2010) and 2) the women
cope on their own and there were few services to enable this, making these choices are subject to material, psychological and
even in the face of the violence they suffered. Women living with structural constraints arising out of the intersections of gender and
HIV face extremes of stigma and discrimination spanning social status over which they have little control (Heise et al., 1998;
everyday microaggressions that communicate social Hale and Vazquez, 2011). Combined, these processes fuel the
contempt, right through to overt discrimination and violence prevalence and persistence of violence against women and limit
(Hale and Vazquez, 2011; Luciano, 2013). That hegemonic agency. The current study has focused on women who are pregnant,
perceptions associate HIV with deviance, contagion and disabled or, living with HIV, however the concept of “ties that bind
sanction was evident in the women’s accounts. Even though us” is an important one for all women. Greater recognition of sexual
the Caribbean has some excellent public education programmes abuse as lifespan trauma and sustained integrated action on violence
aimed at tackling HIV discrimination, women believed that prevention for children and women is therefore essential.
disclosure of their status would lead to them being shunned and
targeted and would generate huge problems for their children.
Women living with HIV cannot leave violent relationships STRENGTHS AND LIMITATIONS
because the fear of the consequences is simply too great.
The strengths of the current study include the foregrounding of
The “Ties That Bind Us”–Early differential experience of sexual violence within a unifying
intersectional framework which includes attention to patriarchy,
Sexualization, Child Sexual Abuse, Cultural agency and social circumstances. This has revealed unique insights
Normalcy and Sexual Violence Against into the specificities of what is a universal problem. Limitations of the
Women study lie in the sampling methods used, for example, only participants
In the present study, women’s experiences of IPV were who had experienced IPV were included and only women who were
exacerbated because of a high level of tolerance for in contact with a support agency were included. This means that the
violence against women and girls within their families and voices of women unable or unwilling to access services or, who may be
communities and, the acceptance of early sexualization. This prevented from accessing services have not been included in this
link between violence victimization in childhood and study; women who may be at even greater risk of violence than those
violence acceptance is confirmed through research on in contact with NGOs. Research is also needed among women who
attitudes and exposure to abuse among 1400 children in are pregnant, disabled and living with HIV and who are not subject to
Barbados and Grenada (Debowska et al., 2018) and studies IPV to determine the ways in which the intersectional framework can
on the prevalence of interpersonal violence (Le Franc et al., generate insights into lives not impacted by violence.
2008). Furthermore, Barrow and Ince (2008) demonstrate
that sexual exploitation is strongly associated with children
becoming sexually active at a young age. Other studies have DATA AVAILABILITY STATEMENT
also reported the association between social acceptance,
cultures of normalcy and child sexual exploitation in The datasets presented in this article are not readily available because
Caribbean settings (e.g., Kempadoo and Taitt, 2006; this dataset includes transcripts of qualitative interviews with women,
Barrow 2008). In the present study, 17 of the 19 several of whom were in violent relationships at the time of the
participants who recalled sexual victimization as children research; given the small size of the countries in which the study was
suggested that these earlier violations, set within a context conducted, there is a risk that some of the women may be identifiable
in which sexual relations between children and adults and this could jeopardize their safety. Requests to access the datasets
sometimes were normalized, may have contributed to the should be directed to a.d.jones@hud.ac.uk.
conditions for later IPV they faced (two participants did not
make this link). Described by one woman as “the ties that
bind us”, this author contends that the interrelationship ETHICS STATEMENT
between violence against children and violence against
women is a key finding, one which can inform policy and The studies involving human participants were reviewed and
the development of more cohesive services. This is supported approved by The University of Huddersfield, School of Human
by other studies (see, for example, Namy et al., 2017). and Health Sciences Research Ethics Committee. The patients/

participants provided their written informed consent to ACKNOWLEDGMENTS

participate in this study.
The author wishes to acknowledge colleagues who were not
involved in the preparation of this manuscript but assisted
AUTHOR CONTRIBUTIONS with the gathering of data and without whom the research
would not have been possible: Ena Trotman Jemmott and
AJ: grant capture, conceptualization, analysis, writing, and Hazel Da Breo and, Denise Tannis, Francia Best and Marissa
editing. Chandler.
FUNDING SUPPLEMENTARY MATERIAL

This research has been supported financially by the European The Supplementary Material for this article can be found online at:
Union (EuropeAid/136243/DD/ACT/Multi—Towards a Future https://www.frontiersin.org/articles/10.3389/fsoc.2021.623661/
Free From Domestic Violence). full#supplementary-material.
Crenshaw, K. (1989). Demarginalizing the Intersection of Race and Sex: A Black

REFERENCES Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and
Antiracist Politics. University of Chicago Legal Forum. Article 8. Available
Ahmed, S., Koenig, M. A., and Stephenson, R. (2006). Effects of Domestic Violence at: https://chicagounbound.uchicago.edu/uclf/vol1989/iss1/8.
on Perinatal and Early-Childhood Mortality: Evidence from North India. Am. Debowska, A., Boduszek, D., Sherretts, N., Willmott, D., and Jones, A. D. (2018).
J. Public Health 96 (8), 1423–1428. doi:10.2105/ajph.2005.066316 Profiles and Behavioral Consequences of Child Abuse Among Adolescent Girls
Ahrens, C. E., Rios-Mandel, L. C., Isas, L., and del Carmen Lopez, M. (2010). and Boys from Barbados and Grenada. Child. Abuse Neglect 79, 245–258.
Talking about Interpersonal Violence: Cultural Influences on Latinas’ doi:10.1016/j.chiabu.2018.02.018
Identification and Disclosure of Sexual Assault and Intimate Partner Einspahr, J. (2010). Structural Domination and Structural Freedom: a Feminist
Violence. Psychol. Trauma Theor. Res. Pract. Pol. 2 (4), 284–295. doi:10. Perspective. Fem. Rev. 94 (1), 1–19. doi:10.1057/fr.2009.40
1037/a0018605 Ellsberg, M., Jansen, H. A., Heise, L., Watts, C. H., and Garcia-Moreno, C. (2008).
Allen, C. F. (2011). Intersections between HIV/AIDS and Violence against Intimate Partner Violence and Women’s Physical and Mental Health in the
Women: Research to Develop Pilot Projects in Barbados and Dominica. WHO Multi-Country Study on Women’s Health and Domestic Violence: an
J. East. Caribbean Stud. 36 (4), 39–59. Retrieved from http://search. Observational Study. The Lancet 371 (9619), 1165–1172. doi:10.1016/s0140-
proquest.com/docview/1040723971?accountid11526. 6736(08)60522-x
Audi, C. A. F., Segall-Corrêa, A. M., Santiago, S. M., Andrade, M. d. G. G., and Franklin, C. A., and Kercher, G. A. (2012). The Intergenerational Transmission of
Pèrez-Escamila, R. (2008). Violência doméstica na gravidez: prevalência e Intimate Partner Violence: Differentiating Correlates in a Random Community
fatores associados. Rev. Saúde Pública 42 (5), 877–885. doi:10.1590/s0034- Sample. J. Fam. Viol. 27, 187–199. doi:10.1007/s10896-012-9419-3
89102008000500013 Guedes, A. (2012). Violence against Women in Latin America and the Caribbean: A
Barrow, C., and Ince, M. (2008). Working Papers in Early Childhood Development Comparative Analysis of Population-based Data from 12 Countries. Published
#47: Early Childhood in the Caribbean. The Hague, Netherlands: Bernard van 2011. http://www.igwg.org/igwg_media/guedes-gbv-lac.pdf.
Leer Foundation. Hague, G., Thiara, R., and Mullender, A. (2011). Disabled Women, Domestic
Barrow, C. (2008). Sexual Identity, HIV, and Adolescent Girls in Barbados. Soc. Violence and Social Care: The Risk of Isolation, Vulnerability and Neglect. Br.
Econ. Stud. 57 (2), 7–26. Retrieved April 21, 2021, from http://www.jstor.org/ J. Soc. Work 41 (1), 148–165. Retrieved April 21, 2021, from http://www.jstor.
stable/27866549. org/stable/43772532. doi:10.1093/bjsw/bcq057
Bombereau, G., and Allen, C. (2008). Social and Cultural Factors Driving the HIV Halcón, L., Blum, R. W., Beuhring, T., Pate, E., Campbell-Forrester, S., and
Epidemic in the Caribbean Literature Review. St. Augustine, Trinidad and Venema, A. (2003). Adolescent Health in the Caribbean: A Regional
Tobago: Caribbean Health Research Council. Portrait. Am. J. Public Health 93 (11), 1851–1857. https://doi.org/10.2105/
Braun, V., and Clarke, V. (2013). Successful Qualitative Research: A Practical Guide AJPH.93.11.1851. doi:10.2105/ajph.93.11.1851
for Beginners. London: SAGE Publications Ltd. Hale, F., and Vasquez, M. (2011). Violence against Women Living with HIV/
Breiding, M. J., and Armour, B. S. (2015). The Association between Disability and AIDS A Background Paper. Washington, DC: Development Connections
Intimate Partner Violence in the United States. Ann. Epidemiol. 25 (6), (DVCN). http://salamandertrust.net/wpcontent/uploads/2012/12/
455–457. Available at www.annalsofepidemiology.org/article/S1047-2797(15) VAPositiveWomenBkgrdPaperMarch2011.pdf.
00127-1/fulltext. doi:10.1016/j.annepidem.2015.03.017 Han, A., and Stewart, D. E. (2014). Maternal and Fetal Outcomes of Intimate Partner
Brownridge, D. A. (2006). Partner Violence against Women with Disabilities. Violence Associated with Pregnancy in the Latin American and Caribbean Region.
Violence Against Women 12 (9), 805–822. Available at http://journals.sagepub. Int. J. Gynecol. Obstet. 124 (1), 6–11. doi:10.1016/j.ijgo.2013.06.037
com/doi/abs/10.1177/1077801206292681. doi:10.1177/1077801206292681 Harris, R. J., Firestone, J. M., and Vega, W. A. (2005). The Interaction of Country of
Burch, R. L., and Gallup Jr., G. G. (2004). Pregnancy as a Stimulus for Domestic Origin, Acculturation, and Gender Role Ideology on Wife Abuse*. Social Sci. Q.
Violence. J. Fam. Violence 19 (4), 243–247. doi:10.1023/b:jofv.0000032634. 86 (2), 463–483. doi:10.1111/j.0038-4941.2005.00313.x
40840.48 Heise, L., and Garcia-Moreno, C. (2002). “Violence by Intimate Partners,” in World
Campbell, J., García-Moreno, C., and Sharps, P. (2004). Abuse during Pregnancy in Report on Violence and Health. Editors E. Krug, L. L. Dahlberg, J. A. Mercy,
Industrialized and Developing Countries. Violence Against Women 10 (7), A. B. Zwi, and R. Lozano (Geneva, Switzerland: World Health Organization),
770–789. doi:10.1177/1077801204265551 87–121.
Capaldi, D. M., Knoble, N. B., Shortt, J. W., and Kim, H. K. (2012). A Systematic Heise, L., Ellsberg, M., and Gottemoeller, M. (1999). Ending Violence against
Review of Risk Factors for Intimate Partner Violence. Partner Abuse 3 (2), Women. Popul. Rep. L. (11), 1–43. Baltimore: John Hopkins University School
231–280. doi:10.1891/1946-6560.3.2.231 of Public Health, Population Information Program.
Carastathis, A. (2016). Intersectionality: Origins, Contestations, Horizons. Lincoln: Hill Collins, P. (2000). Black Feminist Thought: Knowledge, Consciousness and the
University of Nebraska Press. doi:10.2307/j.ctt1fzhfz8 Politics of Empowerment. 2nd Edn. London: Routledge.

Hill Collins, P. (2019). Intersectionality as Critical Social Theory. Durham: Duke Namy, S., Carlson, C., O’Hara, K., Nakuti, J., Bukuluki, P., Lwanyaaga, J., et al. (2017).
University Press. doi:10.1215/9781478007098 Towards a Feminist Understanding of Intersecting Violence against Women and
Hughes, K., Bellis, M. A., Jones, L., Wood, S., Bates, G., Eckley, L., et al. (2012). Children in the Family. Soc. Sci. Med. 184, 40–48. doi:10.1016/j.socscimed.2017.04.042
Prevalence and Risk of Violence against Adults with Disabilities: A Systematic PAHO/WHO (2007). Perspectives in Health, Washington, DC: PAHO/WHO, 11.
Review and Meta-Analysis of Observational Studies. The Lancet 379 (9826), Schröttle, M., and Glammeier, S. (2013). Intimate Partner Violence against
1621–1629. Available at www.thelancet.com/journals/lancet/article/PIIS0140- Disabled Women as a Part of Widespread Victimization and Discrimination
6736(11)61851-5/fulltext. doi:10.1016/s0140-6736(11)61851-5 over the Lifetime: Evidence from a German Representative Study. Int. J. Conflict
Humphries, B. (2008). Social Work Research for Social Justice. Macmillan Violence 7 (2), 232–248. doi:10.4119/ijcv-3021
International Higher Education. Silverman, J. G., Decker, M. R., Reed, E., and Raj, A. (2006). Intimate Partner
Jeremiah, E. (2006). Motherhood to Mothering and beyond Maternity in Recent Violence Victimization Prior to and during Pregnancy Among Women
Feminist Thought. J. motherhood Initiat. Res. 8 (1/2), 21–33. Residing in 26 U.S. States: Associations with Maternal and Neonatal Health.
Jones, A., and Trotman-Jemmott, E. (2009). Perceptions of, Attitudes to and Am. J. Obstet. Gynecol. 195 (1), 140–148. doi:10.1016/j.ajog.2005.12.052
Opinions on Child Sexual Abuse in the Eastern Caribbean. Barbados: Sumi, C., Crenshaw, K., and McCall, L. (2013). Toward a Field of Intersectionality
UNICEF. https://www.unicef.org/infobycountry/files/Child_Sexual_Abuse_ Studies: Theory, Applications, and Praxis. Signs: Intersectionality: Theorizing
in_the_Eastern_Caribbean_Final_9_Nov.pdf. Power Empowering Theor. 38 (4), 785–810. doi:10.1086/669608
Jones, A. D., Da Breo, H., Trotman Jemmott, E., Joseph, D., and Möller, C. (2017). Taillieu, T. L., Brownridge, D. A., Tyler, K. A., Chan, K. L., Tiwari, A., and Santos, S.
Twenty-one Lessons: Preventing Domestic Violence in the Caribbean. C. (2015). Pregnancy and Intimate Partner Violence in Canada: A Comparison
Huddersfield: University of Huddersfield Press. of Victims Who Were and Were Not Abused during Pregnancy. J. Fam.
Jones, A. D., Trotman Jemmott, E., Maharaj, P. E., and Da Breo, H. (2014). An Violence 31, 567–579. doi:10.1086/669608
Integrated Model for Preventing Child Sexual Abuse: Perspectives from Latin UNAIDS/WHO (2007). “Situación de la epidemia del SIDA: Informe especial sobre
America and the Caribbean. Basingstoke, England: Palgrave Macmillan. doi:10. la prevención del VIH.” (Situation of the AIDS Epidemic: A Special Report on
1057/9781137377661 HIV Prevention). Geneva: UNAIDS/WHO. www.universia.net.co/HIV/AIDS/
Kempadoo, K., and Taitt, A. (2006). Gender, Sexuality and Implications for HIV/ view-document/document-648.html.
AIDS in the Caribbean: A Review of Literature and Programmes. Bridgetown, UNICEF (2006). Barbados Health and Family Education for Children Survey,
Barbados: UNIFEM and IDRC. https://caribbean.unwomen.org/en/materials/ Health and Family Education for Children Programme. Barbados: UNICEF
publications/2006/6/gender-sexuality-and-hiv-in-the-caribbean. Office for the Eastern Caribbean Area.
Kempadoo, K. (2004). Sexing the Caribbean: Gender, Race, and Sexual Labor. New UNICEF (2008). Barbados Health and Family Education for Children Survey,
York: Routledge, 272. Health and Family Education for Children Programme. Barbados: UNICEF
la Riviére Zijdel, L. (2004). “Disabled Women and Non–disabled Women, Strategies of Office for the Eastern Caribbean Area.
Action within the European Context.” Cited in European Parliament 2006. European UNICEF (2012). Sexual Violence Against Children in the Caribbean: Progress Report
Parliament Resolution on the Situation of People with Disabilities in the Enlarged 2012. Bridgetown, Barbados: UNICEF Office for the Eastern Caribbean Area.
European Union: The European Action Plan 2006−2007. United Nation (2007). The United Nations Office on Drugs and Crime and the Latin
Le Franc, E., Samms-Vaughan, M., Hambleton, I., Fox, K., and Brown, D. (2008). America and the Caribbean Region of the World Bank. Other Urban Study.
Interpersonal Violence in Three Caribbean Countries: Barbados, Jamaica, and https://caribbean.unfpa.org/en/news/gender-thematic-brief.
Trinidad and Tobago. Rev. Panam Salud Publica 24 (6), 409–421. doi:10.1590/ World Health Organization (2013). Global and Regional Estimates of Violence
S1020-49892008001200005 against Women: Prevalence and Health Effects of Intimate Partner Violence and
Luciano, D. (2013). Addressing Violence Against Women Living with HIV in Latin Non-partner Sexual Violence. Geneva: World Health Organization.
America and the Caribbean. Virginia: Rewire. https://rewire.news/article/2013/
05/19/addressing-violence-against-women-living-with-hiv-in-latin-america- Conflict of Interest: The author declares that the research was conducted in the
and-the-caribbean/. absence of any commercial or financial relationships that could be construed as a
McMahon, S., Huang, C-C., Boxer, P., and Postmus, J. L. (2011). The Impact of potential conflict of interest.
Emotional and Physical Violence during Pregnancy on Maternal and Child
Health at One Year Post-partum. Child. Youth Serv. Rev. 33 (11), 2103–2111. Copyright © 2021 Jones. This is an open-access article distributed under the terms of
doi:10.1016/j.childyouth.2011.06.001 the Creative Commons Attribution License (CC BY). The use, distribution or
Milletich, R. J., Kelley, M. L., Doane, A. N., and Pearson, M. R. (2010). Exposure to reproduction in other forums is permitted, provided the original author(s) and
Interparental Violence and Childhood Physical and Emotional Abuse as the copyright owner(s) are credited and that the original publication in this journal is
Related to Physical Aggression in Undergraduate Dating Relationships. cited, in accordance with accepted academic practice. No use, distribution or
J. Fam. Viol. 25 (7), 627–637. doi:10.1007/s10896-010-9319-3 reproduction is permitted which does not comply with these terms.

ORIGINAL RESEARCH
published: 09 July 2021
doi: 10.3389/fpsyg.2021.606774
Swedish and Norwegian Police

Interviewers’ Goals, Tactics, and
Emotions When Interviewing
Suspects of Child Sexual Abuse
Mikaela Magnusson 1*, Malin Joleby 1 , Timothy J. Luke 1 , Karl Ask 1 and
Marthe Lefsaker Sakrisvold 1,2
1
Department of Psychology, University of Gothenburg, Gothenburg, Sweden, 2 Research Department, Norwegian Police
University College, Oslo, Norway
As the suspect interview is one of the key elements of a police investigation, it has
received a great deal of merited attention from the scientific community. However,
suspect interviews in child sexual abuse (CSA) investigations is an understudied research
area. In the present mixed-methods study, we examine Swedish (n = 126) and
Norwegian (n = 52) police interviewers’ self-reported goals, tactics, and emotional
experiences when conducting interviews with suspected CSA offenders. The quantitative
analyses found associations between the interviewers’ self-reported goals, tactics,
and emotions during these types of suspect interviews. Interviewers who reported
Edited by:
Taina Laajasalo, experiencing more negative emotions were more likely to employ confrontational tactics.
National Institute for Health and Specifically, anger was positively associated with the goal of obtaining a confession
Welfare, Finland
and with aggressive tactics like raising one’s voice and emphasizing the seriousness
Reviewed by:
Pekka Santtila,
of the crime. Frustration and disgust displayed similar patterns. Somewhat contrasting
New York University Shanghai, China these quantitative results, the thematic analysis identified a strong consensus that
Noora Ellonen, emotions should not and do not affect the police interviewers’ work. Furthermore, the
University of Tampere, Finland
police interviewers described a range of strategies for managing emotions during the
*Correspondence:
Mikaela Magnusson interview and for processing their emotional reactions afterwards. The present findings
mikaela.magnusson@psy.gu.se highlight the relevance of emotional processes in CSA suspect interviews and provide an
initial exploration of the potentially complex relationship between the goals, tactics, and
Specialty section:
emotional experiences of police interviewers who question CSA suspects.
Keywords: police, interrogation, suspect, child sexual abuse, Norway, Sweden, investigative interview, emotion
Received: 15 September 2020 INTRODUCTION

Accepted: 04 June 2021
Published: 09 July 2021 Criminal investigations concerning alleged child sexual abuse (CSA) presents a challenge for legal
Citation: practitioners worldwide. As corroborative evidence is often scarce, the statements provided by
Magnusson M, Joleby M, Luke TJ, child witnesses and suspects typically constitute the primary sources of information (Ernberg
Ask K and Lefsaker Sakrisvold M
et al., 2018). During the last four decades, a considerable amount of research has increased our
(2021) Swedish and Norwegian Police
Interviewers’ Goals, Tactics, and
knowledge of forensic child interviewing during CSA investigations (see e.g., Brubacher et al.,
Emotions When Interviewing Suspects 2019). Likewise, a large body of psychological research on suspect interviewing has developed over
of Child Sexual Abuse. the last several decades. Much of this literature is focused on risk factors for false confessions
Front. Psychol. 12:606774. (Kassin et al., 2010). Other parts of the literature have been aimed at developing interviewing
doi: 10.3389/fpsyg.2021.606774 methods that increase the amount of accurate information obtained while questioning suspects

1 July 2021 | Volume 12 | Article 606774
Magnusson et al. Interviewing Suspects of Child Sexual Abuse
(Mac Giolla and Granhag, 2017). This latter part of the literature confrontational, sex offenders described experiencing less open-
has demonstrated that non-accusatory techniques that are minded attitudes, cooperation, and more coercion than did
focused on gathering information, rather than obtaining murderers. A majority of the sex offenders experienced stress
confessions, are more effective than confrontational, accusatory during the interview, and significantly more sex offenders
techniques (e.g., minimization; Meissner et al., 2014; Luke described a lack of being respected compared to the murderers.
and Alceste, 2020). However, research is scarce regarding Moreover, police interviews marked with dominance (compared
practitioners’ experiences of interviewing suspected CSA to interviews marked with humanity) were associated with a
offenders. In the present study, we aimed to address this gap higher proportion of denials. However, it is unclear whether a
by examining Swedish and Norwegian police interviewers’ self- dominant approach led to denials, or if denying resulted in a
reported goals, tactics, and emotions when questioning suspects negative response by the interviewer. On a similar note, a study
of sexual abuse against a child. Given the low prosecution based on interviews with convicted sex offenders in Australia
rate for CSA offenses, there is an urgency for research reported that sex offenders perceived ethical interviewing and
exploring the current procedures used when investigating displays of humanity to increase the likelihood of a confession
these crimes, and their accordance with evidence-based methods (Kebbell et al., 2010). Despite the findings described above, the
(Ernberg et al., 2020). use of empathy in police interviews with suspects of sexual abuse
The lack of research on how to interview suspected CSA has shown no direct effect on the amount of information elicited
offenders poses numerous problems. First, investigating cases (Oxburgh and Ost, 2011).
of CSA can be challenging, as it includes being exposed to Although we lack data to draw firm conclusions, it is possible
potentially traumatic material or information (Burns et al., 2008; that the type of crime influences the way the interviewer conducts
Powell et al., 2014). This may cause these investigations to the suspect interview, which in turn might affect the likelihood
be emotionally demanding and can even lead to secondary of eliciting information. This raises the question of how police
traumatic stress and burnout among professionals (Perez et al., interviewers are affected by the nature of the crime, and how
2010; Bourke and Craun, 2014; Turgoose et al., 2017; Seigfried- their emotions potentially affect their goals and tactics. Many
Spellar, 2018). Second, perpetrators of CSA differ from non- emotional demands are put on law enforcement and extensive
sex offenders in numerous ways (e.g., more likely to suffer research has examined the causes and consequences of stress
from mental illness, anxiety, low self-esteem, and have greater and traumatic work experiences (e.g., Toch, 2002; Perez et al.,
social deficits, but less likely to have lifestyle instability and 2010; Seigfried-Spellar, 2018). The focus has mainly been on how
antisocial disorder; Whitaker et al., 2008), which may impact work-related stressors impact the psychological health of police
on the suitability of different interview tactics. Third, convicted officers. Thus, we have identified some important aspects that
CSA perpetrators have reported a strong internal need to have received limited attention, namely if and how emotions
confess to their crimes, compared to general violent offenders influence how police interviewers perform their work. This is of
or those convicted of raping adults (Gudjonsson and Sigurdsson, particular interest in cases of CSA, as these investigations have
2000). Yet, in Swedish court cases involving CSA, only 31% of been shown to be perceived as especially distressing (Huey and
defendants admitted guilt during their trial (Magnusson et al., Kalyal, 2017).
2018). According to the Pathways model (Ward and Siegert, Emotions can influence human behavior and decision making
2002), there are four psychological mechanisms involved in in a number of ways that may impact the conduct of
sexual offending against children. One of the mechanisms— suspect interviews. First, specific emotions are associated with
cognitive distortions—may constitute an obstacle for offenders cognitive appraisals (Scherer and Moors, 2019) that may shape
to recognize that they have committed a crime. The interview interviewers’ perception of guilt and responsibility. For instance,
techniques used by police need to challenge any distorted anger can increase the perception of criminal intent behind
excuses, encourage offenders to admit to their wrongdoings, ambiguous actions and promote punitiveness (Ask and Pina,
while at the same time making sure to minimize the risk 2011) and may increase the attribution of guilt to a suspect under
of obtaining false confessions. Establishing a respectful and investigation (Sambrano et al., 2020). Second, emotional states
supportive environment is critical to increase the suspect’s influence the type of information processing that an individual
confidence to engage in the process and provide an honest and is likely to engage in. For instance, it has been shown that
detailed response (Read et al., 2009). police investigators who experience anger tend to process case
The use of different interviewing tactics might vary depending materials more superficially, and thus rely more on heuristics,
on the type of crime. A study examining 59 police interviews than colleagues who experience sadness (Ask and Granhag,
from the UK (Oxburgh et al., 2014) found that suspects of 2007). Third, affective states have been found to influence
CSA were asked significantly more inappropriate questions people’s strategies in social interactions. For instance, people
compared to suspects of child or adult murder. Furthermore, in a happy (vs. sad) state tend to rely on more direct, less
a Swedish study (Holmberg and Christianson, 2002) that elaborate strategies when spontaneously interacting with, making
compared self-reported interview experiences between convicted requests from, negotiating with, and trying to persuade others
sex offenders (including offenses against both children and (Forgas, 2002). Relatedly, specific emotions are associated with
adults) and murderers found several differences between distinct action tendencies, which facilitate behavior congruent
crime types. Although only a minority of murderers and sex with the current emotional state (Frijda et al., 1989). For
offenders described their interviewers as aggressive or explicitly instance, anger is associated with the tendency to approach and

confront the source of the emotion (Carver and Harmon-Jones, crimes point toward the use of both information-gathering
2009), whereas fear is associated with avoidance and withdrawal and more confrontational methods among Swedish police
(Lerner and Keltner, 2001). interviewers (Granhag and Magnusson, 2017; Mac Giolla and
Studies that directly examine the relationship between Granhag, 2017). Similar to Norway, Sweden has not implemented
emotional experiences and interviewer behavior are scarce. One any specialized interviewing guidelines for questioning suspected
notable example, highly relevant to the present study, is the recent CSA offenders. However, police interviewers can take a course
study by Sambrano et al. (2020). They experimentally induced on investigating crimes against children which include some
specific emotions (happiness, sadness, and anger) and observed training in interviewing suspects (E. Norrman, course leader,
their influence on mock detectives’ preference for interrogation personal communication, September 9, 2020).
tactics in a hypothetical crime scenario. Across two experiments, The present mixed-methods survey aimed to descriptively
while all groups showed a preference for benevolent (empathetic) examine Swedish and Norwegian police interviewers’ self-
over hostile (guilt-confirming) tactics, this preference was more reported interview goals, tactics, and emotions when questioning
pronounced among sad participants than among angry and suspects in CSA cases. Using a series of quantitative analyses,
happy participants. This finding is consistent with the theoretical we aimed at examining the interrelationships between interview
assumption that emotions associated with certainty appraisals goals, tactics, and experienced emotions. Due to differences
(e.g., anger, happiness) promote confirmation-seeking behaviors, between countries with regard to the implementation of a
whereas uncertainty-related emotions (e.g., sadness) promote research-based investigative interview model, we were also
more elaborate, open-minded thinking, and action (Tiedens and interested in comparing the reports of goals, tactics, and
Linton, 2001; Lerner and Tiedens, 2006). In sum, these results emotions given by Swedish and Norwegian police interviewers.
raise the possibility that the behavior of police interviewers Lastly, we aimed to qualitatively explore how practitioners report
questioning CSA suspects may be influenced by the emotions managing different types of emotions that may arise during
they experience in response to the nature and content of the case. interviews with CSA suspects.
The current data collection took place in Sweden and Norway.
Although the neighboring countries have much in common MATERIALS AND METHODS
with regard to their societal structures, legislations, and criminal
investigation practices, there are some key differences in terms of The study was pre-registered on the Open
their focus on research-based interviewing (Fahsing et al., 2016). Science Framework (https://osf.io/jtac5/?view_only=
In Norway, a shift toward more research-based police practices e3b14c4113144dd3a857448c4ec6b913). The registration
began after a number of judicial scandals involving coercive included a recruitment strategy and plan for exploratory
interrogation techniques. In 2002, the Police University College analyses. We followed the recruitment plan as specified,
commissioned a national model for investigative interviewing but we have deviated from the exploratory analysis plan in
(KREATIV; Fahsing and Rachlew, 2009) that was heavily order to simplify the presentation of results and to address
influenced by the PEACE model from England and Wales (Milne reviewer comments.
and Bull, 1999). The KREATIV training program focuses on
teaching research-based information-gathering techniques for Participants and Recruitment
interviewing victims, witnesses, and suspects (Fahsing et al., A total of 127 Swedish police interviewers responded to the
2016). The program is now an integral part of the basic survey. However, one participant did not meet the current
education to all Norwegian police officers. In addition, the inclusion criterion of having experience conducting suspect
Norwegian Police University College offers specialized courses in interviews during the last 5 years. The final Swedish sample thus
investigative interviewing of adults and the handling of sexual consisted of 126 participants (105 women, 21 men, M age = 45.2
crimes investigations which include some interview training years, SD = 9.1), with 101 police officers (2 years of training at
(Detective superintendent C. Tombre, personal communication, the police academy) and 24 civilian officers (academic degree in
August 30, 2019). However, there are no specific guidelines for other relevant fields, for example, criminology or psychology).
suspect interviews in CSA cases. While substantial organizational The interviewers had between 0 and 25 years (M = 7.0, SD
changes have occurred through the implementation of KREATIV, = 5.6) of experience conducting interviews with CSA suspects.
studies are lacking regarding the quality of Norwegian interview The majority (80.2%) had finished the course on investigating
practices following the reform. crimes against children which includes training on interviewing
Unlike Norway, Sweden has not implemented a national suspects. The Swedish survey was distributed during a national
interview model that all police interviewers are required to conference for specialized child interviewers, with a response
follow. However, Swedish police interviewers are by law not rate of 79% of the attendees. According to the Prosecution
allowed to use threats, coercive techniques including sleep Development Centre Gothenburg (2016), there are around 305
or food deprivation, promises of leniency, or to deceive the active child interviewers in Sweden. Based on this estimate, our
suspect (the Swedish Code of Judicial Procedure, chap 23, §6). sample would represent 41.3% of the current population of child
Furthermore, the police academies across Sweden focus primarily interviewers in Sweden.
on information-gathering techniques and advise against using With regard to the Norwegian data collection, a total of
accusatory methods (Granhag et al., 2013). Field and laboratory 52 police interviewers participated in the survey (45 women,
research from suspect interviews involving a variety of different 7 men, M age = 40.0 years, SD = 8.02). All except three

participants, who chose not to report their employment status, indicate how often they used each tactic. The Likert scales ranged
had a background as police officers (this requires a bachelor’s from (1) Never to (5) Always. The order in which the tactics
degree from the Police University College, or equivalent training were presented was reverse counterbalanced. The fourth section
before the current training program was implemented). The consisted of a list of interview goals formulated as statements (see
Norwegian participants had between 0 and 24 years (M = 8.9, Table 2) adapted from a previous police survey. The respondents
SD = 6.5) of experience interviewing suspects in CSA cases. All were asked to indicate to what extent they agreed or disagreed
participants except one (98.1%) had completed the KREATIV with each of the goals. The scale ranged from (1) Do not agree at
training program, and 40.3% had completed the training course all to (5) Completely agree. The order in which the different goals
on investigating sexual crimes which includes some suspect were presented was reverse counterbalanced.
interview training. Due to varied recruitment methods, we were In the fifth section, the respondents were asked to assess how
unable to calculate a response rate for the Norwegian sample. often they found themselves in emotionally demanding situations
during their work with interviewing CSA suspects (1 = Never,
5 = Always), and to what extent their work with interviewing
Survey
CSA suspects is emotionally demanding (1 = To a very low
The survey consisted of six different sections. First, the survey
extent, 5 = To a very high extent). The respondents were also
included a cover page with information about the research
asked to describe how often they experienced a range of different
project and the participants’ ethical rights (e.g., participation is
emotions during their interviews with CSA suspects. The scale
voluntary and can be withdrawn at any time). All participants
was adapted from Ask and Pina (2011), and the scale step for each
were asked to provide their informed consent to take part in
item ranged from (1) Never, to (5) Always. The order in which
the survey. The cover page also included a screening question
the emotions were presented was reverse counterbalanced, with
asking if they had experience conducting interviews with CSA
one-half of the participants rating the emotions in reversed order.
suspects during the last 5 years. Participants who did not have
Lastly, the respondents were asked an open-ended question about
this experience were informed that they were not eligible for
how they manage different emotions that may arise inside them
participation and thanked for their time. The second section
during a suspect interview in CSA cases.
consisted of demographic questions about the participants’
The survey was created in Swedish and translated to
experience and previous training in forensic interviewing.
Norwegian using forward and back-translations. One of the
The third section comprised a list of different interview
authors, a native Norwegian researcher fluent in Swedish,
tactics for questioning CSA suspects adapted from the current
initially translated the survey in consultation with another
research literature (see Table 1). The participants were asked to
researcher fluent in both Norwegian and Swedish. After the
survey was translated, a third researcher fluent in both languages
TABLE 1 | List of different interview tactics.
reviewed the translations, and all three agreed upon the final
version of the translation. The Norwegian survey was thereafter
Tactic Survey statement translated back to Swedish by a native Swedish speaker fluent
in Norwegian. Some minor changes were made during this
Treat friendly You treat the suspect in a friendly way
stage of the process. Both versions of the survey were pre-
Raise voice You raise your voice during the conversation with tested prior to data collection for clarity and terminology. The
the suspect
demographic questions regarding training were based on the
Care about feelings You show that you care about the suspects feelings
course curriculums offered in each country.
Interrupt You interrupt the suspect while he/she is talking
Common interests You try to find common interests to talk to the
suspect about
Procedure
Boss You show who’s boss
Design
We used a QUAN+qual embedded mixed-methods design
Show respect You show respect for the suspect as a human being
(Creswell and Clark, 2017) in which quantitative and qualitative
Confront with info You confront the suspect with information about the
case
Emphasize seriousness You put pressure by emphasizing the seriousness of
TABLE 2 | List of different interview goals.
the situation
Let suspect talk You let the suspect finish talking even if he/she talks Goals Survey statement
about irrelevant details
Active listening You actively listen to what the suspect says Confession Receiving a confession from the suspect is a high
Show what you think You show the suspect what you think about the priority
criminal act Full report Receiving the suspect’s full report is a high priority
Unanticipated You ask unanticipated questions Unknown information Receiving previously unknown information during an
questions interrogation is a high priority
Claim suspect has You claim to know that the suspect has more Explanation Receiving the suspect’s explanations regarding
more info information then he/she tells you existing investigative information is a high priority
Statements as presented in the survey. Participants indicated how often they used each Statements as presented in the survey. Participants indicated to what extent they agreed
tactic on a Likert scale ranging from Never (1) to Always (5). to each goal on a Likert scale ranging from Do not agree (1) to Completely agree (5).

data answered different research questions, were collected at thematic structure. To assess the inter-rater reliability of the
the same occasion, analyzed separately, and then merged. In coding system, a research assistant coded 20% of the data
this study, quantitative data (e.g., Likert scale responses) from following the thematic structure. The inter-rater reliability,
the survey were used to investigate Swedish and Norwegian estimated using Gwet’s AC1 for each subtheme, showed very
police interviewers’ goals, tactics and emotions, and the potential good agreement (M = 0.954, range = 0.936–0.964). Lastly,
relationship between variables. The qualitative data (open- quotations to illustrate the qualitative process were selected,
ended responses) from the survey were used to explore edited to facilitate reading, and translated to English.
how police interviewers manage emotions that arise during
suspect interviews.
RESULTS
Data Collection Quantitative Analysis: Self-Reported
Verbal information and a pen-and-paper version of the survey Interview Goals, Tactics, and Emotions
was presented to approximately 160 Swedish practitioners Interview Goals
during a national conference for specialized child interviewers Swedish police interviewers reported prioritizing obtaining a
in March 2019. The same recruitment method was used confession (M = 2.61, SD = 1.11) to a greater extent than
during a Norwegian national conference in October 2019 Norwegian interviewers (M = 2.09, SD = 0.98), d = 0.47 [0.14,
for approximately 100 specialized child interviewers. However, 0.79]. Interviewers in the Swedish (M = 4.50, SD = 0.76) and
unlike in Sweden, the majority of the Norwegian child Norwegian (M = 4.69, SD = 0.47) samples reported prioritizing
interviewers did not carry out interviews with CSA suspects. Due to a similar extent getting a complete account from the suspect,
to this unforeseen circumstance, we carried out a second data d = −0.28 [−0.60, 0.05]. Norwegian interviewers reported
collection with Norwegian practitioners between March and June prioritizing obtaining previously unknown information (M =
2020. For practical reasons following the coronavirus outbreak, 4.58, SD = 0.64) to a greater extent than Swedish interviewers
we used an online version of the survey. Invitations to participate (M = 4.06, SD = 0.97), d = −0.56 [−0.89, −0.23]. Swedish
and a link to the survey was sent out to five large police districts interviewers reported prioritizing getting an explanation for
across Norway. The survey was also distributed on social media existing investigative information (M = 4.60, SD = 0.72) to a
via Linked-In, Twitter, and Facebook. The participants did not greater extent than Norwegian interviewers (M = 4.21, SD =
receive any compensation for taking part in the survey. 0.82), d = 0.49 [0.17, 0.82].
Analyses Interview Tactics

Quantitative Analyses Table 3 displays descriptive statistics for each sample’s self-
For the quantitative analyses, we combined the data from the reported interview tactics, as well as standardized effect sizes
Swedish and Norwegian samples (N = 178). We analyzed police estimating the differences between each sample for each tactic.
interviewers’ responses to the quantitative items primarily using Both the Swedish and Norwegian samples reported using
bivariate correlations, to examine the relationships between nearly all of the listed tactics with some regularity. All tactics
goals, tactics, and emotions. Using the emotion adjectives for were reported with mean and median frequencies of use
which participants provided endorsement ratings, we created five above the lower endpoint of the scale (i.e., 1), with the
composites based on the conceptual groupings of the emotion exception of “showing the suspect what you think about the
words: anger (angry, mad; α = 0.81; all reliability coefficients crime” (Mdn = 1).
calculated using polychoric correlations), disgust (disgusted, Some notable differences between the Swedish and Norwegian
sickened; α = 0.66), interest (engaged, interested; α = 0.76), fear samples were evident. For example, Swedish interviewers
(nervous, scared, worried; α = 0.82), sadness (sad, sorrowful; reported raising their voice, d = 0.64 [0.31, 0.97], and
α = 0.60), and frustration (frustrated, irritated; α = 0.62). emphasizing the seriousness of the offense, d = 0.62 [0.29, 0.94],
Additionally, we examined the extent to which interviewers more frequently than Norwegian interviewers, and Swedish
from the Swedish and Norwegian samples differed in terms of interviewers reported approaching the suspect in a friendly
their self-reported goals, tactics, and emotions by calculating manner less frequently than Norwegian interviewers, d = −0.50
standardized mean differences (d) between the samples. [−0.82, −0.17]. Norwegian interviewers also reported finding
common interests with the suspect, d = −0.65 [−0.97, −0.32],
Qualitative Analyses and saying that the suspect had more information than they
For the qualitative analyses, we combined the Swedish (N = were providing, d = −0.56 [−0.89, −0.23], more frequently than
120) and Norwiegan (N = 33) police interviewers’ responses to Swedish interviewers.
the open-ended question about strategies for managing emotions
during CSA suspect interviews. The responses were analyzed Emotions
using data-driven thematic analysis (Braun and Clarke, 2006). Table 4 displays descriptive statistics for each sample’s self-
Initially, two of the authors separately created code labels reported emotions, as well as standardized effect sizes
that closely matched the content of the written responses. All estimating the differences between each sample for each
codes were cross-compared and merged to a thematic structure emotion composite. Both the Swedish and Norwegian
including themes and sub-themes. Disagreements were resolved samples reported experiencing nearly all the listed emotions
through discussion, and the data was re-coded into the final with some regularity. All emotion composites had means

TABLE 3 | Descriptive statistics and sample differences for reported interview tactic usage.
Sweden Norway
Strategy M SD Mdn Range M SD Mdn Range d[95% CI]
Treat friendly 4.83 0.37 5 4, 5 5.00 0.00 5 5, 5 −0.50 [−0.82, −0.17]

Raise voice 1.95 0.63 2 1, 3 1.55 0.54 2 1, 3 0.64 [0.31, 0.97]
Care about feelings 3.86 0.86 4 2, 5 3.90 0.63 4 3, 5 −0.05 [−0.38, 0.27]
Interrupt 2.16 0.54 2 1, 4 2.12 0.38 2 1, 3 0.09 [−0.24, 0.41]
Common interests 2.58 1.13 3 1, 5 3.35 1.14 3 1, 5 −0.65 [−0.97, −0.32]
Boss 3.22 1.10 3 1, 5 3.28 1.07 3 1, 5 −0.05 [−0.38, 0.27]
Show respect 4.89 0.44 5 1, 5 5.00 0.00 5 5, 5 −0.30 [−0.62, 0.03]
Confront with info 4.10 0.81 4 1, 5 4.35 0.62 4 3, 5 −0.32 [−0.65, 0.00]
Emphasize seriousness 2.98 0.86 3 1, 5 2.42 0.89 2 1, 4 0.62 [0.29, 0.94]
Let suspect talk 3.68 0.70 4 1, 5 3.46 0.64 3 2, 5 0.32 [0.00, 0.65]
Active listening 4.88 0.33 5 4, 5 4.85 0.36 5 4, 5 0.10 [−0.22, 0.43]
Show what you think 1.37 0.58 1 1, 3 1.46 0.58 1 1, 3 −0.15 [−0.48, 0.17]
Unanticipated questions 3.15 0.75 3 1, 5 3.06 0.72 3 1, 4 0.12 [−0.21, 0.44]
Claim suspect has more info 2.10 0.92 2 1, 5 2.77 1.62 2 1, 5 −0.56 [−0.89, −0.23]
n = 126 for Sweden and n = 52 for Norway. M = mean, SD = standard deviation, Mdn = median, d[95% CI] = standardized mean difference with 95% confidence interval.
TABLE 4 | Descriptive statistics and sample differences for reported experienced emotions.
Sweden Norway
M SD Mdn Range M SD Mdn Range d [95% CI]
Anger 2.23 0.74 2.00 1.00, 4.00 2.03 0.75 2.00 1.00, 3.50 0.26 [−0.07, 0.58]
Disgust 2.27 0.81 2.00 1.00, 4.50 1.86 0.53 2.00 1.00, 3.00 0.54 [0.22, 0.87]
Sadness 2.20 0.76 2.00 1.00, 4.00 2.09 0.80 2.00 1.00, 4.00 0.14 [−0.19, 0.47]
Fear 1.75 0.56 1.67 1.00, 4.00 1.80 0.61 1.67 1.00, 3.00 −0.09 [−0.42, 0.23]
Interest 4.38 0.60 4.50 1.00, 5.00 4.11 0.61 4.00 2.00, 5.00 0.44 [0.11, 0.77]
Frustration 2.75 0.53 3.00 1.00, 4.00 2.68 0.66 3.00 1.00, 4.50 0.11 [−0.21, 0.44]
n = 126 for Sweden and n = 52 for Norway. M = mean, SD = standard deviation, Mdn = median, d [95% CI] = standardized mean difference with 95% confidence interval.
and medians above the lower endpoint of the scale. correlation with the four goals. The goal of obtaining a confession
Swedish interviewers reported to more frequently feel was significantly positively correlated with raising one’s voice,
interested, d = 0.44 [0.11, 0.77], and disgusted, d = emphasizing the seriousness of the offense, showing the suspect
0.54 [0.22, 0.87], during suspect interviews, compared to what you think about the crime, and claiming that the suspect
Norwegian interviewers. has more information. The goal of obtaining a full report
Overall, participants on average reported that they was significantly positively correlated with confronting the
experienced emotionally demanding situations in their work suspect with information and engaging in active listening and
rarely or sometimes (M = 2.62, SD = 0.80). Additionally, significantly negatively correlated with interrupting the suspect.
Swedish police interviewers (M = 2.73, SD = 0.77) reported The goal of obtaining previously unknown information was
being more frequently in emotionally demanding situations significantly positively correlated with finding common interests
compared to Norwegian interviewers (M = 2.37, SD = 0.82), d with the suspect, confronting the suspect with information,
= 0.45 [0.12, 0.78]. On average, participants reported that they and with claiming the suspect has more information. The goal
found their work emotionally demanding sometimes (M = 3.08, of obtaining an explanation for the offense was significantly
SD = 1.05). However, Swedish police interviewers (M = 2.94, SD positively correlated with showing the suspect who is boss and
= 0.98) reported that their work is emotionally demanding to a confronting the suspect with information.
lesser extent compared to Norwegian interviewers (M = 3.40, Table 6 displays the correlations between the six emotion
SD = 1.14), d = −0.43 [−0.76, −0.11]. composites and four self-reported goals for suspect interviews.
As can be seen, self-reported experiences of anger and
Correlations Between Goals, Tactics, and Emotions frustration were significantly positively correlated with the
Table 5 displays the correlations between interview goals and goal of obtaining a confession. Additionally, experiences of
self-reported use of tactics. Several tactics have a significant disgust were negatively correlated with the goal of obtaining

TABLE 5 | Bivariate correlations and 95% confidence intervals of interview tactics and goals.
Confession Full report Unknown information Explanation
Treat friendly −0.038 [−0.184, 0.109] 0.103 [−0.045, 0.246] −0.018 [−0.164, 0.129] −0.115 [−0.258, 0.032]
Raise voice 0.173 [0.027, 0.312] −0.062 [−0.207, 0.085] −0.140 [−0.281, 0.006] 0.011 [−0.136, 0.157]
Care about feelings −0.023 [−0.169, 0.124] 0.126 [−0.021, 0.268] 0.030 [−0.117, 0.176] 0.073 [−0.075, 0.217]
Interrupt 0.042 [−0.106, 0.187] –0.154 [–0.294, –0.008] −0.010 [−0.156, 0.137] 0.008 [−0.139, 0.154]
Common interests −0.043 [−0.189, 0.104] 0.045 [−0.103, 0.190] 0.170 [0.024, 0.309] −0.037 [−0.183, 0.110]
Boss 0.136 [−0.011, 0.277] 0.034 [−0.114, 0.179] 0.032 [−0.115, 0.178] 0.158 [0.012, 0.298]
Show respect −0.022 [−0.168, 0.125] 0.105 [−0.043, 0.248] 0.033 [−0.114, 0.179] −0.024 [−0.170, 0.123]
Confront with info 0.075 [−0.073, 0.219] 0.156 [0.010, 0.296] 0.170 [0.024, 0.309] 0.265 [0.123, 0.396]
Emphasize seriousness 0.328 [0.191, 0.453] −0.058 [−0.203, 0.089] 0.058 [−0.089, 0.203] 0.102 [−0.045, 0.245]
Let suspect talk 0.023 [−0.124, 0.170] 0.117 [−0.030, 0.259] −0.013 [−0.159, 0.134] 0.041 [−0.107, 0.186]
Active listening −0.023 [−0.169, 0.125] 0.166 [0.019, 0.305] −0.075 [−0.219, 0.072] −0.019 [−0.165, 0.128]
Show what you think 0.166 [0.020, 0.305] −0.092 [−0.236, 0.055] 0.084 [−0.063, 0.228] 0.021 [−0.126, 0.168]
Unanticipated questions 0.114 [−0.034, 0.256] 0.026 [−0.121, 0.172] −0.015 [−0.161, 0.132] 0.009 [−0.138, 0.155]
Claim suspect has more info 0.164 [0.017, 0.303] 0.103 [−0.045, 0.246] 0.268 [0.126, 0.399] 0.066 [−0.082, 0.210]
Correlations whose 95% CI excludes 0 are displayed in boldface.
TABLE 6 | Bivariate correlations and 95% confidence intervals of emotion composites and interview goals.
Anger Disgust Sadness Fear Interest Frustration
Confession 0.231 [0.087, 0.365] 0.128 [−0.019, 0.270] 0.095 [−0.053, 0.238] −0.091 [−0.235, 0.056] 0.075 [−0.072, 0.220] 0.171 [0.025, 0.310]
Full report 0.058 [−0.089, 0.203] −0.084 [−0.228, 0.063] 0.033 [−0.114, 0.179] 0.080 [−0.068, 0.224] 0.098 [−0.049, 0.242] 0.100 [−0.047, 0.243]
Unknown −0.033 [−0.178, 0.115] –0.150 [–0.291, –0.004] 0.030 [−0.117, 0.176] 0.009 [−0.138, 0.155] −0.031 [−0.177, 0.116] −0.002 [−0.149, 0.145]
information
Explanation 0.079 [−0.069, 0.223] 0.069 [−0.078, 0.214] 0.093 [−0.054, 0.237] 0.019 [−0.128, 0.166] 0.197 [0.052, 0.334] −0.001 [0.148, 0.146]
previously unknown information, and experiences of interest do you manage different emotions that may arise within you
were positively correlated with the goal of obtaining an in connection to an interview of a CSA suspect?” The thematic
explanation for the offense. analysis revealed that a majority of police interviewers expressed
Table 7 displays the correlations between the six emotion that their emotions did not belong in the interview room, which
composites and the self-reported use of interview tactics. is captured in the main theme Putting emotions aside during the
Self-reported experiences of anger were significantly positively interview. Furthermore, most police interviewers described how
correlated with raising one’s voice, emphasizing the seriousness they tried to manage their emotions in different ways outside of
of the crime, and asking unanticipated questions, and anger the interview room, which is captured in the main theme Coping
was significantly negatively correlated with expressing care for with emotions afterwards. Each theme had three subthemes,
the suspect’s feelings. Experiences of disgust were positively see Table 8.
correlated with raising one’s voice, emphasizing the seriousness
of the crime, asking unanticipated questions, and disgust was Putting Emotions Aside During the Interview
negatively correlated with treating the suspect in a friendly The consensus among respondents seemed to be that the
manner. Sadness was not significantly correlated with the self- interviewer’s emotions should not be expressed during a suspect
reported use of any interview tactics. Fear was significantly interview. Some police interviewers described this as an easy task
negatively correlated with showing respect for the suspect (e.g., “It is a job I am used to and have learned to handle”),
and with claiming the suspect has more information. Interest whereas other practitioners described how they in different ways
was significantly positively correlated with expressing care for had to work actively on putting their emotions aside by using
the suspect’s feelings. Frustration was significantly positively different strategies. These strategies could be categorized into
correlated with raising one’s voice and with emphasizing the three different subthemes: (1) Focus on one’s professional role, (2)
seriousness of the offense. See the human behind the act, and (3) Take a break.
Qualitative Analysis: How Police Focus on One’s Professional Role

Interviewers Manage Emotions The most frequently mentioned way to manage emotions during
A total of 120 Swedish and 33 Norwegian police interviewers the interview was reminding oneself about one’s professional role.
provided a written response to the open-ended question “How The answers provided insight into different ways this could be

TABLE 7 | Bivariate correlations and 95% confidence intervals of interview tactics and emotion composites.
Anger Disgust Sadness Fear Interest Frustration
Treat friendly −0.094 [−0.237, 0.054] –0.176 [–0.315, –0.030] −0.015 [−0.161, 0.132] 0.017 [−0.130, 0.164] 0.122 [−0.025, 0.264] −0.014 [−0.160, 0.133]
Raise voice 0.280 [0.139, 0.410] 0.278 [0.137, 0.408] −0.025 [−0.171, 0.122] 0.115 [−0.032, 0.258] −0.097 [−0.240, 0.050] 0.229 [0.086, 0.364]
Care about –0.159 [–0.299, –0.013] −0.116 [−0.258, 0.031] 0.102 [−0.045, 0.245] 0.137 [−0.010, 0.278] 0.177 [0.031, 0.316] −0.080 [−0.252, 0.038]
feelings
Interrupt 0.090 [−0.057, 0.234] 0.095 [−0.053, 0.238] −0.031 [−0.177, 0.117] −0.042 [−0.187, 0.105] −0.109 [−0.252, 0.038] 0.050 [−0.097, 0.196]
Common −0.006 [−0.152, 0.141] −0.015 [−0.161, 0.132] 0.045 [−0.102, 0.190] 0.090 [−0.057, 0.234] 0.120 [−0.027, 0.262] −0.075 [−0.220, 0.072]
interests
Boss 0.097 [−0.051, 0.240] −0.012 [−0.158, 0.135] −0.020 [−0.166, 0.127] −0.100 [−0.243, 0.047] 0.061 [−0.086, 0.206] −0.035 [−0.181, 0.112]
Show −0.062 [−0.207, 0.085] −0.054 [−0.200, 0.093] −0.050 [−0.196, 0.097] –0.200 [–0.337, –0.055] 0.081 [−0.067, 0.225] −0.008 [−0.155, 0.139]
respect
Confront 0.047 [−0.101, 0.192] 0.045 [−0.103, 0.190] −0.100 [−0.243, 0.048] −0.067 [−0.211, 0.081] 0.096 [−0.051, 0.240] −0.022 [−0.168, 0.125]
with info
Emphasize 0.323 [0.185, 0.448] 0.280 [0.139, 0.410] 0.141 [−0.006, 0.281] 0.139 [−0.008, 0.280] −0.060 [−0.205, 0.088] 0.335 [0.198, 0.459]
seriousness
Let suspect 0.020 [−0.127, 0.166] −0.007 [−0.153, 0.140] 0.005 [−0.142, 0.152] −0.103 [−0.246, 0.044] 0.134 [−0.013, 0.275] −0.015 [−0.161, 0.132]
talk
Active −0.090 [−0.234, 0.057] −0.073 [−0.217, 0.074] −0.055 [−0.200, 0.093] −0.081 [−0.225, 0.066] 0.076 [−0.072, 0.220] −0.110 [−0.253, 0.037]
listening
Show what 0.068 [−0.080, 0.212] 0.066 [−0.082, 0.211] −0.029 [−0.175, 0.118] −0.020 [−0.166, 0.127] −0.058 [−0.203, 0.089] 0.092 [−0.056, 0.235]
you think
Unanticipated 0.160 [0.014, 0.300] 0.197 [0.051, 0.334] 0.005 [−0.142, 0.151] −0.057 [−0.202, 0.090] 0.079 [−0.068, 0.223] 0.115 [−0.032, 0.257]
questions
Claim 0.100 [−0.047, 0.244] 0.007 [−0.140, 0.154] −0.021 [−0.167, 0.126] –0.171 [–0.310, –0.025] −0.122 [−0.264, 0.025] −0.064 [−0.209, 0.083]
suspect has
more info
done. Oftentimes, remaining professional involved suppressing

TABLE 8 | Overview of the thematic structure and the proportion of responses
one’s personal feelings: categorized into each sub-theme. (n = 153 respondents).
Theme Sub-theme Responses

You adopt a professional role where your own emotions are
put aside. The goal is to get the suspect to talk, to put his/her
Putting emotions aside Focus on one’s professional role 85 (55.6%)
experiences and knowledge about what happened into words.
during the interview See the human behind the act 10 (6.5%)
Show respect—gather the information that is needed to clarify
Take a break 8 (5.2%)
guilt or innocence. My emotions are not expressed during
the interview situation (Norwegian police officer, 17 years of Coping with Share emotions with others 98 (64.1%)
experience interviewing CSA suspects). emotions afterwards Take care of oneself 25 (16.3%)
Reflect on the experience 9 (5.9%)
Other strategies were keeping their guard up and not letting The thematic analysis is based on open survey answers from 120 Swedish and 33
emotions arise in the first place, or not putting any personal Norwegian police interviewers.
judgement on the criminal act. Many interviewers found it

helpful to focus on the task in order to avoid being emotional, and Most officers attempted to suppress their feelings, but they also
some described how it could also be a way of motivating oneself acknowledged that this could be a difficult task. One police officer
when facing tasks that were more challenging than others: explained how interviewing suspects of CSA could be especially
stressful, particularly in cases where the same police officer had
first interviewed the child victim:
What can sometimes be difficult is to ask intimate questions
related to the sexual parts, or to ask about more details on a
specific topic (because you pressure the one being questioned). It is okay to have emotions, but they must be handled the
It is a kind of invasion of another person’s private sphere, even right way. If I experience too much emotion, I should leave the
though the situation has the framework that it has. I deal with interview to my colleague. This is especially important if I have
this by convincing myself that it is natural to talk about this, and conducted a child interview that evoked a lot of feelings (Swedish
that the situation requires it. I do what I can to both appear and police officer, 9 years of experience interviewing CSA suspects).
sound neutral. And sometimes I explain the reason for asking
these questions (Norwegian police officer, 5 years of experience According to some police interviewers, controlling one’s
interviewing CSA suspects). emotions was an ability developed by experience and thus got

easier over time. A few police interviewers revealed that they it pass’. Breathe. Take a break.” (Swedish police officer, 7 years of
occasionally had failed to do so, which could result in them experience interviewing CSA suspects).
raising their voice against the suspect. These situations were
described as a “loss of control” and thus not a planned strategy. Coping With Emotions Afterwards
Contrasting this common belief that emotions should not be Interviewing suspected CSA offenders could evoke a range
expressed, a few argued that showing emotions could be a way of emotions. However, as discussed in the previous main
of highlighting the seriousness of the situation. One police officer theme, most practitioners agreed that these emotions should
explained that it was not the initial strategy, but something one not influence their interview procedures. Furthermore, some
might have to resort to in exceptional cases: described that they rarely experienced any strong emotions
during the interview, but instead experienced that negative
I enter the room with an attitude of not knowing what has emotions could sometimes appear afterwards. The police
happened. I have received a story from a child and now I interviewers discussed different methods for coping with
want to know the suspect’s story. If the suspect does not emotions after leaving the interview room. These strategies
take the child’s story seriously, I might emphasize the gravity could be categorized into three different sub-themes: (1) Share
of the situation (Swedish police officer, 6 years of experience emotions with others, (2) Take care of oneself, and (3) Reflect on
interviewing CSA suspects). the experience.
See the Human Behind the Act Share Emotions With Others
Building a positive, open, and respectful environment was The most frequently reported strategy for managing emotions
described as a way of gaining as much information as possible was to share their feelings afterwards. Many police interviewers
during the interview. This could be accomplished by trying found support among their co-workers. Oftentimes, the police
to feel empathy for the suspect as a human being, without interviewers could vent to their closest colleagues who were
neglecting the seriousness of the crime that the suspect was knowledgeable about the situation. Since their co-workers shared
accused of committing. similar experiences, this form of collegial support was described
as very important for their emotion management.
Some emotions can be good, like feeling sympathy for the person
behind the crime. /. . . / The slightest sense on their part that you
I can talk to my co-workers after the interview about how I felt
feel anger, contempt etc., means that you will be met with the
and how I feel now. I experience that it helps to talk about the
same, and you will not get any answers. It is possible to get answers
emotions that arise. All my co-workers have a great understanding
to very intimate questions and to push on specific questions if
of the emotions that can be present. No one needs to feel stupid,
the person is treated with respect and without judgment (Swedish
or like it is wrong or something like that (Swedish civilian officer,
police officer, 10 years of experience interviewing CSA suspects).
less than 1 year of experience interviewing CSA suspects).
In some cases, it could be difficult to feel empathy for the suspect.

The Swedish police interviewers also talked about the value
The answers revealed how it could be useful to try focusing on the
of attending individual or group supervision sessions (typically
suspect as a multifaceted person, with both good and bad sides,
led by a psychologist). This form of regular supervision
and not being judgmental.
was described in positive terms. However, some practitioners
described that they did not have access to it on a regular
It helps me to think “he did not know better,” “his home life is bad,”
“lack of upbringing,” to have an explanation to myself (which I of
basis. In contrast, the Norwegian police interviewers did not
course do not say). It is difficult when you want them to feel guilt mention regular supervision as a strategy for managing emotions.
(if they do not express it) and at the same time not make them feel Lastly, a few police interviewers reported that they sometimes
bad (they are probably already feeling bad, plus it is not our job to shared their feelings with their partner or close friends. This
handle such a thing) (Swedish police officer, 1 year of experience could be challenging, however, as the police interviewers
interviewing CSA suspects). had to leave out case-specific information for sensitivity and
confidentiality reasons.
Take a Break
If negative emotions did arise during the interview, several police Take Care of Oneself
interviewers described how they tried to regain their focus by Another commonly reported strategy for managing emotions
pausing shortly. Some mentioned how a break could relieve was to take care of their well-being, both mentally and physically.
problematic emotions before continuing the interview: “If I Many police interviewers stated that they went for a run or did
feel frustrated I take a break and talk to my colleague in the other physical exercises after the interview to clear their mind and
back room, or a break with fresh air for us both” (Norwegian release any potential negative emotions. A few described other
police officer, 6 years of experience interviewing CSA suspects). forms of physical activity, including going on walks outside in
Others described how they could make sure to get a mental the fresh air. Some mentioned the importance of sleep for their
break in the middle of the interview by taking a sip of water, mental well-being. Having hobbies and a rich private life was also
taking a deep breath, flipping through some papers or asking the described as a protective factor for managing emotions that can
prosecutor if he/she has any questions: “Note the feeling and ‘let arise during their line of work.

Training and physical activity outside works if you feel it in your Swedish interviewers. This interpretation would be consistent
body. Being together with friends and family and keeping yourself with one of the systemic differences between the countries:
busy with other things (Norwegian police officer, 10 years of In Sweden, the same interviewer is often involved in both the
experience interviewing CSA suspects). child victim and suspect interview. In contrast, Norwegian
child interviewers often try to avoid interviewing the suspect
One police officer highlighted that although physical activity or in the same case (Detective superintendent Inger-Lise Brøste,
distractions was helpful, it did not replace the need for directly personal communication, 30 August 2019). This difference
addressing the emotions. became evident during data collection, where many Norwegian
child interviewers did not meet the inclusion criterion of having
“Exercising, doing completely other stuff, seeing the good world. experience conducting suspect interviews during the last 5 years.
This is, however, a kind of escape, and it is important to discuss It seems plausible that a more extensive personal involvement
the emotions in the group first.” (Swedish police officer, 5 years of in a case would generate more intense emotional reactions.
experience interviewing CSA suspects). This explanation is speculative, but it raises the interesting
possibility that reducing personal involvement may be an
Reflect on the Experience effective measure to minimize unwanted emotional interference
Lastly, some police interviewers discussed the importance of in criminal investigations.
reflecting upon the interview process and their performance. The observed correlations between emotions and interviewing
This could include self-reflection on how to improve for the goals and tactics were typically weak. This is perhaps to be
next interview. expected considering that respondents were asked to report
generalities across several cases rather than reflect on any
I always watch the video-recording of the interview and reflect on particular case. Nevertheless, we were able to detect a few patterns
how it felt, and about my performance. I think this creates some that deserve attention: First, the frequency of experiencing
distance to the case and my feelings, and I can learn from any anger was positively associated with the goal of obtaining
potential mistakes (Swedish police officer, 5 years of experience a confession and with aggressive tactics like raising one’s
interviewing CSA suspects).
voice, emphasizing the seriousness of the crime, and not
caring about the suspect’s feelings. From an emotion-theoretical
Others emphasized the importance of reflecting on their own perspective, these correlations make sense. The experience
reactions, in order to understand them and learn from their of anger is associated with action tendencies that facilitate
experiences until the next case: “I reflect on my own emotions. confrontational and confirmation-seeking behavior (Lerner and
Evaluate why they arose and what it triggered in me as a person.” Tiedens, 2006; Carver and Harmon-Jones, 2009). Second,
(Swedish civilian officer, 5 years of experience interviewing the reported frequency of experiencing frustration displayed
CSA suspects). a similar pattern as anger, which may indicate that some
interviewers resort to aggressive tactics when facing frustrating
DISCUSSION obstacles in the interview room. Third, disgust was similarly
associated with aggressive tactics and negatively, albeit weakly,
The present study sought to examine police interviewers’ associated with the goal of obtaining previously unknown
self-reported goals, tactics, and emotional experiences information. The fact that disgust is associated with avoidant
when questioning CSA suspects in a Scandinavian context. action tendencies (Frijda et al., 1989) might explain why the
The Swedish and Norwegian interviewers reported largely emotion correlates with a reduced interest in further information.
similar profiles of emotional experiences when conducting In addition, feelings of disgust are intimately associated with
these interviews. Reassuringly, interviewers appear to feel the act of moral condemnation (Chapman and Anderson, 2013),
predominantly interested. This was the only feeling that, on which may explain the link between the emotion and use of
average, occurs “often” according to the respondents’ self-ratings. aggressive interview tactics.
Negative emotions typically occur only “sometimes” or even It is important to note that our correlational data cannot
“rarely.” Of these, feelings of frustration seem to be the most distinguish between the different causal mechanisms that may
prevalent. Although the exact cause of frustration is not apparent produce the observed relationships between emotions and
from these ratings, they do imply that a failure to meet one’s interview goals and tactics. For instance, the possibility that
objective is not an uncommon experience when interviewing emotions precede the employment of goals and tactics (e.g.,
CSA suspects. Fear-related emotions, like feeling worried and anger increases the motivation to obtain a confession) is equally
scared, were the least frequently reported, suggesting that plausible as the opposite causal direction (e.g., confession-
interviewers rarely experience a complete lack of control or a seeking behavior produces frustrating suspect responses).
sense of threat in the interview room. Moreover, the correlations might be due to unspecified
Two notable differences were observed between Swedish individual differences among interviewers (e.g., personality
and Norwegian interviewers’ reported emotions: Swedish characteristics may influence both the likelihood of experiencing
respondents reported experiencing disgust and interest more anger and the preference for aggressive interview tactics).
frequently than their Norwegian counterparts. This might reflect This limitation notwithstanding, our correlational findings
a higher degree of personal involvement in CSA cases among highlight the relevance of emotional processes in CSA suspect

interrogations and raise the need for further exploration of the open-ended responses, information-gathering tactics were
the topic. much more common than accusatory techniques. Similarly, the
Despite the quantitative results indicating that emotions do quantitative results indicate that information-gathering goals
play a role in CSA suspect interviews, the participants asserted (e.g., collecting previously unknown information and getting
in their open-ended responses that emotions should not and explanations for investigative information) were prioritized to a
do not affect their work. The qualitative analysis showed strong greater extent than obtaining a confession. Moreover, interview
consensus regarding the importance of not expressing emotions tactics focused on collecting information and building rapport
during the suspect interview, or letting emotions influence (e.g., active listening, showing respect, and letting the suspect
their interview procedures. To achieve these aims, the police talk) were among those reported to be used most frequently,
interviewers described several strategies for managing their consistent with observational research examining the tactics used
emotional reactions and coping with emotions after the suspect in countries that have implemented investigative interviewing
interview. It is possible that a similar thematic structure would be approaches (see, e.g., Soukara et al., 2009).
applicable if asking police interviewers about managing emotions Although the police organizations in Sweden and Norway
in investigations of other violent crimes. However, cases share some similarities, there are significant systematic
involving children are often perceived as especially distressing differences that could help shed light on the variations in
(Brown et al., 1999; Oxburgh et al., 2015; Huey and Kalyal, 2017). tactics and goals across countries. With the implementation of
A few officers explicitly mentioned some unique aspects of CSA KREATIV, the shift toward a research-based interview model
investigations that could be particularly challenging, for instance, occurred nearly two decades ago in Norway. While Sweden is
having to ask intimate questions of a sexual nature. While this currently working toward implementing a national model, it will
could be perceived as invading the private sphere of the suspect, take time before the entire police force has received training in
it is a crucial task in a CSA investigation, illustrating the high the technique. The lack of a research-based interview model may
demand put on the professionals. help explain some of the differences observed between the two
In line with previous research emphasizing the importance of countries. For example, we observed that Swedish interviewers
emotional support from co-workers as a coping aid (Burns et al., reported more frequently using confrontational tactics, such
2008; Powell et al., 2014), both Swedish and Norwegian officers as raising one’s voice and emphasizing the seriousness of the
frequently expressed that sharing emotions with colleagues was crime, whereas Norwegian interviewers reported using these
helpful. Interestingly, Swedish police officers also commonly tactics less. Swedish interviewers also reported prioritizing
mentioned the benefits of attending regular psychology sessions obtaining confessions to a higher degree than their Norwegian
provided by their work, which was not discussed by the counterparts. These differences may be due, in part, to the
Norwegian officers. In previous studies, the attitudes toward KREATIV model emphasizing gathering information rather
receiving psychological support have varied, with the main than seeking a confession (Fahsing et al., 2016). As Sweden
reason for dissatisfaction being that psychologists lack specialized continues to work toward implementing its new interview
knowledge about investigating sensitive cases and thus came model, these differences may diminish over time.
across as uncomfortable talking about specific topics (Powell Furthermore, both countries are currently lacking specialized
et al., 2014). A fear of the possible repercussions of utilizing interview guidelines focused on investigations involving alleged
psychological support (e.g., that it is considered to be something CSA. Given the nature of these cases (e.g., the scarcity of
negative if applying for promotion), and a stigma associated corroborative evidence: Ernberg et al., 2018), the development
with officers seeking help has also been reported (Cullen et al., of research-based interview guidelines adapted for suspected
2020; Foley et al., 2021). Furthermore, having understanding CSA offenders may facilitate the police in their daily work.
and knowledgeable supervisors is also important (Powell et al., Indeed, research-based guidelines for CSA cases are especially
2014), as inadequate organizational support has been reported as important, given the quality of currently available practical
a factor underlying burnout (Fansher et al., 2020). recommendations. For example, the Reid Technique—
Another aspect highlighted in the open-ended responses, was an accusatory interrogation approach widely criticized by
that the interview with the child victim might evoke many researchers and legal scholars (see, e.g., Kassin et al., 2010)—
emotions, which was described as a complicating factor when offers a manual on interrogations in CSA cases (Buckley,
trying to remain neutral in the subsequent interview with 2015), closely based on its more general system of interrogating
the suspect. While this contrasts previous findings that police suspects (Inbau et al., 2013), which advocates the use of
interviewers used fewer negative emotional utterances during problematic tactics (e.g., minimization; Luke and Alceste, 2020).
the interview of a CSA suspect if they had first interviewed the
child (Oxburgh et al., 2006), it is possible that police interviewers LIMITATIONS
manage to remain neutral, but that it is highly emotionally taxing
for them. Some methodological limitations of the current research need
The qualitative analysis also provided some insight into some to be addressed. First, the findings are correlational, and we
police interviewers’ goals of the suspect interview. Several officers can therefore not draw any conclusions regarding causality
described how their main task was to get the suspect to talk or directionality. Second, the data are based on retrospective
and provide his or her view on the allegations, while a few self-reports from practitioners. Survey data can be affected
described how they emphasized the seriousness of the crime in by a range of response biases, including social desirability
order to evoke feelings of guilt in the suspect. Thus, based on effects (Tourangeau et al., 2000). Furthermore, we asked the

participants about their general perceptions and experiences of access to corroborative evidence in CSA investigations (Ernberg
interviewing CSA suspects. Their goals, tactics, and emotional et al., 2018), more research is needed to better understand how
experiences might differ between interviews depending on the to effectively interview CSA suspects. The potential link between
situation at hand (e.g., the amount of corroborative evidence interviewers’ goals, tactics, and emotional experiences should be
available, whether the suspect is cooperative or uncooperative). kept in mind during the development of such guidelines.
Due to limitations with self-reports, systematic field research
is needed to study the techniques used by Swedish and DATA AVAILABILITY STATEMENT
Norwegian practitioners inside the interview rooms. Finally,
our sample was relatively small (total N = 176), and as a The quantitative data supporting the conclusions of this article
consequence, our ability to estimate small and subtle differences will be made available in aggregated form by the authors on
is limited. The small sample is especially concerning given request, without undue reservation.
that the correlations we observed between goal, tactics, and
emotions were small in size. That being said, given that ETHICS STATEMENT
there are relatively few police interviewers in Sweden and
Norway who conduct interviews with suspects of CSA, our Ethical review and approval was not required for the study on
sample nevertheless represents a substantial portion of the human participants in accordance with the local legislation and
total population. institutional requirements. The patients/participants provided
their written informed consent to participate in this study.
CONCLUSIONS
AUTHOR CONTRIBUTIONS
This study set out to explore Swedish and Norwegian police
interviewers’ goals, tactics, and emotional experiences when MM, MJ, and TL designed the study. MM, MJ, and MLS collected
questioning suspected CSA offenders. The quantitative data the data. MM and MJ carried out the qualitative analyses. TL and
showed correlations between officers’ self-reported emotional KA carried out the quantitative analyses. All authors discussed
experience and the tactics they use and the goals they prioritize, the results and worked on the manuscript.
suggesting that emotions play a role in the interview of suspected
CSA offenders. In the qualitative analysis, however, participants SUPPLEMENTARY MATERIAL
asserted that emotions should not and do not affect their work.
They described a range of different strategies that could be used The Supplementary Material for this article can be found
to put emotions aside during the interview and instead manage online at: https://www.frontiersin.org/articles/10.3389/fpsyg.
them afterwards. Given the low prosecution rates and limited 2021.606774/full#supplementary-material
REFERENCES Carver, C. S., and Harmon-Jones, E. (2009). Anger is an approach-related affect:

evidence and implications. Psychol. Bull. 135, 183–204. doi: 10.1037/a0013965
Ask, K., and Granhag, P. A. (2007). Hot cognition in investigative judgments: Chapman, H. A., and Anderson, A. K. (2013). Things rank and gross in
the differential influence of anger and sadness. Law Hum. Behav. 31, 537–551. nature: a review and synthesis of moral disgust. Psychol. Bull. 139, 300–327.
doi: 10.1007/s10979-006-9075-3 doi: 10.1037/a0030964
Ask, K., and Pina, A. (2011). On being angry and punitive: how anger Creswell, J. W., and Clark, V. L. P. (2017). Designing and Conducting Mixed
alters perceptions of criminal intent. Soc. Psychol. Personal. Sci. 2, 494–499. Methods Research. Thousand Oaks, CA: Sage Publications.
doi: 10.1177/1948550611398415 Cullen, O., Zug Ernst, K., Dawes, N., Binford, W., and Dimitropoulos, G. (2020).
Bourke, M. L., and Craun, S. W. (2014). Coping with secondary traumatic stress: “Our laws have not caught up with the technology”: understanding challenges
differences between U.K. and U.S. child exploitation personnel. Traumatology and facilitators in investigating and prosecuting child sexual abuse materials in
20, 57–64. doi: 10.1037/h0099381 the United States. Laws 9:28. doi: 10.3390/laws9040028
Braun, V., and Clarke, V. (2006). Using thematic analysis in psychology. Qual. Res. Ernberg, E., Magnusson, M., and Landström, S. (2018). Court evaluations of young
Psychol. 3, 77–101. doi: 10.1191/1478088706qp063oa children’s testimony in child sexual abuse cases. Legal Criminol. Psychol. 23,
Brown, J., Fielding, J., and Grover, J. (1999). Distinguishing traumatic, 176–191. doi: 10.1111/lcrp.12124
vicarious and routine operational stressor exposure and attendant adverse Ernberg, E., Magnusson, M., and Landström, S. (2020). Prosecutors’
consequences in a sample of police officers. Work. Stress. 13, 312–325. experiences investigating alleged sexual abuse of preschoolers.
doi: 10.1080/02678379950019770 Psychol. Crime Law 26, 687–709. doi: 10.1080/1068316X.2019.
Brubacher, S. P., Peterson, C., La Rooy, D., Dickinson, J. J., and Poole, D. A. 1709639
(2019). How children talk about events: implications for eliciting and analyzing Fahsing, I., Jakobsen, K., and Öhrn, H. (2016). “Investigative interviewing of
eyewitness reports. Dev. Rev. 51, 70–89. doi: 10.1016/j.dr.2018.12.003 suspects in Scandinavia,” in International Developments and Practices in
Buckley, D. M. (2015). How to Identify, Interview and Motivate Child Abuse Investigative Interviewing and Interrogation, Vol. 2, Suspects, eds D. W.
Offenders to Tell the Truth, 2nd Edn. Eagle River, WI: Hanh Printing, Inc. Walsh, G. E. Oxburgh, A. D. Redlich, and T. Myklebust (London, UK:
Burns, C. M., Morley, J., Bradshaw, R., and Domene, J. (2008). The Routledge), 180–192.
emotional impact on and coping strategies employed by police teams Fahsing, I. A., and Rachlew, A. (2009). “Investigative interviewing in the
investigating internet child exploitation. Traumatology 14, 20–31. Nordic region,” in International Developments in Investigative Practices, eds T.
doi: 10.1177/1534765608319082 Williamson, B. Milne and S. P. Savage (Devon: Willian Publishing), 66–91.

Fansher, A. K., Zedaker, S. B., and Brady, P. Q. (2020). Burnout among forensic homicide, filicide and child sexual abuse. Psychiatry Psychol. Law 21, 903–917.
interviewers, how they cope, and what agencies can do to help. Child Maltreat. doi: 10.1080/13218719.2014.918078
25, 117–128. doi: 10.1177/1077559519843596 Oxburgh, G., Ost, J., Morris, P., and Cherryman, J. (2015). Police officers’
Foley, J., Hassett, A., and Williams, E. (2021). “Getting on with the job”: a perceptions of interviews in cases of sexual offences and murder
systematized literature review of secondary trauma and post-traumatic stress involving children and adult victims. Police Pract. Res. 16, 36–50.
disorder (PTSD) in policing within the United Kingdom (UK). Police J. doi: 10.1080/15614263.2013.849595
doi: 10.1177/0032258X21990412. [Epub ahead of print]. Oxburgh, G., Williamson, T., and Ost, J. (2006). Police officers’ use of emotional
Forgas, J. P. (2002). Feeling and doing: affective influences on interpersonal language during child sexual abuse investigations. J. Investig. Psychol. Offen.
behavior. Psychol. Inq. 13, 1–28. doi: 10.1207/S15327965PLI1301_01 Profil. 3, 35–45. doi: 10.1002/jip.41
Frijda, N. H., Kuipers, P., and Ter Schure, E. (1989). Relations among emotion, Perez, L. M., Jones, J., Englert, D. R., and Sachau, D. (2010). Secondary
appraisal, and emotional action readiness. J. Pers. Soc. Psychol. 57, 212–228. traumatic stress and burnout among law enforcement investigators
doi: 10.1037/0022-3514.57.2.212 exposed to disturbing media images. J. Police Crim. Psychol. 25, 113–124.
Granhag, P. A., and Magnusson, M. (2017). “Förhör med misstänkt [Suspect doi: 10.1007/s11896-010-9066-7
interrogation],” in Klinisk Forensisk Psykologi [Clinical Forensic Psychology], eds Powell, M., Cassematis, P., Benson, M., Smallbone, S., and Wortley, R. (2014).
A. Eriksson and K. Sturidsson (Lund: Studentlitteratur), 75–104. Police officers’ strategies for coping with the stress of investigating internet child
Granhag, P. A., Strömwall, L. A., and Cancino Montecinos, S. (2013). Polisens exploitation. Traumatology 20, 32–42. doi: 10.1037/h0099378
Förhör med Misstänkta. Svensk Utbildning i Internationell Belysning. [Police Prosecution Development Centre Gothenburg (2016). Kvalitet i barnförhör
Interrogations With Suspects. Swedish Education From an International [Quality of forensic interviews with children]. Utvecklingscentrum Göteborg.
Perspective]. Stockholm: Rikspolisstyrelsen. Read, J. M., Powell, M. B., Kebbell, M. R., and Milne, R. (2009). Investigative
Gudjonsson, G. H., and Sigurdsson, J. F. (2000). Differences and similarities interviewing of suspected sex offenders: a review of what constitutes best
between violent offenders and sex offenders. Child Abuse Neglect 24, 363–372. practice. Int. J. Police Sci. Manage. 11, 442–459. doi: 10.1350/ijps.2009.11.4.143
doi: 10.1016/s0145-2134(99)00150-7 Sambrano, D., Masip, J., and Blandón-Gitlin, I. (2020). How emotions affect
Holmberg, U., and Christianson, S. Å. (2002). Murderers’ and sexual offenders’ judgement and decision making in an interrogation scenario. Legal Criminol.
experiences of police interview and their inclination to admit or deny crimes. Psychol. 26, 62–82. doi: 10.1111/lcrp.12181
Behav. Sci. Law 20, 31–45. doi: 10.1002/bsl.470 Scherer, K. R., and Moors, A. (2019). The emotion process: event appraisal
Huey, L., and Kalyal, H. (2017). ‘We deal with human beings’ The emotional and component differentiation. Annu. Rev. Psychol. 70, 719–745.
labor aspects of criminal investigation. Int. J. Police Sci. Manag. 19, 140–147. doi: 10.1146/annurev-psych-122216-011854
doi: 10.1177/1461355717717996 Seigfried-Spellar, K. C. (2018). Assessing the psychological well-being and coping
Inbau, F. E., Reid, J. E., Buckley, J. P., and Jayne, B. C. (2013). Criminal mechanisms of law enforcement investigators vs. digital forensic examiners
Interrogation and Confessions, 5th Edn. Burlington, MA: Jones and Bartlett. of child pornography investigations. J. Police Crim. Psychol. 33, 215–226.
Kassin, S. M., Drizin, S. A., Grisso, T., Gudjonsson, G. H., Leo, R. A., and Redlich, doi: 10.1007/s11896-017-9248-7
A. D. (2010). Police-induced confessions: risk factors and recommendations. Soukara, S., Bull, R., Vrij, A., Turner, M., and Cherryman, J. (2009). What really
Law Hum. Behav. 34, 3–38. doi: 10.1007/s10979-009-9188-6 happens in police interviews of suspects? Tactics and confessions. Psychol.
Kebbell, M., Alison, L., Hurren, E., and Mazerolle, P. (2010). How do sex offenders Crime Law 15, 493–506. doi: 10.1080/10683160802201827
think the police should interview to elicit confessions from sex offenders? Tiedens, L. Z., and Linton, S. (2001). Judgment under emotional certainty
Psychol. Crime Law 16, 567–584. doi: 10.1080/10683160902971055 and uncertainty: the effects of specific emotions on information
Lerner, J. S., and Keltner, D. (2001). Fear, anger, and risk. J. Pers. Soc. Psychol. 81, processing. J. Pers. Soc. Psychol. 81, 973–988. doi: 10.1037/0022-3514.81.
146–159. doi: 10.1037/0022-3514.81.1.146 6.973
Lerner, J. S., and Tiedens, L. Z. (2006). Portrait of the angry decision maker: how Toch, H. (2002). Stress in Policing. Washington, DC: American
appraisal tendencies shape anger’s influence on cognition. J. Behav. Decis. Mak. Psychological Association.
19, 115–137. doi: 10.1002/bdm.515 Tourangeau, R., Rips, L. J., and Rasinski, K. (2000). The Psychology of Survey
Luke, T. J., and Alceste, F. (2020). The mechanisms of minimization: Response. Cambridge: Cambridge University Press.
how interrogation tactics suggest lenient sentencing through pragmatic Turgoose, D., Glover, N., Barker, C., and Maddox, L. (2017). Empathy, compassion
implication. Law Hum. Behav. 44, 266–285. doi: 10.1037/lhb00 fatigue, and burnout in police officers working with rape victims. Traumatology
00410 23, 205–213. doi: 10.1037/trm0000118
Mac Giolla, E., and Granhag, P. A. (2017). “Interviewing suspects,” in Forensic Ward, T., and Siegert, R. J. (2002). Toward a comprehensive theory of child
Psychology: Crime, Justice, Law and Interventions, eds G. M. Davies and A. R. sexual abuse: a theory knitting perspective. Psychol. Crime. Law. 8, 319–351.
Beech (Chichester: Wiley-Blackwell), 412–448. doi: 10.1080/10683160208401823
Magnusson, M., Ernberg, E., Landström, S., and Granhag, P. A. (2018). Taking Whitaker, D. J., Le, B., Hanson, R. K., Baker, C. K., McMahon, P. M.,
the stand: defendant statements in court cases of alleged sexual abuse Ryan, G., et al. (2008). Risk factors for the perpetration of child sexual
against infants, toddlers and preschoolers. Psychol. Crime Law 24, 744–759. abuse: a review and meta-analysis. Child Abuse Neglect 32, 529–548.
doi: 10.1080/1068316X.2018.1424845 doi: 10.1016/j.chiabu.2007.08.005
Meissner, C. A., Redlich, A. D., Michael, S. W., Evans, J. R., Camilletti,
C. R., Bhatt, S., et al. (2014). Accusatorial and information-gathering Conflict of Interest: The authors declare that the research was conducted in the
interrogation methods and their effects on true and false confessions: a absence of any commercial or financial relationships that could be construed as a
meta-analytic review. J. Exp. Criminol. 10, 459–486. doi: 10.1007/s11292-014- potential conflict of interest.
9207-6
Milne, B., and Bull, R. (1999). Investigative Interviewing: Psychology and Practice. Copyright © 2021 Magnusson, Joleby, Luke, Ask and Lefsaker Sakrisvold. This is an
Chichester: Wiley. open-access article distributed under the terms of the Creative Commons Attribution
Oxburgh, G., and Ost, J. (2011). The use and efficacy of empathy in police License (CC BY). The use, distribution or reproduction in other forums is permitted,
interviews with suspects of sexual offences. J. Investig. Psychol. Offen. Profil. 8, provided the original author(s) and the copyright owner(s) are credited and that the
178–188. doi: 10.1002/jip.143 original publication in this journal is cited, in accordance with accepted academic
Oxburgh, G., Ost, J., Morris, P., and Cherryman, J. (2014). The impact practice. No use, distribution or reproduction is permitted which does not comply
of question type and empathy on police interviews with suspects of with these terms.

ORIGINAL RESEARCH
published: 19 August 2021
doi: 10.3389/fpsyg.2021.624331
Greater Knowledge Enhances

Complainant Credibility and
Increases Jury Convictions for Child
Sexual Assault
Jane Goodman-Delahunty 1*, Natalie Martschuk 2 , Eunro Lee 3 and Annie Cossins 4
1
Newcastle Law School, University of Newcastle, Newcastle, NSW, Australia, 2 Griffith Criminology Institute, Griffith
University, Brisbane, QLD, Australia, 3 School of Health and Biomedical Sciences, The Royal Melbourne Institute of
Technology, Melbourne, VIC, Australia, 4 School of Law, Society & Criminology, Faculty of Law, University of New South
Wales, Kensington, NSW, Australia
Child sexual assault (CSA) cases reliant on uncorroborated testimony yield low conviction
rates. Past research demonstrated a strong relationship between verdict and juror CSA
knowledge such as typical delays in reporting by victims, and perceived victim credibility.
This trial simulation experiment examined the effectiveness of interventions by an expert
witness or an educative judicial direction in reducing jurors’ CSA misconceptions.
Participants were 885 jurors in New South Wales, Australia. After viewing a professionally
acted video trial, half the jurors rendered individual verdicts and half deliberated in
groups of 8–12 before completing a post-trial questionnaire. Multilevel structural equation
Edited by: modeling exploring the relationship between CSA knowledge and verdict demonstrated
Noora Ellonen, that greater CSA knowledge after the interventions increased the odds ratio to convict by
University of Tampere, Finland
itself, and that the judicial direction predicted a higher level of post-trial CSA knowledge in
Reviewed by:
Ana M. Martín,
jurors than other expert interventions. Moreover, greater CSA knowledge was associated
University of La Laguna, Spain with heightened credibility perceptions of the complainant and a corroborating witness.
Mikaela Magnusson,
At the conclusion of the trial, the more jurors knew about CSA, the higher the perceived
University of Gothenburg, Sweden
credibility of both the complainant and her grandmother, and the more likely jurors were
*Correspondence:
Jane Goodman-Delahunty to convict the accused.
Jane.Delahunty@newcastle.edu.au
Keywords: jury decision making, child sexual abuse, educative information, expert evidence, judicial directions,
witness credibility, deliberation, multilevel structural equation modeling
Specialty section:
INTRODUCTION
In Australia, child sexual assault (CSA) cases typically result in low conviction rates, possibly
Received: 31 October 2020 because of a lack of corroborative evidence to prove the alleged sexual abuse (Cossins, 2020) but
Accepted: 16 July 2021 also because of research findings suggesting a strong relationship between juror misconceptions
Published: 19 August 2021
about CSA, such as expectations that the victim will resist and immediately report the abuse (Quas
Citation: et al., 2005; Cossins, 2008; Cossins et al., 2009), low assessments of complainant credibility (Gabora
Goodman-Delahunty J, Martschuk N,
et al., 1993), and a high acquittal rate (Wundersitz, 2003; Fitzgerald, 2006; Goodman-Delahunty
Lee E and Cossins A (2021) Greater
Knowledge Enhances Complainant
et al., 2017a,b). Several studies have documented juror uncertainty and/or lack of knowledge
Credibility and Increases Jury about children’s reactions to sexual assault which is incongruent with responses of sexually abused
Convictions for Child Sexual Assault. children, especially when the abuser is known to the complainant (Cossins et al., 2009; Seymour
Front. Psychol. 12:624331. et al., 2013). Jurors may disregard counter-intuitive evidence which contradicts common CSA
doi: 10.3389/fpsyg.2021.624331 beliefs and stereotypes, and may rely on misperceptions and erroneous stereotypes in the absence

1 August 2021 | Volume 12 | Article 624331
Goodman-Delahunty et al. Jury Deliberation in a CSA Case
of forensic evidence (Cossins and Goodman-Delahunty, 2013). are prosecuted (NSW Bureau of Crime Statistics Research, 2013).
This view is consistent with dual processing social persuasion A recent crime report showed an increase of 5.3% in sexual
theories of jury decision making showing that jurors may resort assault and indecent assault incidents in the past 60 months
to quick, heuristic peripheral information processing in the (and an increase of 9.4% in 24 months), while rates of other
absence of motivation or time to engage in more effortful central major offenses remained stable or decreased, with the exception
processing of substantive, scientific information (Salerno et al., of domestic violence (NSW Bureau of Crime Statistics Research,
2017). 2020). Despite their prevalence, CSA cases and adult sexual
Prior studies have examined the effectiveness of specialized assault cases produce the lowest conviction rates at trial (61%
knowledge to counteract individual jurors’ CSA misconceptions and 66% respectively), compared to 89% for all other offenses
(Cossins et al., 2009; Goodman-Delahunty et al., 2010, 2011a). (e.g., 70% for assault, 73% for robbery, 94% for illicit drugs)
Specialized educative summaries were derived from results of (Fitzgerald, 2006; Browne, 2017; Cashmore et al., 2019). The
empirical studies of the counter-intuitive behaviors of sexually number of CSA convictions decreased by 8% between 2016 and
abused children as well as information about children’s memory, 2020 (NSW Bureau of Crime Statistics and Research, 2021).
reliability and suggestibility (Goodman-Delahunty et al., 2017a).
Two intervention sources presented the educative information
to mock jurors: (a) expert witness evidence; or (b) the presiding
The Influence of Individual Juror
judge, in the form of a specially drafted educative judicial Characteristics in Cases of Child Sexual
direction to jurors. Expert evidence in CSA trials is permitted Assault
in five out of six Australian jurisdictions1 , but remains under- Meta-analyses of the influence of individual juror characteristics
utilized in practice (Shead, 2014). The second source is not in specific cases have generally yielded results showing little
permissible under Australian law, but is allowed in New Zealand difference between students and non-students (Bornstein and
when a complainant is <6 years of age [Section 21(3), Evidence Greene, 2017). However, in certain types of cases, jurors with
Act 1908 (NZ)]. A comparison of the effectiveness of these two different pretrial attitudes and beliefs produce different verdicts.
interventions is important to inform legal practitioners about the For example, individual jurors who were more authoritarian and
efficacy of expert evidence and to consider law reform proposals who supported capital punishment were more prone than other
regarding judicial directions. Other legal mechanisms to reduce jurors to convict; and in sentencing decisions, students were
jury bias, such as jury selection, were not tested since prospective more lenient than non-student mock jurors (Field and Barnett,
jurors cannot be questioned before empanelment in Australian 1978). A meta-analysis that tested the influence of individual
courts, where juror selection is limited to a few peremptory juror characteristics on verdicts showed a weak but statistically
challenges based on the appearance of the juror, or challenges for significant gender effect regardless of crime type, such that
cause (Horan and Goodman-Delahunty, 2010). women were more likely to convict than men; in particular,
Our prior research using written simulated CSA trial among student mock-jurors (Devine and Caughlin, 2014).
materials demonstrated that specialized knowledge reduced CSA Gender effects in sex offense cases are commonplace (Krauss
misconceptions, enhanced credibility ratings of the complainant et al., 2012). One meta-analysis of sexual assault cases showed
and increased the conviction rate (Cossins and Goodman- that women were more prone than men to convict (Schutte and
Delahunty, 2013). However, these findings were based on Hosch, 1997). Women in the role of mock-jurors rated child
individual mock-juror decisions without jury group deliberation. complainants more credible than did their male counterparts
Since deliberation is a critical legal procedure expected to reduce (Tabak and Klettke, 2014). Although some past studies have
jury errors and individual biases (Levett and Devine, 2017), yielded mixed results based on juror demographics such as age
adding deliberation to a jury research program is vital to assess its and gender (possibly attributable to the study sample, Devine
impact on misconceptions that may influence case outcomes in and Caughlin, 2014), a robust relationship between susceptibility
CSA trials. An important question is whether changes in research to CSA misconceptions and verdict emerged among jury-
procedures, such as adding group deliberation, interact with eligible undergraduate students and community volunteers who
substantive variables to influence the outcomes of simulated jury served as mock-jurors (Goodman-Delahunty et al., 2011a).
studies because “the presence of interaction effects may indicate Their demographic profiles and beliefs about CSA differed from
that aspects of the research method limit the external validity or those of citizens who respond to a summons for jury duty,
generalizability of the research conclusions” (Penrod et al., 2011, in terms of age, educational levels, parenting experience and
p. 197). Accordingly, research testing the relationship between other attitudinal measures (Goodman-Delahunty et al., 2017b).
construct, external, and ecological validity is needed (Wiener Accordingly, further research into the impact of juror attitudes
et al., 2011). In this study, we consider the example of CSA, and demographic features on case outcomes in CSA cases is
since these cases constitute the highest proportion of all criminal required to determine whether CSA cases comprise a specific type
offenses in Australian courts where the most indictable offenses of legal case where variations in jury beliefs and demographic
composition are associated with trial outcomes, and to assess
1 The Evidence Act, 1995, section 79(2) admits expert evidence as an exception
the generalizability of results obtained with jury eligible students
to the opinion rule which generally excludes opinion evidence. These provisions
(Henrich et al., 2010) and other non-jury community samples
apply in the ACT, NSW, the Northern Territory, Tasmania, the Commonwealth to actual jury samples (MacCoun, 2005; McCabe et al., 2010;
and Victoria. Lieberman et al., 2011).

The Form of Expert Evidence Proffered in s57; Jury Act, 1899 (Tas), s48(2); Juries Act, 2000 (Vic), s46; Juries
CSA Cases Act, 1957 (WA), s41).
To develop policies to guide courts in appointing expert witnesses Overall, deliberation studies have yielded diverse outcomes.
in CSA cases, research on the most effective expert witnesses Although one meta-analysis reported that the impact of expert
is helpful (Cossins and Goodman-Delahunty, 2013). Several evidence did not differ as a function of deliberation (Nietzel et al.,
studies have examined the impact of the type of expertise 1999), none of the studies examined jury deliberation in a CSA
and credentials of expert witnesses who testify in CSA cases case following the presentation of expert witness testimony. Some
(Kovera et al., 1994, 1997; Klettke et al., 2009). An expert indication of the effectiveness of expert evidence emerged in a
who is an experimental psychologist will provide a summary of study of an adult sexual assault: student mock-jurors rated the
relevant research pertinent to the case, often described as “social complainant more credible when exposed to expert evidence,
framework” testimony (Monahan et al., 2009), but typically and preferred an expert who linked evidence to case facts over
does not offer an explicit ultimate opinion as to whether the one who did not (Brekke and Borgida, 1998). To date, few
complainant was or was not sexually assaulted (Faigman et al., studies have examined deliberation about CSA. One exception
2014). By comparison, a clinical psychologist who interviews the is an Australian study that incorporated online rather than in
complainant may offer an ultimate opinion about whether the person deliberations (Tabak et al., 2013; Tabak and Klettke, 2014).
child has been sexually abused. Past studies which varied the Deliberation content analyses revealed a focus on the perceived
credentials of social framework vs. diagnostic experts revealed truthfulness of the victim, the context of the allegation, the
that jurors were more readily persuaded by a clinician than an behavior of the victim and defendant, and the inconclusive nature
experimental researcher in death penalty and in sex offender of word-against-word evidence.
cases (Krauss and Sales, 2001; Krauss et al., 2012). Prior research reviews indicated that the strength of pre-
Results of our prior CSA jury simulation study using written deliberation attitudes may be reduced by deliberation (Penrod
trial materials yielded no statistically significant difference et al., 2011; Levett and Devine, 2017). For example, the magnitude
between the perceived credibility of a social framework vs. a of correlations of predeliberation attitudes and verdict among
diagnostic expert, but the conviction rate following evidence empaneled jurors serving on a criminal case was larger than that
from the diagnostic expert (71%) exceeded that in response following deliberation (Moran and Comfort, 1982). However,
to the social framework expert (65%; Odds Ratio [OR] = mock-jury research on other topics conducted in North America
1.32) and was significantly higher than that in the jury group (Devine, 2012), South Korea (Park et al., 2005) and Taiwan
exposed to no expert evidence (37.5%; OR = 4.08) (Goodman- (Huang and Lin, 2014) demonstrated a “leniency effect” following
Delahunty et al., 2011a). A case study examining juror responses jury deliberation, with more decisions to acquit when verdicts
to expert evidence in a real CSA trial that ended in acquittal before and after group deliberation were compared (Peter-
revealed that jurors complained that the social framework Hagene et al., 2019). Similarly, a recent study of a case of historical
expert had never seen the child, and was unhelpful because child sexual assault showed that individual pretrial attitudinal
his evidence was non-specific (Horan and Goodman-Delahunty, biases were associated with mock-juror decisions (culpability and
2020). Moreover, where the social framework evidence was a poor verdict) at the individual juror level, while this effect disappeared
fit for specific case facts, some jurors cited those gaps as reasons at the jury group level (Goodman-Delahunty and Martschuk,
to discount the evidence and credibility of the complainant 2020). One explanation posited for the lower postdeliberation
(Horan and Goodman-Delahunty, 2020). Further research is conviction rates is that informational biases held by individual
needed to examine whether observed differences in conviction jurors exert a stronger influence than normative biases, or
rates are attributable to unmeasured differences in the perceived conformity effects in group decision making (Peter-Hagene
credibility of the two types of experts, or the fact that the et al., 2019). These researchers called for studies exploring
diagnostic expert interviewed the child, and whether this finding the qualitative differences observed between individual juror
is replicated in a sample of actual jurors. Use of videotaped and jury groups verdicts, using techniques such as multi-level
trial materials in which the same actor portrays either a social modeling (Lovis-McMahon, 2015). Whether deliberation after
framework or a diagnostic expert can tease apart these factors. juries are exposed to specialized CSA knowledge will decrease the
conviction rate is the topic of the present study.
A common feature of mock jury research is the use of
written trial summaries or case transcripts instead of a live trial
The Impact of Deliberation in Child Sexual enactment. Differences between types of trial simulations may
Assault Cases affect mock-jurors’ engagement with the case and their capacity
As was noted above, a key component of jury decision-making to assess witness credibility, a key component of the decision-
is deliberation with fellow jurors. Interestingly, in six Australian making process to reach a verdict, particularly in word-against-
jurisdictions, majority verdicts are permitted in relation to some word cases. In CSA cases, the method of evidence presentation
offenses. Jurors are first instructed to deliberate to a unanimous can impact jury responses in assessing the credibility of the
verdict, and if unable to do so within a specified time, they complainant and accused (Eaton et al., 2001; Landström et al.,
are then instructed that a majority verdict of 11-1 or 10-2 will 2010). Comparisons of written and videotaped methods of
suffice, depending on the jurisdiction (Jury Act, 1977 (NSW), presentation in simulated trials showed less impact of juror
s55F; Criminal Code Act, 1983 (NT), s368; Juries Act, 1927 (SA), attitudes when more realistic videotaped trial materials were

used (Nietzel et al., 1999; Penrod et al., 2011). Reviews of the male (hypothesis 1a), younger (hypothesis 1b) and less well-
method of trial presentation yielded mixed outcomes (Penrod educated counterparts (hypothesis 1c);
et al., 2011), including results of a meta-analysis of cases that 2. Exposure to educative interventions will increase juror post-
included presentations of expert witness testimony (Nietzel et al., trial CSA knowledge compared to that of jurors who are not
1999). Accordingly, a videotaped simulated trial may diminish exposed to any intervention (hypothesis 2);
the influence of jury beliefs about CSA on credibility assessments, 3. Jurors with less CSA knowledge will make unfavorable
as was observed in response to written trial simulation materials. assessments of the complainant’s credibility and acquit the
The trial simulation experience is a further issue to consider accused while jurors with more CSA knowledge will assess
when conducting mock-jury research. For example, if materials the complainant’s credibility favorably and convict the accused
are administered online or in a laboratory setting, participating (hypothesis 3);
mock-jurors miss the experience of coming to court, engaging 4. The credibility of the diagnostic psychologist will be rated
with court personnel, attending the court’s jury orientation more favorably than that of the social framework expert
and induction training, all of which emphasize the solemnity and will produce statistically significantly more convictions
and gravity of jurors’ responsibilities, and have been shown to compared to jury groups exposed to other educative
influence jury motivations (Bornstein and McCabe, 2005) and interventions (hypothesis 4);
attitudes (O’Brien et al., 2008). Some researchers contend that 5. The acquittal rate among deliberating jurors will exceed that
these additional contextual features of ecological validity should of non-deliberating jurors (hypothesis 5);
not be overlooked (Vidmar, 2008; Ceci et al., 2010) as they may 6. Credibility perceptions of the complainant and a
impact mock-jurors’ motivation and the decision criteria applied corroborative witness will mediate the effect of post-trial
in assessing the consequences of the verdict. The process of CSA knowledge on the likelihood to convict (hypothesis 6).
group deliberation has yielded conflicting outcomes about jurors’
cognitive performance (Peter-Hagene et al., 2019). METHOD
Aims of the Present Study Participants
The present study had four aims: First it explored the Participants were 885 jurors (58% men, 42% women) who
demographics of a sample of non-empaneled jurors and reported for jury duty in the District and Supreme Courts of
measured their attitudes to CSA to discern whether individual New South Wales (NSW), Australia, but were later excused. Aged
juror characteristics were systematically related to jury verdicts. between 18 and 74 years (M = 43.4, SD = 13.33), more than half
Second, it compared the effectiveness of educative interventions the participants held a university degree (61.7%), 17.1% had a
in increasing juror CSA knowledge, namely (a) a specially tertiary level diploma, 7.2% had a trade certificate, 10.5% finished
drafted judicial direction, and (b) social framework vs. diagnostic high school, and 3.6% reported fewer than 12 years of formal
expert testimony. Third, the study compared individual juror education. English was the first language for 84% of participants.
verdicts with those of deliberating jurors to assess the impact More than half the participants reported that they were a parent
of decision type. A fourth exploratory aim was to test or guardian of a child (55%).
whether the perceived credibility of the child complainant and
corroborative witness, separate and apart from CSA knowledge, Research Design and Procedure
contributed to decisions to convict as psychological mediators. A mixed study design was applied in which the first variable, CSA
The dependent measures were CSA knowledge, assessments of knowledge, was a within-subjects factor, and two variables were
witness credibility, and verdicts. between-subjects factors, namely Decision Type (Deliberation in
The study incorporated ecologically realistic components of jury groups vs. Non-deliberating individual juror decision) and
the jury task into the research method by (a) inviting jurors who Source of Specialized CSA Knowledge (None, Judicial directions,
reported for jury duty to participate in a simulated trial (in lieu of Social framework expert, Diagnostic expert). A total of 443
jury eligible students and community volunteers); (b) conducting deliberating jurors and 442 non-deliberating jurors were assigned
the study in District and Supreme Courts of NSW (in lieu of to one of eight experimental groups. See Table 1.
online or laboratory settings); (c) using a professionally-acted All jurors completed a pretrial questionnaire. After
video trial (in lieu of written trial materials); and (d) inviting the trial, non-deliberating jurors completed a post-trial
half the participants who watched the video trial to deliberate in questionnaire about CSA knowledge, rated the credibility of the
jury groups. In addition, the study tested a fresh set of case facts, complainant, the corroborating lay witness (her grandmother),
based on a real CSA case, to determine the external validity or the expert witness (where applicable), the judge, and rendered
generalizability of prior findings to other CSA cases. individual verdicts.
Deliberating participants were allocated to one of 43 juries,
with 10 or 11 juries per experimental condition. Juries, comprised
Research Hypotheses
of 8–12 jurors, were instructed to choose a foreperson, deliberate
This study tested seven hypotheses drawn from the foregoing
as a jury and render a unanimous decision before completing
literature review:
the same post-trial questionnaire. Participants were given a
1. Female jurors, older jurors and better educated jurors will maximum of 90 minutes to reach a verdict. Because the courts
rate the complainant’s credibility more favorably than their excused jurors from jury duty and invited them to participate in

TABLE 1 | Illustration of experimental groups and number of participants in each condition.
No intervention Judicial direction Social framework expert Diagnostic expert Overall
Deliberating juries n = 109 n = 109 n = 109 n = 116 n = 443

Non-deliberating jurors n = 113 n = 115 n = 108 n = 106 n = 442
Overall n = 222 n = 224 n = 217 n = 222 N = 885
the study just before lunch hour, all deliberating participants were professional opinion, is Bridget’s behavior consistent with that
provided with sandwiches. Deliberations lasted between 11 and of a child who has been sexually abused?
87 min (M = 42.43, Mdn = 40.4, SD = 22.49). Experimental psychologist: There are factors in this case which
are consistent with the research findings indicative of child
Study Materials sexual abuse.
A simulated trial was scripted based on an actual CSA
The verbatim testimony of the diagnostic expert was:
case involving a 12-year-old female complainant, Bridget. The
accused, the complainant’s grandfather, was charged with one Prosecution: Based on your experience and your interview with
count of sexual penetration. The case facts were constant in Bridget, in your professional opinion is Bridget’s account of
all experimental conditions. The complainant reported that her events and behavior consistent with that of a child who has
grandfather penetrated her with his finger while she was sitting been sexually abused?
on a chair reading a book in the living room. Her grandmother Clinical psychologist: Yes, it is.
was outside in the garden at the time. Before entering the living
room, the grandmother heard the complainant say, “Grandpa, Questionnaire on Juror Knowledge About Child
stop it, it hurts.” When she entered the room, the complainant’s Sexual Assault
pants were down and the accused was doing up the belt on his A questionnaire to assess participants’ CSA knowledge
pants. The complainant ran to her grandmother, crying, and (Goodman-Delahunty et al., 2017a,b; Reliability: ρy =
made an immediate disclosure of sexual assault. She was 13 years 0.763 ; Cronbach’s α = 0.67) was administered before and
old when she testified at trial. after jurors viewed the simulated trial. The nine items in the
The video trial lasted 40–55 min depending on the questionnaire were based on empirical findings and measured
intervention condition. Professional actors played the roles jurors’ knowledge about behavioral responses to sexual abuse
of the parties, the witnesses and the judge. In all conditions, the and the suggestibility of children. Participants provided their
trial included opening and closing addresses by the prosecution agreement on a 5-point Likert scale from (1) strongly disagree to
and defense, evidence-in-chief and cross-examination of the (5) strongly agree. A lower score indicated less CSA knowledge
complainant and a corroborating witness for the prosecution and greater endorsement of CSA misconceptions. Examples
(the complainant’s grandmother), and the judge’s summing-up. of items measuring the first factor, Impact of Sexual Abuse on
Educative specialized CSA knowledge was presented by Children (Reliability: ρy = 0.70; α = 0.63), are: (a) “A sexually
a social framework expert (an experimental psychologist), a abused child typically cries out for help and tries to escape”; (b)
diagnostic expert (a clinical psychologist) or the presiding judge “A child victim of sexual abuse will avoid his or her abuser”;
during her summing up. The educative information summarized or (c) “Child victims of sexual abuse respond in a similar way
empirical findings on counter-intuitive behaviors of sexually to the abuse.” Examples of items measuring the second factor,
abused children, developmental aspects of children’s memory, Contextual Influences on CSA Reports (Reliability: ρy = 0.80;
their reliability in reporting sexual abuse and suggestibility when α = 0.67), are: (a) “Children are easily coached to make false
questioned by adults (The Supplementary Material contains claims of sexual abuse”; (b) “Repeating questions such as ‘What
the full trial script.) The judge reported these findings but happened? What else happened?’ leads children to make false
made no statement that the behavior of the complainant was abuse claims.”
consistent with that of a sexually abused child. Both experts
presented the educative information after the complainant’s The Witness Credibility Scale
evidence. Both stated explicitly that the complainant’s behavior The Witness Credibility Scale (WCS; Brodsky et al., 2010;
was consistent with that of a child who has been sexually abused2 . Cronbach’s α = 0.95) was used to measure jurors’ perceptions
In addition, the diagnostic expert stated that he had interviewed of the credibility of the witnesses and the judge. The WCS
the complainant. contains 20 semantic differential items measured on a 10-point
The verbatim testimony of the social framework expert was:
3 The reliability coefficient ρy was used because the CSA-KQ questionnaire format
Prosecution: Based on your review of the research findings, and
violated assumptions for Cronbach’s α (Cronbach, 1951; Raykov, 1997, 2002),
your examination of the police interviews of Bridget, in your leading to an underestimation of the reliability of the CSA-KQ (Lord and Novick,
1968; Raykov, 1997, 2002), while ρy was specifically developed to counteract these
2 Prior researchers noted that this language conflates commonality with relevance limitations by considering true variability, error variances, and indicator-construct
(Lyon and Koehler, 1996). We used it because it remains a standard admitted weights (see Goodman-Delahunty et al., 2017a, p. 402). We added Cronbach’s α for
formulation of expert evidence on this topic. readers more familiar with this measure.

continuum, and participants are instructed to rate the witness on 27.86, SD = 5.54). No pretrial differences in CSA knowledge
each of paired contrasting adjectives such as from (1) unfriendly emerged among deliberating groups.
to (10) friendly; (1) dishonest to (10) honest; (1) inarticulate to Preliminary analyses tested the association between the
(10) well-spoken; or (1) illogical to (10) logical. The WCS includes different independent variables (e.g., age, gender, educational
four subscales reflecting the perceived confidence (α = 0.89), level, intervention condition) and dependent variables (e.g.,
likeability (α = 0.86), trustworthiness (α = 0.93), and knowledge knowledge about CSA, perceptions of the witnesses, factual
(α = 0.86) of a witness, respectively. Credibility was assessed by culpability ratings, verdict). Depending on the nature of the
removing one item with the descriptor “scientific” from the WCS variables and the combination of the association, correlations
since this item was inapplicable to lay witnesses. A higher total between continuous, or continuous and categorical variables, χ2
score indicated greater perceived credibility of the witness, with a analyses between categorical or binary variables, and paired t-
maximum possible score of 190. test analyses for pretrial and post-trial CSA knowledge. Further,
Before conducting the analyses, we tested the principal analyses of covariance (ANCOVA) between demographic
component analyses of the WCS with our data, revealing that information, intervention condition and CSA knowledge scores
all items loaded on the same component. In addition, the WCS were conducted, as well as separate between-subject ANCOVAs
subscales were highly correlated with each other, and with the that examined the effects of source of intervention and
overall witness credibility score, creating problems of multi- decision type on juror CSA knowledge and perceived witness
collinearity, i.e., correlations ranged from r = 0.40 to r = 0.68 for credibility. ANCOVA results that yielded different outcomes to
the complainant, and r = 0.50 to r = 0.70 for the grandmother. the multilevel structural equation modeling (SEM; see below) are
Subscale correlations with the overall witness credibility score summarized in the online Supplementary Material.
ranged from r = 0.74 to r = 0.85 for the complainant, and r Multilevel SEM examined the relationship between CSA
= 0.78 to r = 0.88 for the grandmother. For these reasons, and knowledge and verdict at both juror and jury levels to
because the combined measure had higher internal consistency accommodate the intercorrelations in the variables among jurors
than each of the subscales alone (see above), we did not conduct in each jury. The credibility perceptions of both the child
analyses with the separate WCS subscales. complainant and the corroborating witness were modeled as
double mediators to test why and how the impact of CSA
Other Dependent Measures knowledge and expert interventions occurred. Demographic
Juries and individual jurors rendered a binary verdict (guilty/not covariates, pre-trial CSA knowledge, decision type (individual
guilty) and provided demographic information (gender, age, vs. jury group), and intervention source were considered as
educational level, parental status). predictors of post-trial CSA knowledge. The non-independence
of the nested jurors within a jury was addressed with multilevel
modeling using Mplus V8, while variance in the verdict at the jury
Analyses level was estimated. The non-deliberating jurors were analyzed as
For multilevel structural equation modeling with mediation one jury, as were the other 43 deliberating jurors.
analyses, a sample size of 440 was initially suggested based
on simulation guidelines (Wolf et al., 2013) for power of
0.94 regarding the direct path and 0.82 for the indirect path. RESULTS
The conventional rule-of-thumb of 10–15 cases per parameter
Preliminary analyses indicated the presence of 12 out of 885
indicated a sample size of 10 × 15 = 150. All power analysis
multivariate outliers. These participants were excluded from all
results were integrated to select a far more conservative power
further analyses. Scores for the perceived credibility of the experts
strategy, by securing 800 participants, particularly to conduct
and the judge violated normality assumptions (positively skewed
multilevel analysis.
with a positive kurtosis). Accordingly, log transformations of
The impact of the educative intervention independently
the values were performed to achieve a normal distribution
of decision type was assessed by calculating individual CSA
(Tabachnick and Fidell, 2013).
knowledge gain scores after controlling for juror pretrial CSA
knowledge. The CSA knowledge gain score was calculated by
subtracting a juror’s post-trial CSA knowledge score from their Juror CSA Knowledge About Child Sexual
pretrial CSA knowledge score. Negative values indicated that Assault
CSA misconceptions increased after exposure to the videotrial; Analyses to test for differences between juror demographics
positive values indicated that CSA misconceptions decreased (gender, age, educational level), their pre- vs. post-trial CSA
after exposure to the videotrial. Pretrial CSA knowledge scores knowledge and perceived witness credibility revealed that female
were added as covariate because of statistically significant jurors in the sample were somewhat more formally educated than
differences at the outset in mean CSA knowledge scores between their male counterparts, with 63.9% of women and 60.8% of men
some of the experimental groups. For example, before trial, holding a university degree, 20.0% of women and 14.4% of men
non-deliberating jurors who were later exposed to specialized had some tertiary and further education (TAFE) diploma, and
CSA knowledge from a diagnostic expert endorsed statistically 11.0% of men and 1.9% of women held a trade certificate. The
significantly fewer CSA misconceptions (M = 30.74, SD = 5.15) remaining 14.1% of women and 13.8% of men reported finishing
than jurors in the similar non-deliberating control group (M = high school or less. No differences in age emerged between men

and women. On average, older jurors were less educated (overall, Increases and decreases in CSA knowledge within each of the
rs = −0.14, p < 0.001). Jurors with a university degree were 43 deliberating juries revealed that CSA knowledge remained
younger on average (M = 41.79 years, SD = 12.51) than jurors consistent or increased in the 10 deliberating juries, and
who had a TAFE diploma (M = 46.82 years, SD = 14.16) or decreased in one deliberating jury where no specialized educative
who had not finished high school (M = 50.63 years, SD = 11.76). CSA information was presented (hypothesis 2). Furthermore,
The average age of jurors holding a trade certificate (M = 43.42 mean CSA knowledge scores increased statistically significantly
years; SD = 13.73), and of jurors who reported the highest level or tended to increase after deliberation in seven out of 32
of education were similar (M = 44.09 years old; SD = 15.01). juries (21.9%), while mean CSA Knowledge scores remained
Juror CSA knowledge was related to their demographic consistent within the remaining 23 juries (71.9%). Mean CSA
characteristics in several ways. Female jurors had statistically knowledge scores declined statistically significantly in two of the
significant greater CSA knowledge, both pretrial (M = 29.47, 32 juries who were exposed to specialized educative knowledge
SD = 5.31) and post-trial (M = 32.23, SD = 5.31) than male (via a judicial direction and a social framework expert). Table 3
jurors [pretrial: M = 27.93, SD = 4.49, t(857) = 4.59, p < 0.001; presents the pre- and post-trial CSA knowledge change scores
post-trial: M = 30.13, SD = 4.95, t(868) = 5.68, p < 0.001] and and verdicts for each jury deliberation group and for non-
perceived the complainant to be more credible (M = 123.09, SD deliberating jurors.
= 23.10) than did male jurors [M = 117.76; SD = 51.93, t(852)
= −3.20, p = 0.001] (hypothesis 1a). There was no main effect Perceived Witness Credibility
of juror gender on the overall conviction rate [χ 2(1, 869) = 2.09, Table 4 displays the perceived witness credibility of the
p = 0.148, φ = 0.05]. This finding was moderated by decision complainant, her grandmother, the expert witness, and the
type. Whereas there was no effect of juror gender on convictions judge as a function of source of intervention and decision
by deliberating jurors [women: 31.1%; men: 31.7%; χ2(1, 436) = type (For detailed statistical analyses, consisting of ANCOVAs,
0.02, p = 0.894, φ = −0.01], women who rendered an individual see Supplementary Material). In the absence of any specialized
verdict were significantly more likely to convict (53.3%) than information, deliberating jurors perceived the complainant to
their male counterparts [41.0%; χ2(1, 433) = 6.27, p = 0.012, φ be more credible than did non-deliberating jurors. Among
= 0.12]4 . non-deliberating jurors exposed to diagnostic expert evidence,
Juror age was positively correlated with perceived the perceived credibility of the complainant exceeded that of
complainant credibility, such that older jurors rated the non-deliberating jurors in the control group (hypothesis 3).
complainant as more credible than younger jurors, r = 0.17, Specialized CSA knowledge presented by the judge or by the
p < 0.001 (hypothesis 1b). Further, juror age was positively social framework expert did not affect the perceived credibility
correlated with CSA knowledge gains between the time of the of the complainant. When participants deliberated as a jury,
pre- and post-trial measures, r = 0.10, p = 0.005. This effect was the perceived credibility of the complainant was constant in all
moderated by decision type, such that CSA knowledge gain was experimental groups. These results are displayed in Figure 1A.
associated with age only among non-deliberating participants, Similarly, credibility of the grandmother, the corroborating
r = 0.14, p = 0.005. There was no correlation between age and witness, was greater when specialized educative information was
CSA knowledge gain among participants who deliberated as a presented by an expert witness of either type than by the judge
jury, p > 0.10. Finally, juror educational level was correlated in a judicial direction. These results are displayed in Figure 1B.
with CSA knowledge both pretrial (overall: rs = 0.19, p < Error bars are 95% confidence intervals.
0.001) and post-trial (overall: rs = 0.18, p < 0.001), such that Decision type was not associated with the expertise of the
jurors with higher educational qualifications held fewer CSA psychologist, although non-deliberating jurors rated the experts
misconceptions. Education was, however, not correlated with as more credible (M = 162.89, SD = 17.63, Mdn = 165) than
CSA knowledge gains or perceived complainant credibility, p > did deliberating jurors (M = 158.69, SD = 19.03, Mdn = 161)
0.05 (hypothesis 1c). Neither juror age nor education level was (hypothesis 4). By contrast, the perceived credibility of the judge
associated with the conviction rate, independently of the decision decreased when the judge provided specialized information in
type. Table 2 displays correlations between juror demographic a judicial direction (M = 163.24, SD = 24.55, Mdn = 169)
characteristics, CSA knowledge and perceived complainant compared to trials in which the same information was provided
credibility (measured by the WCS) separately for deliberating by an expert, irrespective of whether the expert was a social
and non-deliberating mock-jurors. framework (M = 168.89, SD = 17.82, Mdn = 172) or diagnostic
Juror CSA knowledge was assessed pre- and post-trial using psychologist (M = 168.51, SD = 17.68, Mdn = 173). In the
the CSA Knowledge Questionnaire. Before trial, jurors had a two latter conditions, the judge was rated more credible than
moderate degree of CSA knowledge, with average scores of ∼28 the experts. However, the judge was perceived as more credible
out of a possible total of 45. The group scores ranged from a low by non-deliberating jurors (M = 167.73, SD = 19.84, Mdn =
of M = 27.86, SD = 5.54 (Control, non-deliberation), to a high of 171) than by deliberating jurors (M = 166.05, SD = 19.08, Mdn
M = 30.74, SD = 5.15 (Intervention 3, non-deliberation). = 171).
The Impact of CSA Knowledge and

4 Differences in degrees of freedom arose for two reasons: (a) some participants Interventions on Verdict
did not answer all questions, and (b) findings are presented for subsamples Results revealed an increase in the conviction rate when jurors
(deliberating juries vs. non-deliberating jurors). rendered individual verdicts without deliberation following

TABLE 2 | Intercorrelations between juror demographic characteristics, CSA knowledge, and perceived complainant credibility.
1 2 3 4 5 6
1. Juror age – −0.11* −0.01 −0.03 0.06 0.15**

2. Juror education −0.15** – 0.19** 0.23** 0.01 0.08
3. Pre-trial CSA knowledge −0.05 0.17** – 0.63** −0.39** 0.17**
4. Post-trial CSA knowledge 0.04 0.16** 0.69** – 0.47** 0.33**
5. CSA knowledge gain 0.14** −0.00 −0.33** 0.46** – 0.23**
6. Complainant credibility 0.16** −0.08 0.26** 0.39** 0.20** –
Correlations below the diagonal are for non-deliberating jurors; correlations above the diagonal are for deliberating jurors. Higher numbers for education indicate more formal education.
For juror education the statistic is Spearman’s Rho, the remainder are Pearson correlations. *p < 0.05; **p < 0.01.
TABLE 3 | Mean pre- and post-trial CSA knowledge and verdict for each deliberating jury and non-deliberating jurors, by experimental condition.
No intervention Judicial direction Social framework expert Diagnostic expert
Mpre Mpost Verdict Mpre Mpost Verdict Mpre Mpost Verdict Mpre Mpost Verdict
Deliberating juries
1 27.78 29.67 6NG 3G 28.00 31.00 9NG 25.25 27.75 12NG 29.17 31.00 12NG
2 27.70 27.80 10NG 26.89 29.11 9NG 29.67 31.50 12NG 29.00 30.67 11NG 1G
3 27.87 31.25 8G 28.55 34.08 12G 29.91 33.45 11G 28.42 32.67 5NG 7G
4 30.11 30.33 2NG 7G 29.40 30.90 10NG 31.82 33.08 9NG 3G 29.33 29.78 9NG
5 28.08 29.91 12G 26.92 33.25 10NG 2G 30.13 32.13 1NG 7G 29.22 36.20 10G
6 27.42 29.67 1NG 11G 28.44 30.89 9NG 28.13 31.25 8NG 28.60 32.55 10NG 2G
7 29.56 32.90 9NG 1G 30.20 29.40 5NG 5G 26.36 28.09 12NG 27.25 29.75 12NG
8 29.14 26.80 9NG 1G 26.75 28.89 8NG 1G 29.82 35.18 1NG 11G 27.63 34.38 8G
9 27.38 27.25 7NG 1G 27.27 29.73 5NG 6G 29.36 30.67 12NG* 27.55 29.64 11NG
10 27.00 28.00 9NG 30.44 32.33 9NG 27.30 29.10 10NG 26.64 31.55 10NG 1G
11 26.42 27.08 5NG 7G 30.00 31.38 7NG 2G – – – 27.00 30.37 8G
All 28.03 29.12 58NG 51G 28.41 31.10 81NG 28G 28.74 31.20 77NG 32G 28.18 31.62 80NG 37G
Non-deliberating jurors
All 27.97 27.87 68NG 44G 28.48 32.38 64NG 51G 28.14 31.07 58NG 49G 30.65 33.46 48NG 58G
G, guilty; NG, not guilty. * Some jurors pressured others to change their views.
TABLE 4 | Perceived credibility of witnesses and the judge by decision type and experimental group.
Decision type Source of educative intervention
No intervention Judicial direction Social framework expert Diagnostic expert
Mean SD Mean SD Mean SD Mean SD
Non-deliberation
Complainant 115.37 29.12 118.86 24.46 118.11 22.83 126.42 23.98
Grandmother 118.35 27.01 114.90 24.57 122.04 26.01 125.97 25.86
Expert – – – – 160.86 18.50 165.19 16.63
Judge 168.00 15.39 162.83 28.07 169.22 15.32 171.16 16.42
Deliberation
Complainant 123.35 22.11 118.43 25.62 121.46 22.44 118.19 20.16
Grandmother 117.89 23.18 113.18 24.30 121.44 21.99 114.54 22.81
Expert – – – – 160.81 18.71 156.33 20.04
Judge 165.91 17.31 164.68 20.30 168.57 20.11 166.09 18.49
The descriptor “scientific” was removed from the WCS to enhance comparability of scores; maximum possible score = 190.
exposure to specialized knowledge from either the judge (44.3%), (38.9%). However, these differences were not statistically
the social framework expert (45.4%) or the diagnostic expert significant [χ 2 (3, N = 440) = 5.40, p = 0.145]. The verdicts
(54.7%), compared to the no intervention control condition of deliberating jurors showed that in comparison with the

FIGURE 2 | Guilty verdict (%) by decision type and source of educative

intervention.
knowledge presented in a judicial direction, one jury with the

highest post-trial CSA knowledge (M = 34.08, SD = 6.07)
convicted; six juries with mixed levels of CSA knowledge
acquitted (five unanimously), and the remaining four juries were
hung. Following exposure to specialized knowledge presented
by a social framework expert, three juries with higher CSA
knowledge scores convicted (one unanimously), six acquitted,
although in two of the six juries CSA knowledge scores were high,
and one jury was hung. Similarly, after exposure to specialized
knowledge presented by a diagnostic expert, three juries with
moderate to high CSA knowledge scores convicted, six juries in
which overall levels of CSA knowledge were low acquitted (four
unanimously), and two juries were hung.
FIGURE 1 | Perceived credibility of the complainant (A) and the grandmother
(B) by decision type and source of educative intervention. Error bars are 95% The Impact of CSA Knowledge,
confidence intervals. Interventions, and Witness Credibility on
Verdict
Multilevel SEM analyzed psychological mechanisms reflecting
the impact of CSA knowledge, educative interventions and
control group (43.1%), the individual conviction rate dropped deliberation effects. Tables 5, 6 show hierarchical multilevel
statistically significantly following exposure to all types of regression SEM results with statistically significant path
interventions [χ 2 (3, N = 440) = 11.08, p = 0.011], as shown in coefficients explaining the impact of CSA knowledge on witness
Figure 2 (hypothesis 5). These results were further qualified by credibility and the subsequent final verdict. Model 1 was a simple
more advanced analyses as described in the following sections. regression model to explain post-trial CSA knowledge. In most
When considering jury group verdicts, almost a quarter of models, the three intervention conditions were contrast coded
deliberations (23.3%) resulted in hung juries (no unanimous or with the control group and multiplied with the deliberation
majority decision was reached in the available time). Overall, variable to test the interaction effects. Models 1–3 tested the
44.2% of juries reached a unanimous verdict and three accepted significance of predictors of post-trial CSA knowledge that
a majority decision, as shown in Table 2. in turn examined its impact on the credibility of the child
In the absence of educative information, three juries voted complainant and the corroborating witness.
unanimously to convict the accused, five juries acquitted (two In Model 1, the interaction effects were omitted due to
unanimously) and three juries were hung. Juries with less CSA convergence issues at Level 1, the individual juror level.
knowledge voted to acquit. Following exposure to specialized Demographic variables of gender, age, education, and pretrial

TABLE 5 | Multilevel SEM analysis predicting CSA knowledge and witness credibility: unstandardized coefficients (b).
Predictor/model fit Model 1 Model 2 Model 3
Juror level predictors of post-trial-CSA-K

Juror age 0.02** 0.02** 0.02*
Female juror 1.05** 1.08** 1.08**
Education 0.27** 0.28** 0.29**
Deliberation −0.44 0.50 −0.54
Intervention contrast
Control vs. Intervention (C1) 0.90** 0.90** 0.90**
Judicial instr. vs. Experts (C2) 0.23** 0.23** 0.23**
Social vs. Diagnostic expert (C3) 0.47** 0.37** 0.37**
Interaction: Deliberation × Intervention
Deliberation × C1 −0.52** −0.48**
Deliberation × C2 −0.31 −0.32
Deliberation × C3 0.13 0.79
Pretrial CSA-K 0.65** 0.65** 0.65**
Juror level paths
Post-trial CSA-K→WCS Victim 1.83**
Post-trial CSA-K→WCS Witness 1.77**
Jury level predictors of post-trial CSA-K
Control vs. Intervention (C1) 0.40** 18.64** −10.80
Judicial instr. vs. Experts (C2) −0.12 11.22** 9.67
Social vs. Diagnostic expert (C3) 0.33 3.43** 65.25
CSA-K intercept 11.72** 11.25** 11.18
CSA-K variance 0.77 0.63 0.32
Model fit: Loglikelihood −5594.51 −9152.18 −9193.78
Akaike information criterion 11217.05 18344.15 18427.47
Sample-size adjusted BIC 11238.90 18375.57 18458.96
CSA-K, Child sexual abuse knowledge; Education on a 5-point scale from 1 “less than Year 12 certificate” to 5 “University degree or higher.” Deliberation: No jury deliberation = −1.
Jury deliberation = 1. Control vs. Intervention (C1): Control group = −3, Three intervention groups = 1 × 3. Judicial direction vs. Experts (C2): Judicial direction group = 2, Social
Framework Expert and Diagnostic Expert = 1 × 2. Social vs. Diagnostic expert (C3): Social Framework Expert = −1, Diagnostic Expert = 1. WCS, Witness Credibility Scale; BIC,
Bayesian Information Criterion. *p < 0.05; **p < 0.01; Number of jurors ranged from 815 to 835; Number of juries = 44.
CSA knowledge were statistically significant and explained the The impact of the interventions was stronger when jurors did
variance in post-trial CSA knowledge. However, decision type not deliberate: jurors without any intervention had lower CSA
did not predict post-trial CSA knowledge, while all intervention knowledge scores (M = 27.87, SD = 5.67) after trial than
contrast variables were statistically significantly associated with jurors exposed to one of the interventions (M = 32.30, SD =
post-trial CSA knowledge. Specifically, jurors in all three 5.40). Among jurors who deliberated, post-trial differences in
intervention groups had statistically significantly higher CSA CSA knowledge scores between the intervention groups (M =
knowledge after the trial than the control group jurors. Jurors 31.32, SD = 4.95) and the control group (M = 29.12, SD =
were more persuaded by the judicial direction (gained more CSA 4.78) were smaller. However, the judicial direction compared to
knowledge) than by the expert testimony, while the diagnostic the expert witness intervention predicted more post-trial CSA
expert intervention predicted higher CSA knowledge scores knowledge, while the diagnostic expert was more persuasive (as
after trial than the social framework expert intervention. Yet, shown in higher post-trial CSA knowledge scores) than the social
at Level 2, the jury level, only juries in the three intervention framework expert. At the jury level, though, the control group
groups showed more CSA knowledge than the control group showed a higher level of post-trial CSA knowledge compared
juries after the trial. The positive effects of the judicial direction to the three intervention groups, while the other intervention
and diagnostic expert were not statistically significant at the effects were in the same direction as at the juror level. When the
jury level. perceived credibility of the corroborating witness was included
In Model 2, all demographic variables were statistically in Model 3, it was statistically significantly explained by post-
significantly associated with post-trial CSA knowledge while trial CSA knowledge. The impact of all other variables and
the intervention contrast between the control group and interaction effects produced statistically similar results. Because
intervention groups was moderated by the impact of deliberation. the jury level effect of the intervention was no longer statistically

TABLE 6 | Multilevel mediation SEMs predicting conviction with post-trial CSA knowledge and mediators of witness credibility: unstandardized coefficients and odd ratios.
Predictor/model fit b or Odds Ratio (OR)
Model 4 Model 5 Model 6 Model 7
Juror level predictors on post-trial-CSA-K

Juror age 0.03** 0.03** 0.03** 0.03**
Female jurors 1.07** 1.08** 1.08** 1.08**
Education 0.28** 0.28** 0.28** 0.28**
Deliberation −0.21 −0.22 −0.22 −0.22
Intervention contrast
Control vs. Intervention (C1) 0.90** 0.90** 0.90** 0.90**
Judicial vs. Experts (C2) 0.23** 0.23** 0.23** 0.23**
Social vs. Diagnostic expert (C3) 0.36** 0.35** 0.36** 0.36**
Interaction: Deliberation × Intervention
Deliberation × C1 −0.47** −0.48** −0.47** −0.47**
Deliberation × C2 −0.31 −0.30 −0.30 −0.30
Deliberation × C3 0.09 0.10 0.09 0.09
Pretrial-CSA-K 0.66** 0.65** 0.65** 0.65**
Juror level paths
Post-trial-CSA-K→WCS Victim 1.80** 1.79** 1.79**
Post-trial-CSA-K→WCS Witness 1.75** 1.76** 1.76**
Post-trial-CSA-K→Convict 1.08**(OR) 1.08**(OR) 1.06**(OR)
WCS Victim→Convict 1.04**(OR) 1.02**(OR) 1.01**(OR)
Indirect via WCS Victim 1.07**(OR) 1.03**(OR) 1.03**(OR)
WCS Witness→Convict 1.05**(OR) 1.04**(OR) 1.04**(OR)
Indirect via WCS Witness 1.09**(OR) 1.07**(OR) 1.07**(OR)
Jury level
Convict threshold 7.88** 9.01** 8.47** 9.67**
Convict between-juries variance 3.57** 3.85** 4.12** 3.50**
Model fit: Loglikelihood −6517.51 −6536.03 −10350.61 −10345.04
Akaike information criterion 13075.02 13112.06 20747.22 20738.07
Sample-Size Adjusted BIC 13106.43 13143.48 20783.35 20775.77
CSA-K: Child sexual abuse knowledge; Education on 5-point scale from 1 “less than Year 12 certificate” to 5 “University degree or higher.” Deliberation: No jury deliberation = −1.
Jury deliberation = 1. Control vs. Intervention (C1): Control group = −3, Three intervention groups = 1 × 3. Judicial direction vs. Experts (C2): Judicial direction group = 2, Social
Framework Expert and Diagnostic Expert = 1 × 2. Social vs. Diagnostic expert (C3): Social Framework Expert = −1, Diagnostic Expert = 1. WCS: Witness Credibility Scale. BIC,
Bayesian Information Criterion. ** p < 0.01; Number of jurors ranged from 815 to 835; Number of juries = 43. CI: Model 7: 95% confidence intervals of the coefficients—Age, 0.004
0.04; Female jurors, 0.65–1.52, Education, 0.11–0.45, Deliberation, −0.66 0.22, C1, 0.88–0.91; C2, 0.21–0.25; C3, 0.30–0.42; Deliberation × C1, −0.71–−0.24; Deliberation × C2,
−0.72 0.11; Deliberation × C3, −0.50–0.69; Post-trial-CSA-K→WCS Victim, 1.52–2.06; Post-trial-CSA-K→WCS Witness, 1.47–2.05; Post-trial-CSA-K→Convict, 1.04–1.09; WCS
Victim→Convict, 1.01–1.02; Indirect via WCS Victim, 0.01–0.04; WCS Witness→Convict, 1.03–1.05; Indirect via WCS Witness, 1.05–1.09. Confidence intervals for other models are
available by request to the authors (omitted due to space limitations).
significant in Model 3 due to the large standard errors of the associated with higher perceived complainant credibility, which
parameter estimates, the variables (i.e., predictors of Post-trial in turn increased the odds ratio to convict. When the credibility
CSA-Knowledge: C1, C2, C3) were omitted from Models 4–7. of the grandmother as a corroborating witness was added, Model
Models 4–7 further examined the relationship between post- 5 showed similar results. Post-trial CSA knowledge was directly
trial CSA knowledge and guilty verdicts in a dual mediation associated with an increased odds ratio to convict with its
model, with credibility ratings of the complainant and her statistically significant indirect effect through perceived witness
grandmother as mediators, and pre-trial knowledge scores credibility. Further, Model 6 and Model 7 compared whether the
and demographic variables as covariates. Model 4 yielded a direct effect of CSA knowledge on the odds ratio to convict was
statistically significant direct effect of post-trial CSA knowledge statistically significant when both credibility variables were tested
on conviction and a statistically significant indirect effect through as dual mediators (Figure 3). In terms of model fit, Model 4 fit the
victim credibility as a psychological mediator. All other effects, data best, with the smallest Akaike Information Criterion (AIC)
including the interaction effects, remained similar to those in and sample-size adjusted Bayesian Information Criterion (BIC)
the previous model. Higher post-trial knowledge scores were among Models 2–5.

FIGURE 3 | Multilevel structural equation modeling: unstandardized path coefficients (bs and Odds ratios in Model 7) for victim and witness credibility mediators of the
impact of CSA knowledge on guilty verdict.
To further test the dual mediation (hypothesis 4), Model 6 in their CSA knowledge score. This path from post-trial CSA
and Model 7 were compared. The results showed that Model knowledge to conviction was statistically significantly mediated
7 was superior to Model 6 with a better model fit, that is, by the perceived credibility of both the complainant and the
smaller AIC and sample-size adjusted BIC, which suggested corroborative witness. First, a one-point increase in the post-trial
partial mediation of the credibility of both the child victim CSA knowledge score was statistically significantly associated
and the corroborating witness. Partial mediation was supported with a score higher by 1.79 points in ratings of the credibility of
because of the statistically significant paths of both the direct the child complainant and with a score higher by 1.76 points in
impact of CSA knowledge and the indirect impact through the rating the credibility of the corroborative witness. Third, higher
perceived credibility upon conviction rates. Older jurors, female ratings of the perceived credibility of the child complainant in
jurors, more educated jurors, and jurors already knowledgeable turn yielded statistically significantly greater odds of conviction.
about CSA were more likely to show greater post-trial CSA That is, jurors were 1.01 times more likely to convict per each
knowledge. Similar patterns to the results in Model 2 were increase of a single point on the credibility scale. Fourth, jurors
observed in this final dual mediation model regarding the impact with a score higher by one point on the perceived credibility of
of interventions during the trial. The impact of all the three the corroborative witness were 1.04 times more likely to convict.
interventions was moderated by deliberation with a statistically These four statistically significant paths comprised statistically
significant interaction effect, deliberation × C1. Specifically, significant indirect effects of CSA knowledge via the perceived
the intervention effect of specialized educative information on credibility of the child complainant, 1.03 times more likely to
CSA knowledge was stronger when jurors did not deliberate convict, and via the perceived credibility of her grandmother,
than when they did. The remaining two interaction terms 1.07 times more likely to convict. This finding suggested a double
were not statistically significant, supporting the main effects of partial mediation, that is, higher post-trial CSA knowledge
interventions. In particular, the judicial direction predicted a was associated with a higher likelihood of a conviction on its
higher level of post-trial CSA knowledge in jurors than other own (a statistically significant direct effect) and also indirectly
expert interventions with the positive coefficient of the contrast through credibility perceptions of both prosecution witnesses
dummy variable, C2. When interventions by the diagnostic (two statistically significant indirect effects).
expert and social framework expert were compared with the In light of the two-factor structure of the CSA knowledge
C3 contrast variable, more post-trial CSA knowledge emerged measure, supplementary multilevel SEM was conducted. Similar
among jurors exposed to the diagnostic expert. partial mediation results emerged: when the first factor of the
In turn, jurors with greater post-trial CSA knowledge were CSA-KQ (the subscale score on the Impact of CSA on Children),
∼1.06 times more likely to convict per each point of increase was used as the main predictor variable, compared to total

CSA-KQ scores, the larger impact of witness credibility on the videotrial, replicating previous findings with Australian
convictions (odds ratios up to 1.06) was statistically significant. community volunteers (Goodman-Delahunty et al., 2010).
Regarding the second factor (subscale scores on Contextual Further, CSA knowledge was significantly correlated with
Influences on CSA Reports), a similar increase in the impact jurors’ formal education levels, such that jurors with higher
of witness credibility (odds ratios up to 1.16) was statistically educational qualifications had statistically significantly more
significant. Both models showed better fit to the data than Model accurate CSA knowledge. Women and older jurors gained
7 in the main analysis. most benefit from the educative trial interventions, reflected in
their higher knowledge gain scores and post-trial reductions in
CSA misconceptions, compared to male and younger jurors.
DISCUSSION These findings on demographic influences in understanding
The present study explored jurors’ pre-trial attitudes toward and responding to CSA knowledge remained consistent in the
complainants in CSA cases and tested the effects of educative multilevel SEMs. Moreover, women and older jurors rated the
interventions to increase jurors’ CSA knowledge by providing complainant statistically significantly more credible than did men
specialized information within a realistic trial setting. Jurors and younger jurors.
attending the District and Supreme Courts in NSW participated Similarly, conviction rates were related to juror gender, such
in the study, and verdicts rendered by non-deliberating that women rendering individual decisions were more likely to
individual jurors were compared with those of jurors who convict the accused than were their male counterparts. A similar
deliberated in jury groups. pattern has been shown meta-analytically in cases involving CSA,
Inspections of jury groups revealed that CSA knowledge adult sexual assault (Schutte and Hosch, 1997) and other types of
gains varied by group both between and within each of the criminal cases (Devine and Caughlin, 2014), namely that women
experimental conditions. These findings indicated that the are more prone to convict than men. This effect disappeared in
combination of the variability in juror CSA misperceptions the present study when jurors deliberated in groups to a verdict.
at the outset of the experiment, the intervention source, and These results indicated that before deliberation, women had a
the decision process were related to final levels of juror CSA higher propensity than men to convict, but the influence of
knowledge and verdicts. deliberation with other jurors exerted a more powerful effect on
Multilevel SEM analyses presented novel insights into their verdict than juror demographic characteristics.
psychological mechanisms showing why and how jurors’ CSA The second hypothesis was confirmed, namely that
knowledge impacted their verdicts. When the variance in jurors’ independently of the educative interventions, jurors who
verdicts was analyzed, at both juror and jury levels, the impact arrived for jury duty with numerous CSA misconceptions would
of CSA knowledge was partially mediated by the perceived rate the complainant low in credibility and tend to acquit the
credibility of both the child complainant and the corroborative accused, while jurors with fewer CSA misconceptions at the
witness. Greater CSA knowledge after the interventions increased outset would rate the complainant’s credibility more favorably
the odds ratio to convict by itself, whereas greater CSA knowledge and be more inclined to convict the accused. Juror pre-trial CSA
was also associated with heightened credibility perceptions as knowledge was positively correlated with perceived complainant
psychological mechanisms which in turn further predicted a credibility, showing a medium to strong effect. As expected,
higher odds ratio to convict the defendant. In these mediation less CSA knowledge was associated with lower credibility
models, older age, being female, more formal education, and ratings of the complainant. Moreover, jurors’ CSA knowledge
more pre-trial CSA knowledge were associated with higher predicted the perceived credibility of the complainant, the
CSA post-trial knowledge scores. Whether or not the jurors corroborating witness, and verdicts, such that jurors with more
had deliberated moderated the effect of interventions, such CSA knowledge were more likely to convict. These results were
that the interventions had a greater impact on non-deliberating consistent in the multilevel models, supporting previous findings
jurors. The judicial direction had a greater impact on post- obtained with psychology undergraduates and community
trial CSA knowledge than diagnostic or social framework expert members, confirming that jury eligible citizens with greater
interventions, while the diagnostic expert was more persuasive CSA knowledge were more likely to assess a child complainant
than the social framework expert. These results are discussed in as credible and more likely to convict (Gabora et al., 1993;
relation to the research hypotheses. Goodman-Delahunty et al., 2010, 2017b).
The Influence of Juror Demographic The Impact of CSA Knowledge on the

Characteristics on the Perceived Perceived Credibility of the Witnesses
Credibility of the Complainant and Verdict Although deliberating jurors who were not exposed to any
Our first hypotheses, that juror demographic characteristics educative CSA knowledge were less likely to convict than non-
would affect perceived witness credibility assessments and deliberating jurors in the parallel control condition, deliberating
verdict, were partially supported in a number of ways. jurors in the control condition perceived the complainant as
Correlations between juror gender and CSA knowledge showed more credible than their non-deliberating counterparts. While
that women’s knowledge of CSA was statistically significantly deliberation appeared to enhance credibility perceptions in the
greater than that of men, before and after exposure to control condition, that effect did not translate into convictions,

suggesting that the deliberation process and possibly group additional variance in CSA knowledge acquisition and final
consideration of the meaning of the criminal standard of proof verdict decisions.
(MacCoun and Kerr, 1988; Wright and Hall, 2007) increased As a novel analytic attempt, the mediation models were tested
jurors’ doubt of the accused’s guilt, or their willingness to using multilevel modeling. These types of rigorous methods
convict, even when they perceived him to be factually culpable. need to be adopted widely to deal with the troublesome
Reluctance by deliberating juries to convict the accused in a CSA interdependence in juror data given their nested nature (Lovis-
case in the face of perceived factual culpability is not unusual McMahon, 2015; Peter-Hagene et al., 2019).
(Goodman-Delahunty and Martschuk, 2020).
Unexpectedly, the educative interventions had little impact on
the perceived credibility of the complainant among deliberating The Impact of Educative Interventions on
jurors: ratings of the complainant’s credibility were constant Juror CSA Knowledge
in all experimental conditions. For non-deliberating jurors, The third hypothesis that the educative interventions would
however, the perceived credibility of the complainant was highest increase jurors’ CSA knowledge was confirmed. In the absence
in response to the diagnostic expert condition and lowest in of any intervention, both deliberating and non-deliberating
the control condition. Further, for non-deliberating jurors, the jurors endorsed fewer CSA misconceptions after viewing the
grandmother’s credibility was influenced more favorably by the videotrial than before, and the decrease in CSA misconceptions
two expert witnesses than by the judicial direction, although no was greater for non-deliberating than for deliberating jurors.
similar effect emerged for the credibility of the complainant. As Similarly, analyses of jurors’ CSA knowledge scores showed that
noted above, the educative interventions impacted the verdict, jurors in the control group acquired less accurate information
but were not the sole predictors of verdict. about CSA in the course of the trial than did their counterparts
The mediation analyses provided an explanation for these who were exposed to specialized educative information.
results. Specifically, it revealed that the extent of jurors’ post- Previous research showed similar statistically significant
trial CSA knowledge predicted the perceived credibility of the increases in the CSA knowledge of jury eligible citizens who
complainant and the grandmother, in parallel with the source were exposed to specialized knowledge interventions (Goodman-
effects of the educative intervention. The interaction between Delahunty et al., 2011a). Unlike the present study, CSA
deliberation and interventions was partially supported in that knowledge in the control group in our previous study was
the impact of judicial or expert interventions was stronger in unchanged post-trial. That study differed in a number of respects
non-deliberating jurors when the variance in the final verdict that may account for this difference, including juror demographic
was analyzed at both juror and jury levels. These findings characteristics (jury eligible students and community members
emphasize the importance of interventions to enhance jurors’ vs. jurors), the presentation mode (written vs. video-recorded
CSA knowledge during the trial while jurors who are deliberating trial) and decision type (individual non-deliberating jurors vs.
may find their group discussion more persuasive than the deliberating juries). The findings in the present study more
expert interventions. More importantly, the mediation analysis closely approximate real juror pre-trial CSA knowledge and
showed that the combination of CSA knowledge scores and responses to educative interventions. Although the educative
increases in the perceived credibility of the complainant and information did not statistically significantly increase CSA
her grandmother was statistically significantly associated with knowledge among the non-deliberating jurors, this information
guilty verdicts. The more jurors knew about CSA post-trial, the protected these juries from susceptibility to CSA misconceptions
higher the perceived credibility of both the complainant and observed to increase among juries in the control condition who
her grandmother, the more likely the jurors were to convict were not exposed to this information.
the accused. Because both the direct impact of CSA knowledge The post-trial increase in juror CSA misconceptions in the
on verdict and the indirect effects through two credibility control condition in which no specialized educative information
scores were statistically significant, the mediation was partial was presented, regardless of whether they deliberated, is an
rather than full. However, the indirect paths through credibility important finding since it is likely to reflect the everyday
variables showed stronger effects than the direct effects. Hence, trial circumstances in the majority of CSA trials conducted in
the credibility perceptions as mechanisms of CSA knowledge Australia and elsewhere. In other words, prosecutors typically
impacting on verdict appeared important, and call for theoretical do not call expert witnesses to provide educative information to
and practical attention. Prior research demonstrated that the jurors about the counter-intuitive behaviors of sexually abused
credibility of a complainant who was the sole prosecution children and the reliability of child witnesses, nor do judges
witness apart from the expert witness, mediated the effect routinely provide this information in a judicial direction. One
of CSA knowledge change on verdict (Goodman-Delahunty potential source of the observed post-trial increase in CSA
et al., 2011a). The interaction impact of jury deliberation may misconceptions was the stereotypical misconceptions introduced
indicate that other factors that were not measured in this by the defense lawyer during her vigorous cross-examination
model should be considered in future studies. Specifically, of the complainant in the simulated trial. For example, the
unique factors within each of the deliberating juries (such complainant was asked about continuing to live with her
as interpretations of the evidence and the criminal standard grandfather after the alleged assault and she confirmed that he
of proof, the overall extent of CSA misconceptions in the took her to school each day. Defense counsel also asserted that the
group, and group-specific dynamic factors) may further explain complainant had given three conflicting versions of the events,

that she had fabricated her assault allegation after coaching by diagnostic expert who stated that he had interviewed the child,
her grandmother, and had the knowledge to do so because she than in response to the social framework expert who had only
had attended sex education classes at school. Thus, jurors’ CSA reviewed the police records. However, among deliberating jurors,
misconceptions, and those introduced by defense lawyer during the number of juries voting to convict the accused did not differ
cross-examination of the complainant and other witnesses, were in response to the type of expert witness.
left unchallenged. Absent exposure to educative information to
correct CSA misconceptions, these misconceptions intensified
when jurors deliberated in jury groups to a verdict. The Persistence of CSA Misconceptions
Following Jury Deliberation
The hypothesis, that CSA misconceptions would decrease after
The Impact of Type of Expert Witness on exposure to one of the three educative interventions, was
Juror CSA Knowledge and Verdict partially confirmed. CSA knowledge scores of jurors in all non-
The fourth hypothesis, that jurors would prefer a diagnostic deliberating conditions either increased slightly or remained
over a social framework expert and that trials involving the stable, unlike those of their deliberating counterparts. Following
diagnostic expert would yield statistically significantly more deliberations, CSA knowledge persisted at a level equivalent to
convictions compared to other educative interventions, was pre-trial CSA knowledge in all intervention groups. The increase
partially confirmed. The simple difference test in jurors’ in CSA misconception scores of deliberating jurors who were
individual verdicts showed the effect of educative judicial and not exposed to any educative information (control condition)
expert intervention did not support the hypothesis. However, far exceeded that of jurors in other deliberating groups. While
further advanced models showed that the control group without the increase in CSA misconceptions was moderated by the
any educative intervention had significantly lower post-trial CSA presence and source of educative interventions, these findings
knowledge which in turn decreased witness credibility perception demonstrated that deliberation did not reliably reduce juror
and subsequently the odds ratio to convict the defendant when errors and CSA misconceptions. In some juries, CSA knowledge
the interaction effect of deliberation and educative intervention increased after a discussion of the case facts as a group, but
and demographic covariates such as education and gender were other juries endorsed CSA misconceptions in the course of their
taken into account, particularly with the variance in the verdict deliberations. Deliberation provided an opportunity for many
variable analyzed both at the juror and jury levels with the jurors to repeat and reinforce CSA misconceptions introduced
strength of multilevel modeling. by defense lawyer during cross-examination of the complainant,
Jurors’ assessments of the experts as measured by the WCS or by other deliberating jurors. These findings reflect the impact
revealed that the perceived credibility of the social framework within each jury of unmeasured factors arising from group
and the diagnostic experts was equivalent. Ratings of their dynamics, such as cohesion (Cialdini and Goldstein, 2004),
credibility may have been similar because the study did not vary norms arising within each group (Schulz et al., 2007), or
attributes of the experts intrinsic to the WCS. Both experts were the impact of dominant individual jurors on group decisions
portrayed by the same actor who conveyed similar evidence (Gordon, 2014).
in a uniform manner in direct and cross-examination. The The frequency of hung juries was greater among deliberating
experts’ level of education and experience in the field establishing jurors who received specialized educative information from the
pertinent credentials were parallel. Thus it is understandable that judge in a judicial direction compared to those who received it
they were perceived by jurors as equivalently likable, confident, from an expert witness. This finding suggested that jurors were
trustworthy and knowledgeable. In testifying, both stated that more polarized by educative information in the form of a judicial
the behavior of the complainant was consistent with that of a direction compared to that provided by an expert witness, whose
sexually abused child, although the social framework expert’s opinion they could disregard.
statement was more cautious than that of the diagnostic expert, The post-trial persistence of CSA misconceptions in
perhaps unnecessarily so. The major difference between the deliberating jurors who were exposed to three different sources
experts was that one reviewed only police records (the social of educative interventions was unexpected. These findings
framework expert), whereas the other reviewed these records may be due to the persistence of discredited information
and personally interviewed the complainant (diagnostic expert). (Anderson et al., 1980), a confirmation bias (Nickerson, 1998),
The fact that the diagnostic expert interviewed the complainant or attitude polarization (Myers and Lamm, 1976), all of which
appeared to enhance the credibility ratings of the complainant. have been tested and observed previously in the context of
Specifically, educative information presented by the diagnostic mock-jury research on topics other than CSA (Salerno et al.,
expert increased the perceived credibility of the complainant 2017; Peter-Hagene et al., 2019). Alternatively, misconceptions
compared to ratings by jurors in the control condition, whereas in statements by the jury foreperson or other influential jurors in
educative information presented by the social framework expert the group may have dominated the discussion (Gordon, 2014).
or the trial judge did not impact the perceived credibility of The findings are also consistent with deliberation theories such
the complainant. However, this effect was statistically significant as the liberation hypothesis which postulates that when the
only for non-deliberating jurors. The conviction rate revealed a evidence is ambiguous, jurors resort to extra-legal information
similar pattern of results: both deliberating and non-deliberating such as their own experiences and beliefs, to reach a verdict
individual jurors tended to convict more often in response to the (MacCoun and Kerr, 1988). Alternatively, the findings may

be attributable to the leniency effect previously observed in simulation materials to include a closer fit between the expert
deliberation (MacCoun and Kerr, 1988; Devine et al., 2009). evidence and the case facts.
However, testing these theories requires analysis of the content One caveat in interpreting the significant interaction effect
of each of the deliberations, a task beyond the scope of the between the expert witness condition and the deliberation
present study. condition is that this finding could have been confounded by a
Notably, the multilevel modeling presented more evidence difference in the order of administration of post-trial verdict and
on general psychological factors and paths in understanding CSA knowledge measures to deliberating and non-deliberating
the variance in jurors’ CSA knowledge and final verdict when jurors. That is, post-trial CSA knowledge was measured after
the intercorrelation of juror data was taken into account at the deliberating jury group verdict and before the individual
both juror and jury levels. Expert interventions and judicial verdict of those jurors, whereas post-trial CSA knowledge of non-
direction statistically significantly increased jurors’ post-trial deliberating jurors was assessed before their individual verdicts
CSA knowledge while psychological mechanisms of perceived were provided.
credibility shed light on the social and cognitive factors Had the juries been allowed more time to deliberate to
that triggered guilty verdicts, thereby extending the previous a final verdict, there may have been fewer hung juries.
literature (Goodman-Delahunty et al., 2010; Powell et al., 2016). Reviewers of prior research applying dual-process heuristic-
systematic or elaboration likelihood models of persuasion to
jury deliberations observed that the more rapid, heuristic or
Limitations of the Study peripheral decision strategy may be triggered by time-pressure.
The observation or videotaping of jury deliberations in real trials Future research analyses of deliberation content applying these
is prohibited. If the effects of deliberation are to be tested, the best persuasion models should explore whether different decision
option for researchers in Australia is to recruit jurors called for strategies were applied by juries who felt pressured to conclude
jury duty to serve on a simulated case so that their deliberations their deliberations, i.e., whether they avoided slower, more
can be videotaped, as was done in this study. Although we detailed reasoning about the information presented in the
increased the external validity of our methodology by recruiting educative interventions.
actual jurors who reported for jury duty, by conducting our Similarly, future research applying dual-process heuristic-
experiments within a court precinct, and by using a professionally systematic or elaboration likelihood models of persuasion to jury
acted videotrial, it can nonetheless be argued that our findings deliberations may benefit by distinguishing juror beliefs from
may not generalize to real juries because jurors knew they juror knowledge to test the impact of contextual effects, such
were participating in a simulated trial (Goodman-Delahunty as group deliberations, on beliefs vs. knowledge. Psychologists
et al., 2011b). Nonetheless, jury deliberations revealed the have observed that global or abstract beliefs and attitudes often
conscientiousness with which jurors engaged in their task, with differ from actions in response to a specific set of case facts
disagreements and anxiety expressed about the consequences of or a particular context (Ajzen and Fishbein, 2005). A lack
a guilty verdict. of information is one common cause of the observed value-
The case facts in the simulated trial were representative of action gap between what people say and what they do. The fact
some counter-intuitive aspects of a typical CSA trial, namely that different results were obtained using only the individual
(a) the perpetrator was someone familiar to the complainant, post-trial questionnaire vs. group discussions in a naturalistic
a family member, rather than a stranger; (b) the complainant setting and a group decision followed by the individual post-trial
continued to have ongoing contact with the perpetrator after the questionnaire is unsurprising (Grzyb, 2016). Jury deliberation
alleged abuse; (c) the abuse took place in familiar setting, the itself may have a number of possible effects. For instance, it may
complainant’s home; and (d) the abuse was not violent. However, (a) reinforce in some jurors their belief (or uncertainty) that a
the case facts also included a number of unrepresentative features judge’s or expert’s statement about children was false, enabling
of CSA cases, i.e., the complaint entailed (a) a single abusive those jurors to dismiss that information so it does not become
event, a one-off instance rather than a series of recurring part of their knowledge base. Alternatively, (b) a juror who may
abusive events; (b) immediate rather than delayed disclosure; have formed a belief that a certain educative statement was true
(c) the child victim was 12 years of age thus less suggestible but who had not yet engaged in more effortful processing to
than many younger children; (d) the victim resisted; and (e) a add it to their knowledge base, may be more readily persuaded
corroborating witness observed the complainant and the accused by alternative arguments in deliberation that counteract that
with their pants down and overheard the victim tell the accused new belief. Or (c) other jurors who accepted certain educative
to stop. Inclusion of the latter series of facts strengthened the statements (formed a belief) and then added that information
evidence against the accused, supporting potential convictions, to their knowledge base may be compelled to dismiss that newly
but decreased the goodness of fit between the case facts and acquired knowledge if the majority view of the jury was to dismiss
typical counter-intuitive specialized CSA knowledge that is often it in order to reach a consensus verdict, whether to convict
the basis for expert evidence (Seymour et al., 2013). Prior research or acquit.
has shown that when this fit is poor, jurors may disregard
the educative evidence or infer that the mismatch and atypical CONCLUSION
features indicate that the complainant was not sexually assaulted
(Horan and Goodman-Delahunty, 2020). Future research testing Our study confirmed that in CSA trials, similar to results in
the effectiveness of educative interventions should tailor trial studies of jury behavior in other selected types of cases, such

as capital crimes (where a conviction-prone bias accompanies DATA AVAILABILITY STATEMENT

support for the death penalty among death-qualifies jurors;
see Bornstein and Greene, 2017), a statistically significant The raw data supporting the conclusions of this article will be
relationship exists between the pre-existing attitudes and made available by the authors, without undue reservation.
demographic characteristics of citizens called for jury duty, their
perceptions of the evidence, assessments of witness credibility, ETHICS STATEMENT
and verdicts. As anticipated, educational interventions in the
form of a judicial direction and expert evidence from a The studies involving human participants were reviewed and
psychologist statistically significantly increased jurors’ CSA approved by Charles Sturt University Human Research Ethics
knowledge, which enhanced the credibility of the complainant Committee. The patients/participants provided their written
and increased the conviction rate. informed consent to participate in this study.
Our findings suggested that specialized information
was best conveyed by a judicial direction than by AUTHOR CONTRIBUTIONS
an expert witness. However, a diagnostic expert who
interviewed the complainant had a greater impact on juror All authors listed have made a substantial, direct and intellectual
perceptions of the complainant’s credibility and verdict contribution to the work, and approved it for publication.
than a social framework expert, possibly because the
diagnostic expert appeared more competent and reliable FUNDING
to express an opinion about the whether the child was
sexually assaulted. Funded by a grant to JG-D and AC (Principal Investigators)
Importantly, the findings demonstrated systematic (2011–2013). Countering misconceptions in child sexual assault
differences between individual juror decisions and jury cases with expert evidence and judicial directions (Discovery
decisions following group deliberations. The fact that Project DP110103706) [Grant]. Australian Research Council
these analyses yielded different outcomes in terms of Discovery Program. https://www.arc.gov.au/grants/discovery-
CSA knowledge increases, the perceived credibility of the program/discovery-projects.
complainant, and the effects of deliberation on verdict
underscores the critical importance of including group SUPPLEMENTARY MATERIAL
deliberation in simulated jury studies, and of using more
sophisticated methods of analysis in jury research that take The Supplementary Material for this article can be found
the non-independence of the nested jurors within a jury online at: https://www.frontiersin.org/articles/10.3389/fpsyg.
into account. 2021.624331/full#supplementary-material
REFERENCES Cashmore, J., Taylor, A., and Parkinson, P. (2019). Fourteen-year trends in the
criminal justice response to child sexual abuse reports in New South Wales.
Ajzen, I., and Fishbein, M. (2005). “The influence of attitudes on behavior,” in The Child Maltreat. 25, 85–95. doi: 10.1177/1077559519853042
Handbook of Attitudes, eds D. Albarracín, B. Johnson, and M. Zanna (Mahwah, Ceci, S. J., Kahan, D. M., and Braman, D. (2010). The WEIRD are even weirder
NJ: Erlbaum), 173–221. than you think: diversifying contexts is as important as diversifying samples.
Anderson, C. A., Lepper, M. R., and Ross, L. (1980). Perseverance of Behav. Brain Sci. 33, 87–88. doi: 10.1017/S0140525X10000063
social theories: the role of explanation in the persistence of discredited Cialdini, R. B., and Goldstein, N. J. (2004). Social influence:
information. J. Pers. Soc. Psychol. 39, 1037–1049. doi: 10.1037/h00 compliance and conformity. Annu. Rev. Psychol. 55, 591–621.
77720 doi: 10.1146/annurev.psych.55.090902.142015
Bornstein, B. H., and Greene, E. (2017). The Jury Under Fire: Myth, Controversy, Cossins, A. (2008). Children, sexual abuse and suggestibility: what laypeople think
and Reform. New York, NY: Oxford University Press. they know and what the literature tells us. Psychiatry Psychol. Law 15, 153–170.
Bornstein, B. H., and McCabe, S. G. (2005). Jurors of the absurd. the role of doi: 10.1080/13218710801886040
consequentiality in jury simulation research. Florida State Univ. Law Rev. Cossins, A. (2020). Closing the Justice Gap for Adult and Child Sexual Assault:
32, 443–467. Available online at: https://ir.law.fsu.edu/lr/vol32/iss2/7 (accessed Rethinking the Adversarial Trial. London: Palgrave Macmillian.
July 27, 2021). Cossins, A., and Goodman-Delahunty, J. (2013). Misconceptions or expert
Brekke, N., and Borgida, E. (1998). Expert psychological testimony in evidence in child sexual assault trials: enhancing justice and jurors’ “common
rape trials: a social-cognitive analysis. J. Pers. Soc. Psychol. 55, 372–386. sense”. J. Judicial Admin. 22, 171–190.
doi: 10.1037/0022-3514.55.3.372 Cossins, A., Goodman-Delahunty, J., and O’Brien, K. (2009). Uncertainty and
Brodsky, S. L., Griffin, M. P., and Cramer, R. J. (2010). The Witness Credibility misconceptions about child sexual abuse: implications for the criminal justice
Scale: an outcome measure for expert witness research. Behav. Sci. Law 28, system. Psychiatry Psychol. Law 16, 435–452. doi: 10.1080/132187109029
892–907. doi: 10.1002/bsl.917 30234
Browne, R. (2017). More child sexual assault cases in court but fewer convictions: Criminal Code Act, 1983 (Northern Territory of Australia), s368.
Justice Peter McClellan, Sydney Morning Herald, 12 April. Available online Cronbach, L. J. (1951). Coefficient alpha and the internal structure of a test.
at: https://www.smh.com.au/national/nsw/more-child-sexual-assault-cases- Psychometrika 16, 297–334. doi: 10.1007/BF02310555
in-court-but-fewer-convictions-justice-peter-mcclellan-20170412-gvj9vb. Devine, D. J. (2012). Jury Decision Making: The State of the Science. New York, NY:
html (accessed July 27, 2021). New York University Press.

Devine, D. J., Buddenbaum, J., Houp, S., Studebaker, N., and Stolle, D. decision making? a social cognitive analysis. J. Appl. Psychol. 82, 178–191.
P. (2009). Strength of evidence, extraevidentiary influence, and the doi: 10.1037/0021-9010.82.1.178
liberation hypothesis: data from the field. Law Hum. Behav. 33, 136–148. Kovera, M. B., Levy, R. J., Borgida, E., and Penrod, S. D. (1994). Expert testimony in
doi: 10.1007/s10979-008-9144-x child sexual abuse cases: effects if expert evidence type and cross-examination.
Devine, D. J., and Caughlin, D. E. (2014). Do they matter? a meta-analytic Law Hum. Behav. 18, 653–674. doi: 10.1007/BF01499330
investigation of individual characteristics and guilt judgments. Psychol. Pub. Krauss, D. A., McGabe, J. G., and Lieberman, J. D. (2012). Dangerously
Policy Law 20, 109–134. doi: 10.1037/law0000006 misunderstood: representative jurors’ reactions to expert testimony on future
Eaton, T. E., Ball, P. J., and O’Callaghan, M. G. (2001). Child witness and defendant dangerousness in a sexual violent predator trial. Psychol. Pub. Policy Law 18,
credibility: child evidence presentation mode and judicial instructions. J. Appl. 18–49. doi: 10.1037/a0024550
Psychol. 31, 1845–1858. doi: 10.1111/j.1559-1816.2001.tb00207.x Krauss, D. A., and Sales, B. D. (2001). The effects of clinical and scientific expert
Faigman, D., Monahan, J., and Slobogin, C. (2014). Group to individual (G2i) testimony on jurors decision making in capital sentencing. Psychol. Pub. Policy
inference in scientific expert testimony. Univ. Chicago Law Rev. 81, 417–480. Law 7, 267–310. doi: 10.1037/1076-8971.7.2.267
Field, H. S., and Barnett, N. J. (1978). Simulated jury trials: students Landström, S., Granhag, P. A., and Hartwig, M. (2010). Children’s live
vs. “real” people vs jurors. J. Social Psychol. 104, 287–293. and videotaped testimonies: how presentation mode affects observers’
doi: 10.1080/00224545.1978.9924071 perception, assessment and memory. Legal Crim. Psychol. 12, 333–348.
Fitzgerald, J. (2006). The Attrition of Sexual Offences From the New South Wales doi: 10.1348/135532506X133607
Criminal Justice System. Crim. Justice Bulletin No. 92. Sydney, NSW: NSW Levett, L. M., and Devine, D. (2017). “Integrating individual and group models
Bureau of Crime Statistics and Research. of juror decision making” in The Psychology of Juries, ed M. B. Kovera
Gabora, N. T., Spanos, N. P., and Joab, A. (1993). The effects of complainant age (Washington, DC: American Psychological Association), 37–57.
and expert psychological testimony in a simulated child sexual abuse trial. Law Lieberman, J. D., Krauss, D. A., and Wiener, R. (2011). When does sample
Hum. Behav. 17, 103–119. doi: 10.1007/BF01044540 matter in juror decision-making research? differences between college
Goodman-Delahunty, J., Cossins, A., and O’Brien, K. (2010). Enhancing student and representative samples of juror. Behav. Sci. Law 29, 325–327.
the credibility of complainant in child sexual assault trials: the effect doi: 10.1002/bsl.987
of expert evidence and judicial directions. Behav. Sci. Law 28, 769–783. Lord, F., and Novick, M. (1968). Factor Analysis and Related Methods. Hillsdale,
doi: 10.1002/bsl.936 NJ: Lawrence Erlbaum.
Goodman-Delahunty, J., Cossins, A., and O’Brien, K. (2011a). A comparison Lovis-McMahon, D. (2015). Multilevel potential outcome models for causal
of expert evidence and judicial directions to counter misconceptions in inference in jury research (Doctoral dissertation). Tempe, AZ: Arizona
child sexual abuse trials. ANZJ Crim. 44, 196–217. doi: 10.1177/00048658114 State University.
05140 Lyon, T. D., and Koehler, J. J. (1996). The relevance ratio: evaluating the probative
Goodman-Delahunty, J., and Martschuk, N. (2020). Mock jury and juror responses value of expert testimony in child sexual abuse cases. Cornell L. Rev. 82, 43–78.
to uncharged acts of sexual misconduct: advances in the assessment of unfair MacCoun, R. J. (2005). Comparing legal fact finders: real and mock, amateur and
prejudice. J. Psychol. 228, 199–209. doi: 10.1027/2151-2604/a000410 professionals. Florida State Univ. Law Rev. 32, 511–506.
Goodman-Delahunty, J., Martschuk, N., and Cossins, A. (2017a). Validation of MacCoun, R. J., and Kerr, N. L. (1988). Asymmetric influence in mock jury
child sexual abuse knowledge questionnaire. Psychol. Crim. Law 23, 391–412. deliberation: jurors’ bias for leniency. J. Pers. Soc. Psychol. 54, 21–33.
doi: 10.1080/1068316X.2016.1258469 doi: 10.1037/0022-3514.54.1.21
Goodman-Delahunty, J., Martschuk, N., and Cossins, A. (2017b). What Australian McCabe, J. G., Krauss, D. A., and Lieberman, J. D. (2010). Reality check: a
jurors know and do not know about evidence of child sexual abuse. Crim. Law comparison of college students and community samples of mock jurors in
J. 41, 86–103. a simulated sexual violent predator civil commitment. Behav. Sci. Law 28,
Goodman-Delahunty, J., Rossner, M., and Tait, D. (2011b). “Simulation and 730–750. doi: 10.1002/bsl.902
dissimulation in jury research: credibility in a live mock trial,” in Qualitative Monahan, J., Walker, L., and Mitchell, G. (2009). The limits of social framework
Criminology: Stories From the Field, eds L. Bartels and K. Richardson evidence. Law Probab. Risk 8, 307–321. doi: 10.1093/lpr/mgp020
(Leichhard, NSW: Hawkins Press), 34–44. Moran, G., and Comfort, J. C. (1982). Neither “tentative” nor “fragmentary”:
Gordon, S. G. (2014). All together now: using principles of group dynamics to train verdict preferences of impaneled felony jurors as a function of
better jurors. Indiana Law Rev. 48:415. doi: 10.2139/ssrn.2479572 attitude towards capital punishment. J. Appl. Psychol. 71, 146–155.
Grzyb, T. (2016). “Why can’t we just ask?” the influence of research methods on doi: 10.1037/0021-9010.71.1.146
results. The case of the “bystander effect”. Polish Psychol. Bull 47, 233–235. Myers, D. G., and Lamm, H. (1976). The group polarization phenomenon. Psychol.
doi: 10.1515/ppb-2016-0027 Bull. 83, 602–627. doi: 10.1037/0033-2909.83.4.602
Henrich, J., Heine, S., and Norenzayan, A. (2010). The weirdest people in the Nickerson, R. S. (1998). Confirmation bias: a ubiquitous phenomenon in
world? Behav. Brain Sci. 33, 61–83. doi: 10.1017/S0140525X0999152X many guises. Rev. Gen. Psychol. 2, 175–220. doi: 10.1037/1089-2680.
Horan, J., and Goodman-Delahunty, J. (2010). Challenging the peremptory 2.2.175
challenge system in Australia. Crim. Law J. 34, 167–186. Nietzel, M. T., McCarthy, D. M., and Kern, M. J. (1999). “Juries: the current state
Horan, J., and Goodman-Delahunty, J. (2020). Expert evidence to counteract of the empirical literature,” in Perspectives in Law and Psychology, Vol. 10.
jury misconceptions about consent in sexual assault cases: failures and lessons Psychology and Law: The State of the Discipline, eds R. Roesch, S. D. Hart, and
learned. UNSW Law J. 43, 707–737. J. R. P. Ogloff (Amsterdam: Kluwer Academic Publishers), 23–52.
Huang, K.-C., and Lin, C.-C. (2014). Mock jury trials in Taiwan: paving the NSW Bureau of Crime Statistics and Research (2013). New South Wales Criminal
ground for introducing lay participation. Law Hum. Behav. 38, 367–377. Court Statistics 2012 89 (Table 3.4). Sydney, NSW NSW: Bureau of Crime
doi: 10.1037/lhb0000080 Statistics and Research.
Juries Act, 1927 (South Australia of Australia), s57. NSW Bureau of Crime Statistics and Research (2020). Recorded crime reports.
Juries Act, 1957 (Western Australia of Australia), s41. Available online at: https://www.bocsar.nsw.gov.au/Pages/bocsar_crime_stats/
Juries Act, 2000 (Victoria of Australia), s46. bocsar_latest_quarterly_and_annual_reports.aspx (accessed July 27, 2021).
Jury Act, 1899 (Tasmania of Australia), s48(2). NSW Bureau of Crime Statistics and Research (2021). Sexual Assault Offences:
Jury Act, 1977 (New South Wales of Australia), s55F. Progress Through the Criminal Justice System. Available online at: https://www.
Klettke, B., Graesser, A. G., and Powell, M. B. (2009). Expert testimony in child bocsar.nsw.gov.au/Pages/bocsar_pages/Sexual-assault.aspx (accessed July 27,
sexual abuse cases: the effects of evidence, coherence and credentials on juror 2021).
decision making. Appl. Cogn. Psychol. 24, 481–494. doi: 10.1002/acp.1565 O’Brien, K., Goodman-Delahunty, J., Clough, J., and Pratley, J. (2008). Juror
Kovera, M. B., Borgida, E., Gresham, A. W., Gray, E., and Regan, P. C. satisfaction and confidence in New South Wales, Victoria and South Australia.
(1997). Does expert psychological testimony inform or influence juror Trends Issues Crime Crim. Just. 354, 1–6.

Park, K. K., Kim, S., Lee, E., and Seo, H. (2005). Social conformity and cognitive Tabak, S. J., and Klettke, B. (2014). Mock jury attitudes towards credibility, age, and
conversation during jury deliberation: a content analysis of deliberation guilt in a fictional child sexual assault scenario. Australian J. Psychol. 66, 47–55.
arguments in the first officially simulated jury trial in Korea. Korean J. Soc. Pers. doi: 10.1111/ajpy.12035
Psychol. 19, 1–6. Tabak, S. J., Klettke, B., and Knight, T. (2013). Simulated jury decision making
Penrod, S. D., Kovera, M. B., and Groscup, J. (2011). “Jury research methods,” in in online focus groups. Qual. Res. J. 3, 102–113. doi: 10.1108/144398813113
Research Methods in Forensic Psychology, eds. B. Rosenbaum and S. D. Penrod 14667
(Hoboken, NJ: Wiley), 191–214. Vidmar, N. (2008). “Civil juries in ecological context: methodological implications
Peter-Hagene, L. C., Salerno, J. M., and Phalen, H. (2019). “Jury decision making” for research,” in Civil Juries and Civil Justice: Psychological and Legal
in Psychological Science and the Law, eds N. Brewer and A. Bradfield Douglass Perspectives, eds B. H. Bornstein, R. L. Wiener, R. Schopp, and S. L. Willborn
(New York, NY: Guildford) 338–366. (Berlin: Springer Science + Business Media), 35–65.
Powell, M., Westera, N., Goodman-Delahunty, J., and Pichler, A. (2016). An Wiener, R. L., Krauss, D. A., and Lieberman, J. D. (2011). Mock jury research:
Evaluation of How Evidence is Elicited from Complainants of Child Sexual Abuse. where do we go from here? Behav. Sci. Law 29, 467–479. doi: 10.1002/bsl.989
Sydney: Royal Commission into Institutional Responses to Child Sexual Abuse, Wolf, E. J., Harrington, K. M., Clark, S. L., and Miller, M. W. (2013).
469. Sample size requirements for structural equation models: an evaluation
Quas, J. A., Thompson, W. C., Alison, K., and Stewards, C. (2005). Do jurors of power, bias, and solution propriety. Educ. Psychol. Meas. 76, 913–934.
‘know’ what isn’t so about child witnesses? Law Hum. Behav. 29, 425–456. doi: 10.1177/0013164413495237
doi: 10.1007/s10979-005-5523-8 Wright, D. B., and Hall, M. (2007). How a “reasonable doubt” instruction
Raykov, T. (1997). Scale reliability, Cronbach’s alpha, and violations of essential affects decisions of guilt. Basic Appl. Soc. Psychol. 29, 91–98.
tau-equivalence with fixed congeneric components. Multivar. Behav. Research doi: 10.1080/01973530701331254
32, 329–353. doi: 10.1207/s15327906mbr3204_2 Wundersitz, J. (2003). Tracking From Police Incident Report to Finalisation in
Raykov, T. (2002). Analytic estimation of standard error and confidence Court. Canberra ACT: Office of Crime Statistics and Research, SA, Child
interval for scale reliability. Multivar. Behav. Research 37, 89–103. Sexual Assault.
doi: 10.1207/S15327906MBR3701_04
Salerno, J. M., Bottoms, B. L., and Peter-Hagene, L. C. (2017). Individual Conflict of Interest: The authors declare that the research was conducted in the
versus group decision making: jurors’ reliance on central and peripheral absence of any commercial or financial relationships that could be construed as a
information to evaluate expert testimony. PLoS ONE 12:e018358. potential conflict of interest.
doi: 10.1371/journal.pone.0183580
Schulz, P. W., Nolan, J. M., Cialdini, R. B., Goldstein, N. J., and Griskevicius, C. Publisher’s Note: All claims expressed in this article are solely those of the authors
(2007). The constructive, destructive, and reconstructive power of social norms. and do not necessarily represent those of their affiliated organizations, or those of
Psychol. Sci. 18, 249–254. doi: 10.1111/j.1467-9280.2007.01917.x
the publisher, the editors and the reviewers. Any product that may be evaluated in
Schutte, J. W., and Hosch, H. M. (1997). Gender differences in sexual assault
this article, or claim that may be made by its manufacturer, is not guaranteed or
verdicts: a meta-analysis. J. Soc. Behav. Pers. 12, 759–772.
Seymour, F., Blackwell, S., Calvert, S., and McLean, B. (2013). Counterintuitive endorsed by the publisher.
expert psychological evidence in child sexual abuse trials in New
Zealand. Psychiatr. Psychol. Law 21, 511–522. doi: 10.1080/13218719.2013. Copyright © 2021 Goodman-Delahunty, Martschuk, Lee and Cossins. This is an
839930 open-access article distributed under the terms of the Creative Commons Attribution
Shead, K. (2014). Responding to historical child sexual abuse: a prosecution License (CC BY). The use, distribution or reproduction in other forums is permitted,
perspective on current challenges and future directions. Curr. Issues Crim. Just. provided the original author(s) and the copyright owner(s) are credited and that the
26, 55–73. doi: 10.1080/10345329.2014.12036007 original publication in this journal is cited, in accordance with accepted academic
Tabachnick, B. G., and Fidell, L. S. (2013). Using Multivariate Statistics, 6th edn. practice. No use, distribution or reproduction is permitted which does not comply
Boston: Pearson. with these terms.

BRIEF RESEARCH REPORT
published: 24 August 2021
doi: 10.3389/fpsyg.2021.600370
Non-intimate Relationships and

Psychopathic Interpersonal and
Affective Deficits as Risk Factors for
Criminal Career: A Comparison
Between Sex Offenders and Other
Offenders
Fabio Ferretti 1 , Andrea Pozza 1 , Fulvio Carabellese 1* , Adriano Schimmenti 2 ,
Gianluca Santoro 2 , Gabriele Mandarelli 3 , Giacomo Gualtieri 1 , Felice Carabellese 3 ,
Roberto Catanesi 3 and Anna Coluccia 1
1
Department of Medical Sciences, Surgery and Neurosciences, University of Siena, Siena, Italy, 2 Kore University of Enna,
Enna, Italy, 3 Forensic Psychiatry, University of Bari, Bari, Italy
Edited by:
Noora Ellonen, Sex-offenders are at risk of criminal recidivism. For the treatment to be truly effective,
University of Tampere, Finland it must be individualized. For this purpose, an accurate assessment should focus
Reviewed by: on criminological, psychological, and psychopathological features. The present study
Matt DeLisi,
Iowa State University, United States compared sex offenders with other offenders on historical experiences (i.e., problems
Richard S. John, with violence, anti-social behaviors, problems with personal relationships, problems
University of Southern California,
with substance use, traumatic experiences, and parenting style). In addition, given the
United States
Laura Jayne Broome, association between life events and psychopathy, we explored whether the relation
Swansea University, United Kingdom between life events and crime type (sexual crime vs. other types of crime) might be
*Correspondence: moderated by psychopathy traits (interpersonal and affective deficits and antisocial
Fulvio Carabellese
carabellese.fulvio@gmail.com behavior). Eighty-eight sex offenders (76% of whom child molesters) and 102 other
offenders were included. The Historical, Clinical and Risk Management - 20 item Version
Specialty section: 3 (HCR-20V3) and Psychopathy Checklist-Revised (PCL-R) were administered. The
Forensic and Legal Psychology, scores of the HCR-20V3 Historical scale items were computed to assess life events.
a section of the journal The scores of the PCL-R factors, F1 Interpersonal affective deficits and F2 Antisocial
behavior, were recorded. The presence of a history of problems with non-intimate
Received: 29 August 2020
Accepted: 03 August 2021
relationships was the only significant risk factor for sexual crime compared with other
Published: 24 August 2021 crimes. Interpersonal and affective deficits provided an increased likelihood of being
Citation: sex offenders as compared with other offenders when problems with non-intimate
Ferretti F, Pozza A, Carabellese F, relationships were possibly/partially or certainly present.
Schimmenti A, Santoro G,
Mandarelli G, Gualtieri G, Keywords: sex offenders, risk factors, psychopathy, personality traits, life events
Carabellese F, Catanesi R and
Coluccia A (2021) Non-intimate
Relationships and Psychopathic
Interpersonal and Affective Deficits as
INTRODUCTION
Risk Factors for Criminal Career:
A Comparison Between Sex
The history of criminal perpetrators, the developmental pattern of their careers, and the risk factors
Offenders and Other Offenders. for criminal behaviors have always been considered a central topic in criminological research.
Front. Psychol. 12:600370. The relationship between the traumatic experiences of the sexually abused victims and the
doi: 10.3389/fpsyg.2021.600370 arising of dysfunctional sexual behaviors was investigated by several scholars, with the conclusion

Ferretti et al. Criminal Career of Sex Offenders
that additional causes underlying sexual offenses must be of the HCR-20V3 provide a series of information on the
searched and that the so-called Victim/Perpetrator Paradigm is subject’s existential path. The items recall some factors among
too reductive (Burton, 2008; Rasmussen, 2012). those mentioned (items H1, H2, H3, H5, H8) that could
Psychiatric disorders play an important role as well. have a significant role in the criminal career. Moreover, the
Schizophrenia and bipolar disorder (Carabellese et al., 2012; historical items of HCR-20V3 take into consideration other
Pozza et al., 2019, 2020) and mood and personality disorders factors, in particular personality disorders (H7) and major mental
(Dunsieth et al., 2004; Coluccia et al., 2020) were detected in sex disorders (H6). In addition to being correlated with violent
offenders. In their research on a sample of 1346 sex offenders, behavior (Carabellese et al., 2020a) and sexual offenses (Somma
Eher et al. (2019) found that 50.1% of them suffered from et al., 2020), such disorders can interact with the psychopathic
personality disorders. Carabellese et al. (2012) confirmed these dimension (Wong and Olver, 2016), another personality factor
results, highlighting the role of schizoid, narcissistic, and avoidant notoriously correlated with the risk of criminal recidivism
personality disorders. (Hanson and Morton-Bourgon, 2005). Cartwright et al. (2018)
The relevance of psychopathy on criminal behavior and suggested that the HCR-20V3 plays an important role in
violent conduct is widely recognized in the literature, and it is the assessment and management of sex offenders beyond the
generally assessed by the gold standard measure Psychopathy assessment of the risk of relapse. The HCR-20V3 has proved to
Checklist-Revised (PCL-R; Hare and Neumann, 2006) which be a valid predictor of violent conduct in perpetrators of crimes,
covers two main factors, i.e., interpersonal and affective deficits and its use as a predictive tool has also been suggested in sex
and antisocial behavior (Carabellese et al., 2008; Leistico et al., offenders (Cartwright et al., 2018). The factors that the HCR-
2008; Campbell et al., 2009; Campobasso et al., 2009; Kennealy 20V3 takes into consideration, especially in the historical items,
et al., 2010; Yang et al., 2010; Ismail and Looman, 2018; Mazzoni represent a valid support for a professional judgment aimed to
et al., 2018). The association between psychopathy and sexual develop a personalized therapeutic project (Yates, 2013; Gannon
crimes emerged in several studies, including a meta-analysis et al., 2019).
conducted by Hanson et al. (2009) which found that the presence The aim of this study was to explore the relevance of life
of psychopathy in sex offenders constitutes a predictive factor events in criminal outcome, comparing sex offenders with other
for criminal recidivism in sexual offenses and other types of offenders about historical experiences assessed by the HCR-
offenses. High scores on the PCL-R were found to be associated 20V3 related to problems with violence, anti-social behaviors,
with experiences of neglect and abuse (Kimonis et al., 2011; problems with personal relationships, problems with substance
Schraft et al., 2013; Ometto et al., 2016; Sevecke et al., 2016). use, traumatic experiences, and parenting style. This is the first
Other studies (e.g., Schimmenti et al., 2020) have shown that the study which explored risk factors in sex offenders by using
relational failure of caregivers in the attachment system, due to HCR-20V3 in its validated Italian version (Caretti et al., 2019)
experiences of neglect, early abandonment or loss of caregivers, and one of the few studies which investigated the discriminant
can lead the child to difficulty in interpersonal self-regulation role of these life events comparing sex offenders (SO) with
skills. These interpersonal difficulties can in turn represent a risk other offenders (OO). In addition, given the association between
factor in adults for aggressive behavior, difficulties in emotion life events and psychopathy, we explored whether the relation
regulation and sexual problems, as it can be found in many cases between certain life events measured by the HCR-20 V3 and the
of psychopathy, sadism, or paraphilia. type of crime (sexual crime vs. other types of crime) might be
Such a kaleidoscope of different situations makes it difficult moderated by psychopathy traits, i.e., interpersonal and affective
to identify those risk factors that can affect the dysfunctional deficits and antisocial behavior.
sexual behavior of sex offenders. Due to the variety of factors that
must be considered when studying sexual crimes, investigating
sex offenders’ criminal career through different risk factors is a MATERIALS AND METHODS
topic that still deserves the attention of researchers. Although the
evidence about these relationships is still controversial, at least Participants and Setting
partly, sex offenders seem to show a higher prevalence of physical, This research is part of a national multicenter project authorized
sexual, and emotional abuse during their childhood (Lee et al., by the Penitentiary Administration Department of the Ministry
2002; Jespersen et al., 2009; Seto and Lalumière, 2010; Drury of Justice. The study was reviewed and approved by the Director
et al., 2019) and a history of family dysfunctions (Lee et al., 2002), and ethical board of Penitentiary Administration Department
poor parenting style (Sigre-Leirós et al., 2016) and symptoms of of Ministry of Justice in accordance with the current ethical
social anxiety (Porter et al., 2015). In their review, Kraanen and standards and in compliance with the rules concerning the
Emmelkamp (2011) reported that sex offenders’ criminal career privacy of data related to the perpetrators of sexual crimes.
is associated with substance abuse, alcohol, and drug misuse. This multicenter study was realized with the collaborations of
Among the psychometric instruments for professional risk the penitentiary institutions of six Italian regions (Lombardia,
assessment, the Historical, Clinical and Risk Management - 20 Veneto, Toscana, Lazio, Puglia e Sicilia). The regions involved
item Version 3 (HCR-20 Version 3; Douglas et al., 2013) were chosen for the presence of university centers that planned
represents a psychometrically sound instrument, designed to the enrollment of the sample, the collection and analysis of the
give a comprehensive framework for the entire process of data. The researchers had previously been trained in the use
risk assessment (Douglas et al., 2013). The historical items of the tools. All the prisoners enrolled in the study received

a final conviction for their crimes. Data were collected during Measures
2015 and 2016 and only the prisoners who had previously The Psychopathy Checklist-Revised (PCL-R; Hare and
given their consent to meet the researchers were included in Neumann, 2006) in its validated Italian version (Caretti et al.,
the study. For legal and privacy-related reasons, data will not 2011) was administered to assess the presence of psychopathy.
be made available. PCL-R factor 1 captures traits dealing with the interpersonal and
The exclusion criteria were applied to both the SO and the OO affective deficits of psychopathy (e.g., shallow affect, superficial
groups and consisted of the lack of provision of the informed charm, manipulativeness, and lack of empathy), whereas
consent and the presence of a lifetime psychiatric diagnosis factor 2 deals with symptoms relating to antisocial behavior
which was considered by the Penitentiary Administration (e.g., criminal versatility, impulsiveness, irresponsibility, poor
an exclusion criterion to prevent the fact that the relation behavior controls, and juvenile delinquency) (Hare et al., 1989).
between life events and sexual crime might be attributed to The scores of the two different components of factors F1 and F2
a psychiatric history as psychiatric disorders are observed in were recorded, as well as the values of the different components
sex offenders (Dunsieth et al., 2004; Carabellese et al., 2012; within factors F1 and F2. A threshold score of PCL-R equal to
Eher et al., 2019). The presence of a lifetime psychiatric or greater than 25 was established to identify the condition of
diagnosis was established by the penitentiary administration if the psychopathy, as indicated in studies conducted on European
individual had received or was receiving a psychiatric treatment, populations (Grann et al., 1998; Andersen et al., 1999; Jüriloo
i.e., psychiatric medications or psychotherapeutic treatments et al., 2013).
prescribed or delivered by a mental health professional in any The HCR-20V3 Italian version (Douglas et al., 2013; Caretti
period during his life. et al., 2019), a tool assessing the risk of violence, was
In Italy, the number of sexual offenses leading to the administered. Some items of the Historical Scale were considered
internment of the offenders in the so-called High-Security to analyze the existence of lifetime problematic experiences that
Forensic Psychiatry Residences (i.e., the psychiatric facilities can be considered risk factors for the development of sexually
hosting socially dangerous offenders with a mental disorder) based crimes. According to the literature about risk factors for
is relatively low (3.4%) (Catanesi et al., 2019). As for the sex sex offending, the following items were selected among those
offenders with mental disorders not serious enough to exclude belonging to the Historical Scale: (a) H1 – problems with
criminal responsibility, in Italy they are sentenced in prison, violence; (b) H2 – problems with other antisocial behavior; (c)
but in this case, we had no data concerning them because the H3 – problems with relationships; (d) H5 – problems with
prisoners with psychiatric disorders were excluded from the study substance use; (e) H8 – problems with traumatic experiences.
by the Penitentiary Administration. Items H1 and H2 are divided in sub-items expressing three
Eighty-eight SO and 102 OO were assessed by researchers age classes: (1) child, 12 years and younger; (2) adolescent, 13–
trained in the administration of the tools, to guarantee 17 years; (3) adult, 18 years and older. The item H3 is divided in
uniformity and homogeneity in data collection. The category two sub-items (intimate and non-intimate relationships), and the
SO was declined according to the definition given by Myers item H8 is also divided in two sub-items (victimization/trauma
et al. (2005), even though the majority (76%) of these and adverse child rearing experiences). Every item is scored 0,
inmates were child molesters. The final sample of 88 sex 1, or 2: 0 if the item is definitely absent, 1 if the item possibly
offenders did not commit any other crimes than sexual is present, or present to a minor/moderate degree, and 2 if the
offenses, possibly in relapse. The data concerning the fact item is definitely present. All the researchers involved had been
that the SOs had been convicted for the first time or they trained through role-playing and internships to administer the
were recidivists, though, was kept confidential by the prison tools before the research was initiated.
management. It must also be said that SOs in Italy are Together with the assessment scales, the anamnestic and
either detained in separate prisons (as happens for example criminological data were collected through the examination of
in Puglia), or they are placed in special sections within the the personal data sheets.
individual prisons.
The category OO included a heterogeneous set of crimes, such
as personal crimes (homicide, assault), property crimes (robbery, Statistical Analysis
fraud), and crimes against the State, but not sex offenses. Descriptive statistics were used to describe subjects’
A preliminary meeting with prisoners and social workers was characteristics (mean and standard deviation for
scheduled before the assessment, to present the research project continuous measures, frequencies, and relative frequencies
and collect participants’ written consent (Mandarelli et al., 2017). for categorical variables). In accordance with the
At this stage, about 18% of prisoners refused to give informed manual of the instrument (Caretti et al., 2019), the
consent and participate. However, the researchers did not know presence of the risk factors was measured by the HCR-
the number of those who had initially refused to participate to the 20V3 Historical item scores which were coded as a
preliminary meeting in which the project was presented. In fact, categorical variable (i.e., response categories: “No” vs.
the request as to whether the prisoners intended to participate “Possibly/Partially” vs. “Yes”).
to the preliminary meeting had been made by the penitentiary The characteristics of incarcerated groups (SO vs. OO) and
institutions and kept confidential. The various stages of sample their risk factors assessed with HCR-20V3 were compared using
recruitment are shown in Supplementary Material. parametric tests for the continuous variables (Student’s t-test,

once verified the assumption of normality with Kolmogorov- were detected in a lower proportion: 30.0% experienced a
Smirnov test) and non-parametric tests for categorical variables victimization or trauma, and 26.8% reported experiences of
(Chi-square and Fisher’s Exact Test). adverse child rearing.
A binary logistic model was used to estimate the effect of Six out of eleven risk factors considered in the analysis
the HCR-20V3 historical risk factors in predicting criminal displayed a significant association with the groups: (a,b) problems
careers, with being SO or OO as the dichotomous dependent with violence (as a child and as an adult), (c,d) problems with
variable. Only those risk factors that showed a significant other antisocial behavior (as a child and as an adult), (e) problems
difference in the comparison of the two groups were included with non-intimate relationships and (f) problems with substance
as predictors in the regression model. The assumptions related use. Prisoners included in the OO group seemed to be portrayed
to the sample size adequacy, the independence of observations, by a history of problems with violence. Considering the subscale
and the lack of multicollinearity among the independent variables “as an adolescent (13–17)” and summing the relative frequencies
were verified. Model fit was assessed through Hosmer-Lemeshow of the two categories “Possibly/Partially” or “Yes”, this risk factor
test. The Wald’s statistic was used to analyze the predictors’ was observed in the 47.1% of the OO group, against the 19.4%
contribution to the explanation of the dependent variable. of the SO group (χ2 (2) = 19.285; p = 0.000). Although a history
The Odds Ratio and their 95% confidence intervals were of problems with violence as an adult was more frequent in
computed for each predictor’s category. A second binary logistic the SO group than in the OO group (respectively: 65.9 and
model was fitted to assess the interaction effects with PCL-Rs’ 58.8%), the lack of this risk factor proved to be double than that
Factors 1 and 2. of the OO group (respectively: 20.5 and 11.8%; χ2 (2) = 7.960;
Data analysis was performed with SPSS-IBM v25 software, p = 0.019). Also the other antisocial behaviors characterized the
setting significance at p < 0.05. OO group: during the adolescence, this problem was detected
(“Possibly/Partially” or “Yes”) in the 56.8% of the OO group
(SO = 27.2%; χ2 (2) = 17.288; p = 0.000), and, as an adult, this
RESULTS percentage was 77.4% (SO = 50.0%; χ2 (2) = 16.615; p = 0.000).
The last risk factor associated with the OO group concerned
Table 1 summarizes the characteristics of the two substance use (χ2 (2) = 25.697; p = 0.000).
groups of subjects (SO vs. OO). The total mean age was Finally, a history of problems with non-intimate relationships
45.12 ± 12.732, and it was significantly higher in the was the only risk factor linked to the criminal career of the SO: in
SO group (t (188) = −3.947; p = 0.000). About half of this group, it was definitely assessed in the 43.2% of the subjects,
the convicted was unemployed (46.8%), and a percentage against the 29.4% of the OO group (χ2 (2) = 7.416; p = 0.025).
of 11.1% showed a diagnosis of psychopathy (PCL- The results of Student’s t-tests showed that as compared
R score ≥ 25). No association was found between the with OO, the group of SO had significantly higher and lower
employment status, or the diagnosis of psychopathy, and scores, respectively, on the PCL-R F1 Interpersonal and Affective
the two groups of subjects. Deficits and PCL-R F2 Antisocial Behavior. In both groups,
Considering the educational attainment, the secondary school the scores on the PCL-R factors were moderately correlated to
was the most frequent (41.1%), followed by high school (38.4%). each other. The results of the comparisons on the PCL-R factor
A higher school level was associated with the SO group scores and the correlations in the two groups are presented in
(χ2 (3) = 24.186; p = 0.000). Chi-Square post hoc tests were Supplementary Material.
performed by checking the significance of the adjusted residuals These six mentioned factors presenting a significant
of the contingency table cells, with Bonferroni correction. association with the study groups were included as predictors
The results of the post hoc tests showed that SO and OO in a binary logistic regression model. The results of the logistic
groups showed significant adjusted residuals for secondary regression analysis are presented in Table 3 where the role of
school and university degree (both, respectively: p = 0.022 and HCR-20V3 historical risk factors is entered as predictors and
p = 0.001). the criminal career (defined as being a SO or OO) is entered as
The results of the comparisons on the HCR-20 V3 between dichotomous outcome. The lack of significance of the Hosmer-
SO and OO on the risk factors carried out by non-parametric Lemeshow test (χ2 (8) = 7.078; p = 0.528) proved the goodness of
tests are presented in Table 2. Post hoc analysis results are the model fitting. The percentage of predicted cases was 73% out
displayed in the note of Table 3. Some of the risk factors of the total group.
assessed as definitely present (coded as “Yes”) showed a high Three risk factors did not show a significant contribution
prevalence in the total group of subjects. A history of problems to the dependent variable: problems with violence (as an
with violence as an adult (18 and over) was detected in the adult) and problems with antisocial behavior (as an adolescent
62.1% of subjects, and problems with antisocial behavior, once and as an adult).
again as an adult, pertained to the 46.3% of the prisoners Two risk factors provided a significant effect on the dependent
enrolled in the study. Problems with relationships showed a variable but in favor of the “other offenders” criminal career:
high prevalence in this sample, intimate and non-intimate as the possible, or partial, presence of problems with violence as
well (respectively: 47.9 and 35.8%). A history of problems an adolescent (β = −1.237; Wald’s χ2 (1) = 5.090; p = 0.024)
with substance use was definitely found in the 34.2% of the and the certain history of substance use (β = −1.615; Wald’s
subjects, whilst risk factors related to traumatic experiences χ2 (1) = 11.092; p < 0.001). According to these results, the

TABLE 1 | “Sex Offenders” (SO) and “Other Offenders” (OO) characteristics: age, employment status, educational attainment, and diagnosis of psychopathy.
Total SO (n = 88) OO (n = 102) Statistic p
Age (mean ± SD) 45.12 ± 12.732 48.90 ± 13.281 41.85 ± 11.232 t(188) = −3.947 0.000
Employment status n.(%) Employed 101 (53.2) 50 (56.8) 51 (50.0) Fisher’s Exact Test 0.383
Unemployed 89 (46.8) 38 (43.2) 51 (50.0)
Educational attainment n.(%) Primary school 13 (6.8) 2 (2.3) 11 (10.8) χ2 (3) = 24.186 0.000
Secondary school 78 (41.1) 26 (29.5) 52 (51.0)
High school 73 (38.4) 39 (44.3) 34 (33.3)
University degree 26 (13.7) 21 (23.9) 5 (4.9)
Diagnosis of psychopathy (PCL-R ≥ 25) n.(%) Yes 21 (11.1) 9 (10.2) 12 (11.8) Fisher’s Exact Test 0.819
No 169 (88.9) 79 (89.8) 90 (88.2)
Bold values represent statistically significant p-values.
detection of these risk factors was associated with a lower β = −0.533, Wald’s χ2 (1) = 13.396, p = 0.000; “Yes”: β = −0.231,
likelihood of being “SO”. Wald’s χ2 (1) = 10.291, p = 0.001).
The presence of a history of problems with non-intimate
relationships was the only significant risk factor explaining the
outcome “SO”. The high value of the Wald’s statistic (6.325) DISCUSSION
expressed its relevant contribution to the regression model.
Looking at the odds ratio, the prisoners experiencing problems The challenge of this study was to border the investigation
with non-intimate relationships were 3.610 (95% CI 1.327–9.819) concerning the association between risk factors and criminal
times more likely to be a “SO” than another type of offender. career, to the identification of those lifetime events that can
Although no significant difference was found between SO and discriminate sex offenders from other offenders. The risk factors
OO on their psychopathological conditions, a second logistic were assessed by a set of items of the HCR-20 V3 Historical
regression was estimated, using the only significant predictor scale. As compared with the SO, the OO group showed a greater
for the SO category in interaction with the two PCL-R factors: likelihood of having a history of problems with violence and
(F1) interpersonal and affective deficits, (F2) antisocial behavior. antisocial behavior (as adolescents and adults), together with
The results of the logistic regression analysis are shown in problems with substance abuse. These findings were in line with
Table 4 where the interaction effects between the significant the literature evidence. Only one risk factor, i.e., the presence
risk factor (i.e., Non-Intimate Relationships) and the two PCL- of problems in non-intimate relationships, discriminated the
R factors (i.e., Interpersonal and Affective Deficits and Antisocial criminal career of sex offenders from other types of criminals. In
Behavior, respectively) were included as predictors and the sexual the HCR-20 V3 manual, non-intimate relationships are depicted
career (i.e., being a SO vs. OO) is included as dichotomous as the bonds with the family members, the friends, or with generic
outcome. The lack of significance of the Hosmer-Lemeshow acquaintances which do not involve the sexual dimension. Social
test (χ2 (6) = 2.896; p = 0.822) confirmed the goodness of isolation, emotional distance, instability or conflict, manipulation
fit, even if the percentage of predicted cases was quite low of others, inappropriate sexualization and violence in non-
(66% out of the total group). In other words, a 29% of the intimate relationships and escalation of problems are indicators
variance between the SO and OO groups was attributed to the of the presence of such problems. As compared with the OO
psychopathy factors when we considered problems with non- group, SO showed higher and lower levels, respectively, of
intimate relationships. psychopathic interpersonal and affective deficits and antisocial
Both PCL-R factors showed a significant interaction with behavior. Finally, the present findings confirmed the importance
the existence of problems with the non-intimate relationships of deficient non-intimate relationships in predicting the criminal
detected with HCR-20 V3, but with opposite effects. Increasing career of sex offenders.
values of the PCL-R Factor 1 (interpersonal and affective deficits) Furthermore, this association was boosted by psychopathic
provided an increased likelihood of being SO when problems traits related to interpersonal and affective deficits, but not by
with the non-intimate relationships were possibly/partially antisocial behavior. The moderator role of interpersonal/affective
or certainly present (“Possibly/partially”: β = 0.358, Wald’s deficits in the relation between problems in non-intimate
χ2 (1) = 9.493, p = 0.002; “Yes”: β = 0.251, Wald’s χ2 (1) = 10.824, relationships and the risk of being SO vs. OO seems to be
p = 0.001). On the contrary, increasing values of the PCL-R Factor partially in line with a meta-analysis which indicates that
2 (antisocial behavior) were associated to a decreased likelihood this psychopathy facet is associated with violence (Kennealy
of being SO when problems with the non-intimate relationships et al., 2010). The present findings suggest that during clinical
were possibly/partially or certainly present (“Possibly/partially”: practice the assessment of SO should focus on social isolation,

TABLE 2 | Comparison among “sex offenders” (SO) and “other offenders” (OO) about the risk factors assessed with HCR-20-historical scale (items: H1a, H1b, H1c,
H2a, H2b, H2c, H3a, H3b, H5, H8a, H8b).
Total SO (n = 88) OO (n = 102) Statistic p
HCR-20V3 H1a Violence: as a child (12 and under) n.(%) No 144 (75.8) 71 (80.7) 73 (71.6) χ2 (2) = 5.436 0.066
Possibly/Partially 32 (16.8) 9 (10.2) 23 (22.5)
Yes 14 (7.4) 8 (9.1) 6 (5.9)
HCR-20V3 H1b Violence: as an adolescent (13–17) n.(%) No 125 (65.8) 71 (80.7) 54 (52.9) χ2 (2) = 19.285 0.000
Yes 25 (13.2) 10 (11.4) 15 (14.7)
HCR-20V3 H1c Violence: as an adult (18 and over) n.(%) No 30 (15.8) 18 (20.5) 12 (11.8) χ2 (2) = 7.960 0.019
Yes 118 (62.1) 58 (65.9) 60 (58.8)
HCR-20V3 H2a Other antisocial behavior: as a child (12 No 151 (79.5) 72 (81.8) 79 (77.5) χ2 (2) = 1.882 0.390
and under) n.(%)
Yes 20 (10.5) 10 (11.4) 10 (9.8)
HCR-20V3 H2b Other antisocial behavior: as an adolescent No 108 (56.8) 64 (72.7) 44 (43.1) χ2 (2) = 17.288 0.000
(13–17) n.(%)
Yes 36 (18.9) 12 (13.6) 24 (23.5)
HCR-20V3 H2c Other antisocial behavior: as an adult (18 No 67 (35.3) 44 (50.0) 23 (22.5) χ2 (2) = 16.615 0.000
and over) n.(%)
Yes 88 (46.3) 34 (38.6) 54 (52.9)
HCR-20V3 H3a Relationships: intimate n.(%) No 43 (22.6) 19 (21.6) 24 (23.5) χ2 (2) = 0.293 0.864
Yes 91 (47.9) 44 (50.0) 47 (46.1)
HCR-20V3 H3b Relationships: non-intimate n.(%) No 68 (35.8) 33 (37.5) 35 (34.3) χ2 (2) = 7.416 0.025
Yes 68 (35.8) 38 (43.2) 30 (29.4)
HCR-20V3 H5 Substance use n.(%) No 88 (46.3) 55 (62.5) 33 (32.4) χ2 (2) = 25.697 0.000
Yes 65 (34.2) 14 (15.9) 51 (50.0)
HCR-20V3 H8a Traumatic experiences: No 97 (51.1) 50 (56.8) 47 (46.1) χ2 (2) = 2.272 0.321
victimization/trauma n.(%)
Yes 57 (30.0) 24 (27.3) 33 (32.4)
HCR-20V3 H8b Traumatic experiences: adverse child No 82 (43.2) 44 (50.0) 38 (37.3) χ2 (2) = 3.866 0.145
rearing experiences n.(%)
Yes 51 (26.8) 23 (26.1) 28 (27.5)
Post hoc tests and the significance of the standardized adjusted residuals (ns = not significant).
a: No-SO: p = 0.000; No-OO: p = 0.000; Possibly/Partially-SO: p = 0,000; Possibly/Partially-OO: p = 0,000; Yes-SO = ns; Yes-OO = ns.
b: No-SO: p = ns; No-OO: p = ns; Possibly/Partially-SO: p = ns; Possibly/Partially-OO: p = ns; Yes-SO: p = ns; Yes-OO: p = ns.
c: No-SO: p = 0.000; No-OO: p = 0.000; Possibly/Partially-SO: p = 0,008; Possibly/Partially-OO: p = 0,000; Yes-SO: p = ns; Yes-OO: p = ns.
d: No-SO: p = 0.001; No-OO: p = 0.001; Possibly/Partially-SO: p = ns; Possibly/Partially-OO: p = ns; Yes-SO: p = ns; Yes-OO: p = ns.
e: No-SO: p = ns; No-OO: p = ns; Possibly/Partially-SO: p = ns; Possibly/Partially-OO: p = ns; Yes-SO: p = ns; Yes-OO: p = ns.
f: No-SO: p = 0.000; No-OO: p = 0.000; Possibly/Partially-SO: p = ns; Possibly/Partially-OO: p = ns; Yes-SO: p = 0.000; Yes-OO: p = 0.000.

TABLE 3 | Results of the binary logistic model fitting: predictive effects of HCR-20V3 Historical items.
Predictors β S.E. Wald’s statistic p OR OR 95% CI
Lower Upper
limit limit
HCR-20V3 H1b Violence: as an adolescent (13–17)

Violence: (13–17) - No Wald’s χ2 (2) = 5.100 0.078
Violence: (13–17) - Possibly/Partially -1.237 0,548 Wald’s χ2 (1) = 5.090 0.024 0.290 0.099 0.850
Violence: (13–17) - Yes -0.389 0.617 Wald’s χ2 (1) = 0.398 0.528 0.678 0.202 2.270
HCR-20V3 H1c Violence: as an adult (18 and over)
Violence: (18 and over) - No Wald’s χ2 (2) = 2.953 0.228
Violence: (18 and over) - Possibly/Partially -0.508 0.608 Wald’s χ2 (1) = 0.700 0.403 0.602 0.183 1.979
Violence: (18 and over) - Yes 0.330 0.495 Wald’s χ2 (1) = 0.445 0.505 1.391 0.527 3.672
HCR-20V3 H2b Other antisocial behavior: as an adolescent (13–17)
Other antisocial behavior: (13–17) - No Wald’s χ2 (2) = 0.756 0.685
Other antisocial behavior: (13–17) - Possibly/Partially -0.480 0.579 Wald’s χ2 (1) = 0.687 0.407 0.619 0.199 1.926
Other antisocial behavior: (13–17) - Yes -0.160 0.694 Wald’s χ2 (1) = 0.053 0.817 0.852 0.218 3.323
HCR-20V3 H2c Other antisocial behavior: as an adult (18 and over)
Other antisocial behavior: (18 and over) - No Wald’s χ2 (2) = 3.669 0.160
Other antisocial behavior: (18 and over) - Possibly/Partially -0.948 0.567 Wald’s χ2 (1) = 2.793 0.095 0.388 0.128 1.178
Other antisocial behavior: (18 and over) - Yes -0.840 0.537 Wald’s χ2 (1) = 2.448 0.118 0.432 0.151 1.236
HCR-20V3 H3b Non-intimate relationships
Relationships: non-intimate - No Wald’s χ2 (2) = 8.127 0.017
Relationships: non-intimate - Possibly/Partially 0.005 0.47 Wald’s χ2 (1) = 0.000 0.991 1.005 0.400 2.524
Relationships: non-intimate - Yes 1.284 0.51 Wald’s χ2 (1) = 6.325 0.012 3.610 1.327 9.819
HCR-20V3 H5 Substance use
Substance use - No Wald’s χ2 (2) = 14.402 0.001
Substance use - Possibly/Partially 0.351 0.515 Wald’s χ2 (1) = 0.465 0.495 1.421 0.518 3.898
Substance use - Yes -1.615 0.485 Wald’s χ2 (1) = 11.092 0.001 0.199 0.077 0.515
Dichotomous dependent variable: type of group (SO vs. OO).

Abbreviations: HCR-20 V3 = Historical, Clinical and Risk Management, OO = Other offenders, OR = Odds Ratio, SO = Sex Offenders.
emotional distance, instability or conflict, manipulation of others, deficits in interpersonal empathy and emotion regulation
inappropriate sexualization and violence, particularly in non- (Gillespie et al., 2018; Schuler et al., 2021), while other types of
intimate relationships. Tailored treatment programs aimed to offenders would show antisocial personality traits (Mazzoni et al.,
prevent relapse in SO (Carabellese et al., 2020b; Gualtieri et al., 2018; Azevedo et al., 2020).
2020b) should address these deficits in non-intimate relationships The greater likelihood of a history of violence and antisocial
and target psychopathic traits, specifically interpersonal and behavior (as adolescents and adults), together with problems with
affective traits. Cognitive behavioral therapy has been found to substance abuse amongst OO as compared with SO, is a result in
be an effective treatment for psychopathic traits (Salekin et al., line with the literature evidence (e.g., Gottfredson et al., 2008).
2010). The fact that SO showed specific interpersonal deficits in The importance of violence during the adolescence as predictor
the domain of non-intimate relationships suggests that a group of offending patterns in adulthood is still debated (Cardwell and
format of treatment might be more helpful in addressing such Piquero, 2018), even if violent behaviors during the childhood,
difficulties in this population. In a group setting, the process for example the presence of aggressions (Juon et al., 2006), seem
of confronting with the experiences of other offenders might to predict serious offending in adulthood. The link between drug
improve emotional awareness and regulation (Jennings and use and criminal behavior has received attention from scholars
Sawyer, 2003; Levenson et al., 2009), and it might reduce the risk (Tonry and Wilson, 1990; Bennett et al., 2008; Gottfredson et al.,
of drop-out from treatment, like for other clinical populations 2008; Liu et al., 2018).
with impulsive traits (Pozza and Dèttore, 2017). Although several studies related to sex offenders’ criminal
The result showing that SO had higher and lower levels of, career provided evidence of the effects of traumatic experiences
respectively, interpersonal and affective deficits and antisocial during their childhood, also defined as Adverse Childhood
behavior suggests that psychopathic traits could be differentially Experiences (ACE), this relationship was not confirmed by
related to sex crimes and other types of crimes. This result the findings illustrated in this paper. Considering the studies
appears consistent with literature data which extensively which compare male sex offenders with the general population
showed that sex offenders would be characterized by severe (Levenson et al., 2016; Kingston et al., 2017), sex offenders

TABLE 4 | Results of the binary logistic model with SO/OO as outcome and the HCR20-V3 item H8 non-intimate as predictor, in interaction with the PCL-R Factor 1
Interpersonal and affective deficits and Factor 2 Antisocial behavior; Odds Ratio and 95% confidence interval of the Odds Ratio.
Predictors β S.E. Wald’s statistic P OR OR 95% CI
Lower Upper
limit limit
Interaction effects between

HCR-20V3 H3b non-intimate
relationships and PCL-R
factor 1
Non-intimate relationships: Wald’s χ2 (2) = 19.766 0.000
No ∗ PCL-R F1 Interpersonal
and Affective Deficits interaction
Non-intimate relationships: 0.358 0.116 Wald’s χ2 (1) = 9.493 0.002 1.430 1.139 1.796
Possibly/Partially ∗ PCL-R F1
Interpersonal and Affective
Deficits interaction
Non-intimate relationships: 0.251 0.076 Wald’s χ2 (1) = 10.824 0.001 1.286 1.107 1.493
Yes ∗ PCL-R F1 Interpersonal
and Affective Deficits interaction
Interaction effects between

HCR-20V3 H3b non-intimate
relationships and PCL-R
factor 2
Non-intimate relationships: Wald’s χ2 (2) = 23.521 0.000
No ∗ PCL-R F2 Antisocial
Behavior interaction
Non-intimate relationships: −0.533 0.146 Wald’s χ2 (1) = 13.396 0.000 0.587 0.441 0.781
Possibly/Partially ∗ PCL-R F2
Antisocial Behavior interaction
Non-intimate relationships: −0.231 0.072 Wald’s χ2 (1) = 10.291 0.001 0.794 0.689 0.914
Yes ∗ PCL-R F2 Antisocial
Behavior interaction
Abbreviations. HCR-20V3 = Historical, Clinical and Risk Management PCL-R = Psychopathy Checklist-Revised, SO = Sexual offenders, OO = Other offenders,
OR = Odds ratio. Bold values represent statistically significant p-values.
showed a larger odd of being victims of sexual abuse, verbal or the HCR-20 V3 used to assess the risk factors. The Historical
physical abuse and emotional neglect, but also signals of adverse Clinical Risk Management-20 (HCRS-20) has been criticized
child-rearing experiences (incarcerated family members, parental for its low field validity in clinical forensic psychiatry settings
dysfunctions, disordered familiar environment). Victimization or (Jeandarme et al., 2017; Tully, 2017). On the contrary, other
other traumatic experiences during the childhood, like physical, studies supported the psychometric validity and the clinical
sexual and emotional abuse, were more prevalent in sex offenders utility of the scale (Judges et al., 2016), highlighting that the
than in other kinds of offenders (Lee et al., 2002; Jespersen et al., scoring subjectivity of the scale did not impair its reliability, above
2009; Seto and Lalumière, 2010; Reavis et al., 2013; Drury et al., all for the HCR-20 V3 historical scale (Rufino et al., 2011).
2019; Gualtieri et al., 2020a). In addition, evidence of adverse A third problem can be associated to the method used
child-rearing experiences was more frequent as well, like a history by this scale to assess the risk factors, which are described
of family dysfunctions or a condition of poor parenting style (Lee by a wide variety of lifetime events such as trauma or
et al., 2002; Reavis et al., 2013; Sigre-Leirós et al., 2016). victimization related to parental and non-parental sexual abuse
Many reasons can explain the divergent findings reported during childhood/adolescence, witnessing of domestic violence,
in this paper. First, the methodological issues that affect this bullying, or lifetime interpersonal victimization. As much
research topic. The review of McMillan et al. (2008), specifically heterogeneity of events is also listed to measure the presence of
carried out on sexual offenders against children, pointed out adverse child-rearing experiences (overly rigid parenting styles,
a series of problems related to the temporal relationship unstable family relations, conviction of parents before the subject
between a putative risk factor and sexual perpetration, the reached 10 years of age, death of a parent during childhood
effects of other variables, and the complexity in modeling or adolescence, separation from parents before 17 years of age,
developmental phenomena. parental substance abuse). Probably, this wide set of experiences
Secondarily, this lack of evidence may be due to the describing a single indicator can reduce its specificity in detecting
measurement scale administered to the participants, in particular the risk factors.

Some limitations of the study should be pointed out. CONCLUSION

Firstly, the recruitment of the sample might be subjected
to a bias as this process took place in agreement with the The role of life events as risk factors for sexual crime is a
Penitentiary Administration of the Ministry of Justice. The first complex and long-debated topic in criminological research. The
stage of recruitment was carried out exclusively by the prison present study is the first contribution which compared SO and
administration, which asked the prisoners if they were interested OO on the risk factors assessed by the HCR-20V3 and explored
in participating to a preliminary meeting with the researchers. the role of psychopathic traits as moderators of the relation
On that occasion, the researchers presented the project in detail between specific life events and the risk of having committed
and asked the prisoners for the informed consent. However, sexual crime or other types of crime. The present findings shed
other information that could have been useful for the purpose of further light on this issue showing the role of deficient non-
the research was not provided to the researchers by the prison intimate relationships in predicting the criminal career of sex
administration. In particular, no data were provided about the offenders, and suggesting that specific psychopathic traits, i.e.,
specific types of crime of the OO group or how many of both interpersonal/affective deficits, but not antisocial behavior, can
the groups of offenders were recidivists. However, it was well- moderate the relation between a history of problems in non-
established that those prisoners belonging to the group of SO did intimate relationships and sexual crime.
not commit any other crimes than sexual ones.
Secondly, the cross-sectional design did not allow us to
draw firm conclusion about the role of the risk factors in DATA AVAILABILITY STATEMENT
the likelihood of being SO or OO, particularly in the risk of
recidivism. Future research should use a longitudinal design to The original contributions presented in the study are included
ascertain problems in non-intimate relationships and whether in the article/Supplementary Material, further inquiries can be
interpersonal/affective deficits can predict a higher risk of directed to the corresponding author.
recidivism in sexual crime as compared with the risk of recidivism
in other types of crime.
Another relevant limitation concerns the fact that individuals ETHICS STATEMENT
with a psychiatric history were excluded. This aspect reduced
The studies involving human participants were reviewed
the external validity of this research. The findings observed
and approved by Director of Penitentiary Administration
in the present sample may be applied to a minority of the
Department of Ministry of Justice. The patients/participants
SO population. In fact, SO with no other criminal record
provided their written informed consent to participate in this
and no lifetime psychiatric diagnosis typically represent only
study.
a small proportion of SO that is lower than 20%, as the
lack of psychiatric comorbidities is generally observed in
a 7–15% proportion of the SO population (e.g., Raymond AUTHOR CONTRIBUTIONS
et al., 1999; Dunsieth et al., 2004). Future studies should
also include individuals with lifetime psychiatric disorders FF performed the statistical analysis. AP interpreted the results
and control the effect of the presence of psychiatric of the analyses and wrote the manuscript with support from FF,
conditions in the risk of being SO or OO, particularly FuC, and FeC. AS and GS took care of the bibliographic research.
severe psychiatric disorders such as psychosis/bipolar and FuC was the corresponding author. GM collected the data. GG
personality disorders (Van Wijk et al., 2007; Coluccia et al., 2015; contributed to the bibliographic research. FeC conceived of the
Chen et al., 2016; Eastman et al., 2019). However, the presented idea and collected the data. RC and AC performed
role of psychiatric disorders in the risk of violence has a general review. All authors contributed to the article and
been questioned as a dimensional approach is considered approved the submitted version.
more informative (Graz et al., 2009). Finally, an area that
deserves further exploration regards the moderator effect of
protective factors in the relation between life events and crime SUPPLEMENTARY MATERIAL
type. It would be interesting to examine whether certain
protective factors such as the capacity to forgive oneself The Supplementary Material for this article can be found
(Barcaccia et al., 2019) could protect the individual from the online at: https://www.frontiersin.org/articles/10.3389/fpsyg.
criminal behavior. 2021.600370/full#supplementary-material
REFERENCES offenders with antisocial personality disorder. PLoS One 15:e0229876.

doi: 10.1371/journal.pone.0229876
Andersen, H. S., Sestoft, D., Lillebaek, T., Mortensen, E. L., and Kramp, P. (1999). Barcaccia, B., Pallini, S., Pozza, A., Milioni, M., Baiocco, R., Mancini, F., et al.
Psychopathy and psychopathological profiles in prisoners on remand. Acta (2019). Forgiving adolescents: far from depression, close to well-being. Front.
Psychiatr. Scand. 99, 33–39. doi: 10.1111/j.1600-0447.1999.tb05382.x Psychol. 10:1725. doi: 10.3389/fpsyg.2019.01725
Azevedo, J., Vieira-Coelho, M., Castelo-Branco, M., Coelho, R., and Figueiredo- Bennett, T., Holloway, K., and Farrington, D. (2008). The statistical
Braga, M. (2020). Impulsive and premeditated aggression in male association between drug misuse and crime: a meta-analysis.

Aggress. Violent Behav. 13, 107–118. doi: 10.1016/j.avb.2008. and offence-related characteristics and treatment engagement.
02.001 J. Sex. Aggress. 25, 105–115. doi: 10.1080/13552600.2018.155
Burton, D. L. (2008). An Exploratory Evaluation of the Contribution of 1501
Personality and Childhood Sexual Victimization to the Development of Eher, R., Rettenberger, M., and Turner, D. (2019). The prevalence of mental
Sexually Abusive Behavior. Sex. Abuse 20, 102–115. doi: 10.1177/107906320831 disorders in incarcerated contact sexual offenders. Acta Psychiatr. Scand. 139,
5352 572–581. doi: 10.1111/acps.13024
Campbell, M. A., French, S., and Gendreau, P. (2009). The prediction of violence Gannon, T. A., Olver, M. E., Mallion, J. S., and James, M. (2019). Does
in adult offenders: a meta-analytic comparison of instruments and methods specialized psychological treatment for offending reduce recidivism? A
of assessment. Crim. Justice Behav. 36, 567–590. doi: 10.1177/009385480933 meta-analysis examining staff and program variables as predictors of
3610 treatment effectiveness. Clin. Psychol. Rev. 73:101752. doi: 10.1016/j.cpr.2019.10
Campobasso, C., Colonna, M., Carabellese, F., Grattagliano, I., Candelli, C., 1752
Morton, R. J., et al. (2009). A serial killer of elderly women: analysis of a Gillespie, S. M., Garofalo, C., and Velotti, P. (2018). Emotion regulation,
multi-victim homicide investigation. Forensic Sci. Int. 185, 7–11. doi: 10.1016/j. mindfulness, and alexithymia: specific or general impairments in sexual, violent,
forsciint.2008.12.023 and homicide offenders? J. Crim. Justice 58, 56–66. doi: 10.1016/j.jcrimjus.2018.
Carabellese, F., Felthous, A. R., Mandarelli, G., Montalbò, D., La Tegola, D., 07.006
Parmigiani, G., et al. (2020a). Women and Men who Committed Murder: Gottfredson, D. C., Kearley, B. W., and Bushway, S. D. (2008). Substance use,
male/Female Psychopathic Homicides. J. Forensic Sci. 65, 1619–1626. doi: 10. drug treatment, and crime: an examination of intra-individual variation in a
1111/1556-4029.14450 drug court population. J. Drug Issues 38, 601–630. doi: 10.1177/00220426080380
Carabellese, F., Urbano, M., Carabellese, F., Gualtieri, G., Pozza, A., Ferretti, F., 0211
et al. (2020b). Il trattamento giudiziario del sex offender: vecchi limiti, nuove Grann, M., Langstrom, N., Tengstrom, A., and Stalenheim, E. G. (1998). Reliability
opportunità. Ital. J. Criminol. 14, 229–241. of File-Based Retrospective Ratings of Psychopathy With the PCL-R. J. Pers.
Carabellese, F., Rocca, G., Candelli, C., La Tegola, D., and Birkhoff, J. (2012). La Assess. 70, 416–426. doi: 10.1207/s15327752jpa7003_2
gestione degli autori di reati sessuali tra psicopatologia e rischio di recidiva. Graz, C., Etschel, E., Schoech, H., and Soyka, M. (2009). Criminal behaviour and
Prospettive trattamentali. Rassegna Ital. di Criminol. 2, 70–83. violent crimes in former inpatients with affective disorder. J. Affec. Disord. 117,
Carabellese, F., Vinci, F., and Catanesi, R. (2008). Compatibility Between Mental 98–103. doi: 10.1016/j.jad.2008.12.007
Disorder and Mental Capacity: analysis of a Particolar Case of Group Sexual Gualtieri, G., Ferretti, F., Masti, A., Pozza, A., and Coluccia, A. (2020a). Post-
Homicide. J. Forensic Sci. 53, 1450–1454. doi: 10.1111/j.1556-4029.2008.0 traumatic Stress Disorder in Prisoners’ Offspring: a Systematic Review and
0897 Meta-analysis. Clin. Pract. Epidemiol. Ment. Health 16, 36–45. doi: 10.2174/
Cardwell, S. M., and Piquero, A. R. (2018). Does Violence in Adolescence 1745017902016010036
Differentially Predict Offending Patterns in Early Adulthood? Int. J. Offender Gualtieri, G., Traverso, S., Pozza, A., Ferretti, F., Carabellese, F., Gusinu, R.,
Ther. Comp. Criminol. 62, 1603–1628. doi: 10.1177/0306624x16688978 et al. (2020b). Clinical risk management in High-Security Forensic Psychiatry
Caretti, V., Ciappi, S., Scarpa, F., Castelletti, L., Catanesi, R., Carabellese, F., et al. Residences. Protecting patients and health professionals: perspectives and
(2019). HCR 20-3 Checklist per la valutazione del rischio di recidiva di un crimine critical issues of the Law 81/2014. Clin. Ter. 171, e97–e100.
violento. Adattamento Italiano. Firenze: Hogrefe. Hanson, R. K., Bourgon, G., Helmus, L., and Hodgson, S. (2009). A Meta-analysis of
Caretti, V., Manzi, G. S., Schimmenti, A., and Seragusa, L. (2011). PCL-R. Hare the Effectiveness of Treatment for Sexual Offenders: Risk, Need, and Responsivity.
Psychopathy Checklist Revised. Firenze: Giunti. Ottawa: Corrections Research, Public Safety Canada.
Cartwright, J. K., Desmarais, S. L., Hazel, J., Griffith, T., and Azizian, A. (2018). Hanson, R. K., and Morton-Bourgon, K. E. (2005). The characteristics
Predictive validity of HCR-20, START, and static-99R assessments in predicting of persistent sexual offenders: a meta-analysis of recidivism studies.
institutional aggression among sexual offenders. Law Hum. Behav. 42, 13–25. J. Consult. Clin. Psychol. 73, 1154–1163. doi: 10.1037/0022-006X.73.6.
doi: 10.1037/lhb0000263 1154
Catanesi, R., Mandarelli, G., Ferracuti, S., Valerio, A., and Carabellese, F. (2019). Hare, R. D., Harpur, T. J., and Hakstian, A. R. (1989). Two-factor conceptualization
The new residential forensic psychiatric system (REMS): a one-year population of psychopathy: construct validity and assessment implications. Psychol. Assess.
study. Ital. J. Criminol. 13, 7–23. doi: 10.7347/RIC.NS2019 1, 6–17. doi: 10.1037/1040-3590.1.1
Chen, Y. Y., Chen, C. Y., and Hung, D. L. (2016). Assessment of psychiatric Hare, R. D., and Neumann, C. N. (2006). “The PCL-R Assessment of Psychopathy:
disorders among sex offenders: prevalence and associations with criminal development, Structural Properties, and New Directions,” in Handbook of
history. Crim. Behav. Ment. Health 26, 30–37. doi: 10.1002/cbm. Psychopathy, ed. C. Patrick (New York: Guilford), 58–88.
1926 Ismail, G., and Looman, J. (2018). Field Inter-Rater Reliability
Coluccia, A., Ferretti, F., Fagiolini, A., and Pozza, A. (2015). Incidenza e fattori of the Psychopathy Checklist–Revised. Int. J. Offender Ther.
di rischio per disturbi psicotici nelle popolazioni migranti in Europa: una Comp. Criminol. 62, 468–481. doi: 10.1177/0306624X1665
meta-analisi di studi trasversali. Ital. J. Criminol. 9, 29–39. 2452
Coluccia, A., Pozza, A., Ferretti, F., Carabellese, F., Masti, A., and Gualtieri, Jeandarme, I., Pouls, C., De Laender, J., Oeic, T. I., and Bogaertsd, S.
G. (2020). Online romance scams: relational dynamics and psychological (2017). Field validity of the HCR-20 in forensic medium security units in
characteristics of the victims and scammers. A scoping review. Clin. Flanders. Psychol. Crime Law. 23, 305–322. doi: 10.1080/1068316x.2016.125
Pract. Epidemiol. Ment. Health 16, 24–35. doi: 10.2174/174501790201601 8467
0024 Jennings, J. L., and Sawyer, S. (2003). Principles and techniques for maximizing
Douglas, K. S., Hart, S. D., Webster, C. D., and Belfrage, H. (2013). HCR-20V3. the effectiveness of group therapy with sex offenders. Sex. Abuse 15, 251–267.
Assessing Risk for Violence. Burnaby: Mental Health, Law, and Policy Institute, doi: 10.1177/107906320301500403
Simon Fraser University. Jespersen, A. F., Lalumière, M. L., and Seto, M. C. (2009). Sexual abuse
Drury, A. J., Elbert, M. J., and DeLisi, M. (2019). Childhood sexual abuse is history among adult sex offenders and non-sex offenders: a meta-
significantly associated with subsequent sexual offending: new evidence among analysis. Child Abuse Negl. 33, 179–192. doi: 10.1016/j.chiabu.2008.0
federal correctional clients. Child Abuse Negl. 95:104035. doi: 10.1016/j.chiabu. 7.004
2019.104035 Judges, R., Egan, V., and Broad, G. (2016). Critique of the Historical Clinical
Dunsieth, N. W. Jr., Nelson, E. B., Brusman-Lovins, L. A., Holcomb, J. L., Beckman, Risk–20, Version 3, Risk Assessment Instrument. J. Forensic Psychol. Pract. 16,
D., Weldge, J. A., et al. (2004). Psychiatric and legal features of 113 men 304–320. doi: 10.1080/15228932.2016.1196102
convicted of sexual offenses. J. Clin. Psychiatry 65, 293–300. doi: 10.4088/jcp. Juon, H. S., Doherty, E. E., and Ensminger, M. E. (2006). Childhood Behavior
v65n0302 and Adult Criminality: cluster Analysis in a Prospective Study of African
Eastman, O., Craissati, J., and Shaw, J. (2019). Young adult sexual Americans. J. Quant. Criminol. 22, 193–214. doi: 10.1007/s10940-006-9
offenders with emerging personality disorders: developmental 008-9

Jüriloo, A., Lauerma, H., Holmalahti, T., Tyni, S., Aarnio, J., Viitanen, P., Pozza, A., Coluccia, A., Kato, T., Gaetani, M., and Ferretti, F. (2019).
et al. (2013). Psychopathic traits in a representative sample of Finnish male The ‘Hikikomori’ syndrome: worldwide prevalence and co-occurring
prisoners. Nord. J. Psychiatry 68, 117–122. doi: 10.3109/08039488.2013.78 major psychiatric disorders: a systematic review and meta-analysis
0259 protocol. BMJ Open 9:e025213. doi: 10.1136/bmjopen-2018-02
Kennealy, P. J., Skeem, J. L., Walters, G. D., and Camp, J. 5213
(2010). Do core interpersonal and affective traits of PCL-R Pozza, A., and Dèttore, D. (2017). Drop-out and efficacy of group versus
psychopathy interact with antisocial behavior and disinhibition to individual cognitive behavioural therapy: what works best for Obsessive-
predict violence? Psychol. Assess. 22, 569–580. doi: 10.1037/a001 Compulsive Disorder? A systematic review and meta-analysis of direct
9618 comparisons. Psychiatry Res. 258, 24–36. doi: 10.1016/j.psychres.2017.0
Kimonis, E. R., Skeem, J. L., Cauffman, E., and Dmitrieva, J. (2011). Are secondary 9.056
variants of juvenile psychopathy more reactively violent and less psychosocially Rasmussen, L. A. (2012). - Victim and Victimizer: the Role of Traumatic
mature than primary variants? Law Hum. Behav. 35, 381–391. doi: 10.1007/ Experiences as Risk Factors for Sexually Abusive Behavior. Isr. J. Psychiatry
s10979-010-9243-3 Relat. Sci. 49, 270–279.
Kingston, D. A., Graham, F. J., and Knight, R. A. (2017). Relations Between Self- Raymond, N. C., Coleman, E., Ohlerking, F., Christenson, G. A., and Miner, M.
Reported Adverse Events in Childhood and Hypersexuality in Adult Male (1999). Psychiatric comorbidity in pedophilic sex offenders. Am. J. Psychiatry
Sexual Offenders. Arch Sex. Behav. 46, 707–720. doi: 10.1007/s10508-016-0 156, 786–788.
873-5 Reavis, J. A., Looman, J., Franco, K. A., and Rojas, B. (2013). Adverse
Kraanen, F. L., and Emmelkamp, P. M. G. (2011). Substance misuse and substance childhood experiences and adult criminality: how long must we live
use disorders in sex offenders: a review. Clin. Psychol. Rev. 31, 478–489. doi: before we possess our own lives? Perm. J. 17, 44–48. doi: 10.7812/TPP/1
10.1016/j.cpr.2010.11.006 2-072
Lee, J. K. P., Jackson, H. J., Pattison, P., and Ward, T. (2002). Developmental risk Rufino, K. A., Boccaccini, M. T., and Guy, L. S. (2011). Scoring subjectivity and
factors for sexual offending. Child Abuse Negl. 26, 73–92. doi: 10.1016/S0145- item performance on measures used to assess violence risk: the PCL-R and
2134(01)00304-0 HCR-20 as exemplars. Assessment 18, 453–463. doi: 10.1177/107319111037
Leistico, A. R., Salekin, R. T., De Coster, J., and Rogers, R. (2008). A 8482
large-scale meta-analysis relating the hare measures of psychopathy to Salekin, R. T., Worley, C., and Grimes, R. D. (2010). Treatment of
antisocial conduct. Law Hum. Behav. 32, 28–45. doi: 10.1007/s10979-007- psychopathy: a review and brief introduction to the mental model
9096-6 approach for psychopathy. Behav. Sci. Law 28, 235–266. doi: 10.1002/bsl.
Levenson, J. S., Macgowan, M. J., Morin, J. W., and Cotter, L. P. (2009). 928
Perceptions of sex offenders about treatment: satisfaction and engagement Schimmenti, A., Carabellese, F. F., and Caretti, V. (2020). “Child maltreatment and
in group therapy. Sex. Abuse 21, 35–56. doi: 10.1177/107906320832 psychopathy,” in The International Handbook of Psychopathic Disorders and the
6072 Law, Volume I: Diagnosis and Treatment, Second Edition, eds A. R. Felthous and
Levenson, J. S., Willis, G. M., and Prescott, D. S. (2016). Adverse Childhood H. Saß (Chichester: John Wiley & Sons, Ltd).
Experiences in the Lives of Male Sex Offenders: implications for Trauma- Schraft, C. V., Kosson, D. S., and McBride, C. K. (2013).
Informed Care. Sex. Abuse 28, 340–359. doi: 10.1177/107906321453 Exposure to violence within home and community environments
5819 and psychopathic tendencies in detained adolescents. Crim.
Liu, L., Chui, W. H., and Chen, Y. (2018). Violent and Non-Violent Justice Behav. 40, 1027–1043. doi: 10.1177/009385481348
Criminal Behavior among Young Chinese Drug Users: a Mixed Methods 6887
Study. Int. J. Env. Res. Public Health 15:432. doi: 10.3390/ijerph1503 Schuler, M., Mohnke, S., Amelung, T., Dziobek, I., Borchardt, V., Gerwinn, H.,
0432 et al. (2021). Empathy in paedophilia and sexual offending against children:
Mandarelli, G., Carabellese, F., Parmigiani, G., Bernardini, F., Pauselli, L., a longitudinal extension. J. Sex. Aggress. 1–18. doi: 10.1080/13552600.2021.
Quartesan, R., et al. (2017). Treatment decision-making capacity in non- 1931721
consensual psychiatric treatment: a multicentre study. Epidemiol. Psychiatr. Sci. Seto, M. C., and Lalumière, M. L. (2010). What is so special about
9, 1–8. male adolescent sexual offending? A review and test of explanations
Mazzoni, G. P., Contena, B., Fanciullacci, S., and Pozza, A. (2018). Analisi through meta-analysis. Psychol. Bull. 136, 526–575. doi: 10.1037/a001
comparativa dei profili di personalità di pazienti affetti da disturbo 9700
borderline e detenuti con diagnosi di disturbo antisociale di personalità Sevecke, K., Franke, S., Kosson, D., and Krischer, M. (2016). Emotional
ottenuti tramite l’MMPI-2. Riv. Psichiatr. 53, 267–273. doi: 10.1708/3000.3 dysregulation and trauma predicting psychopathy dimensions in female and
0006 male juvenile offenders. Child Adolesc. Psychiatry Ment. Health 10:43. doi: 10.
McMillan, D., Hastings, R. P., Salter, D. C., and Skuse, D. H. (2008). Developmental 1186/s13034-016-0130-7
risk factor research and sexual offending against children: a review of some Sigre-Leirós, V., Carvalho, J., and Nobre, P. J. (2016). Early parenting styles and
methodological issues. Arch. Sex. Behav. 37, 877–890. doi: 10.1007/s10508-007- sexual offending behavior: a comparative study. Int. J. Law Psychiatry 46,
9193-0 103–109. doi: 10.1016/j.ijlp.2016.02.042
Myers, W. C., Marrero, L., and Herkov, M. J. (2005). “Forensic Psychiatry Somma, A., Fossati, A., Carabellese, F., Santoro, G., Schimmenti, A.,
and Forensic Psychology,” in Encyclopedia of Forensic and Legal Caretti, V., et al. (2020). Dysfunctional Personality Traits and
Medicine, ed. J. Payne-James (Cambridge: Elsevier Academic Press), Demographic Variables Associated with Violence Risk in Male
444–451. Sexual Offenders: a Study on Italian Adult Inmates. J. Forensic
Ometto, M., de Oliveira, P. A., Milioni, A. L., Dos Santos, Psychiatry Psychol. 32, 242–260. doi: 10.1080/14789949.2020.185
B., Scivoletto, S., Busatto, G. F., et al. (2016). Social skills 7425
and psychopathic traits in maltreated adolescents. Eur. Child Tonry, M., and Wilson, J. Q. (1990). Drugs and Crime. Chicago: University of
Adolesc. Psychiatry 25, 397–405. doi: 10.1007/s00787-015- Chicago Press.
0744-y Tully, J. (2017). HCR-20 shows poor field validity in clinical forensic psychiatry
Porter, S., Newman, E., Tansey, L., and Quayle, E. (2015). Sex offending and social settings. Evid. Based Ment. Health 20, 95–96. doi: 10.1136/eb-2017-10
anxiety: a systematic review. Aggress. Violent Behav. 24, 42–60. doi: 10.1016/j. 2745
avb.2015.04.005 Van Wijk, A. P., Blokland, A. A. J., Duits, N., Vermeiren, R. R. J. M.,
Pozza, A., Coluccia, A., Gualtieri, G., Carabellese, F., Masti, A., and Ferretti, and Harkink, J. (2007). Relating psychiatric disorders, offender and offence
F. (2020). Post-traumatic stress disorder secondary to manic episodes with characteristics in a sample of adolescent sex offenders and non-sex
hypersexuality in bipolar disorder: a case study of forensic psychotherapy. Clin. offenders. Crim. Behav. Ment. Health 17, 15–30. doi: 10.1002/cbm.
Neuropsychiatry 17, 181–188. 628

Wong, S. C. P., and Olver, M. E. (2016). “Risk reduction treatment of psychopathy Publisher’s Note: All claims expressed in this article are solely those of the authors
and applications to mentally disordered offenders,” in Violence in Psychiatry, eds and do not necessarily represent those of their affiliated organizations, or those of
K. D. Warburton and S. M. Stahl (Cambridge: Cambridge University Press), 32 the publisher, the editors and the reviewers. Any product that may be evaluated in
3–331. this article, or claim that may be made by its manufacturer, is not guaranteed or
Yang, M., Wong, S., and Coid, J. (2010). The efficacy of violence prediction: a meta- endorsed by the publisher.
analytic comparison of nine risk assessment tools. Psychol. Bull. 136, 740–767.
doi: 10.1037/a0020473 Copyright © 2021 Ferretti, Pozza, Carabellese, Schimmenti, Santoro, Mandarelli,
Yates, P. M. (2013). Treatment of sexual offenders: research, best practices, and Gualtieri, Carabellese, Catanesi and Coluccia. This is an open-access article
emerging models. Int. J. Behav. Consult. Ther. 8, 89–95. doi: 10.1037/h0100989 distributed under the terms of the Creative Commons Attribution License (CC BY).
The use, distribution or reproduction in other forums is permitted, provided the
Conflict of Interest: The authors declare that the research was conducted in the original author(s) and the copyright owner(s) are credited and that the original
absence of any commercial or financial relationships that could be construed as a publication in this journal is cited, in accordance with accepted academic practice. No
potential conflict of interest. use, distribution or reproduction is permitted which does not comply with these terms.

ORIGINAL RESEARCH
published: 04 November 2021
doi: 10.3389/fpsyg.2021.606797
Acceptance of Sexual Interest in

Minors in Self-Referred Individuals
Under Treatment – An Exploratory
Pilot Study
Ute Lampalzer* , Safiye Tozdan, Fritjof von Franqué and Peer Briken
Institute for Sex Research, Sexual Medicine and Forensic Psychiatry, University Medical Center Hamburg-Eppendorf,
Hamburg, Germany
Some therapists/scientists argue that “acceptance” of sexual interest in minors (SIM),

i.e., the integration of the sexual preference into the individual self-concept, is a
prerequisite for dealing with SIM in a responsible way. However, if one assumes
that – even in some persons – SIM might change over time, “acceptance” could also
Edited by:
Nadia Marie Wager,
run counter to therapeutic targets because the motivation to change as well as the
University of Huddersfield, specific self-efficacy for modifying SIM might be reduced. This exploratory pilot study
United Kingdom analyzes the relationship between acceptance of SIM and (1) dynamic risk for contact
Reviewed by: sexual reoffending, (2) SIM and frequency of the use of child/adolescent (sexual abuse)
Ross M. Bartels,
University of Lincoln, United Kingdom imagery, (3) frequency of sexual desire/behavior toward children/adolescents, and (4)
Katerina Klapilova, the change of the level of acceptance of SIM during the course of treatment. The
National Institute of Mental Health,
Czechia
majority of the participants (N = 79) was not exclusively interested in children (85%) and
*Correspondence:
used child pornography but did not commit child sexual abuse (54%). Acceptance of
Ute Lampalzer SIM, frequency of the use of child/adolescent (sexual abuse) imagery and frequency
u.lampalzer@uke.de of sexual desire/behavior toward children/adolescents are assessed via self-report
Specialty section:
questionnaires, dynamic risk for contact sexual reoffending is measured by STABLE-
This article was submitted to 2007. Pretreatment data are analyzed via Spearman’s correlation (N = 79). Intragroup
analysis compares acceptance of SIM from pre- and posttreatment (n = 35). There
Frontiers in Psychology was no correlation between acceptance of SIM and dynamic risk for contact sexual
Received: 15 September 2020 reoffending. However, there was a medium, positive correlation between acceptance
Accepted: 06 October 2021 of SIM and the frequency of the use of legal imagery of children, a positive correlation
Published: 04 November 2021
between the item “My inclination is an integral part of my personality” and the frequency
Citation:
Lampalzer U, Tozdan S,
of the use of legal imagery of children, and a positive correlation between acceptance
von Franqué F and Briken P (2021) of SIM and the frequency of sexual activities with minors. Acceptance of SIM did not
Acceptance of Sexual Interest
change during the course of treatment. The results suggest that “acceptance” of SIM
in Minors in Self-Referred Individuals
Under Treatment – An Exploratory has to be discussed in a differentiated way, i.e., as possibly being associated with
Pilot Study. positive and negative outcomes as well.
Front. Psychol. 12:606797.
doi: 10.3389/fpsyg.2021.606797 Keywords: child abuse material, pedophilia, prevention of child sexual abuse, psychotherapy, risk of offending

Lampalzer et al. Acceptance of Sexual Interest in Minors
INTRODUCTION explanations” (Cantor and Fedoroff, 2018, p. 205): “Sex crime

rates are dropping [. . .]. The incidence of sex crime rates
Findings from an online survey with 8,718 German males decreases as people age [. . .]. The likelihood that (known)
indicate that 4.1% have sexual fantasies about children and that high-risk sex offenders will re-offend decreases the longer they
3.2% have offended against prepubescent children. But only 0.1% commit no crime [. . .]. The self-report of men and women
reported a pedophilic sexual preference (Dombert et al., 2016). with paraphilic disorders [. . .]. As people grow older, their
The sites of the German network “Kein Täter werden” (means: interests shift to partners similar in age” (Cantor, 2018, p. 205).
not become an offender) offer treatment for people seeking He argues that these findings reflect “aging populations,” “the
therapeutic help because of sexual interest in minors (SIM) and decrease in sex drive that accompanies aging,” “the absence
distress or a risk of sexual (re-)offending (Netzwerk “Kein Täter of critical thinking,” and “sexual behavior with partners of
werden”, 2018). It is a formal requirement in this network that increasing age as they [people] themselves age” (Cantor, 2018,
individuals who receive treatment are currently not in contact p. 205). Researchers who assume that pedophilia is immutable
with the criminal justice system. usually infer that treatment should “focus on [. . .] management”
The construct “acceptance,” as it is understood in this of pedophilia (Seto, 2017, p. 18), i.e., on developing the
exploratory pilot study, means that sexual preference is accepted skills that are needed to regulate and control pedophilic urges
as a “stable and therefore constantly challenging part of the own (Institute for Sexology and Sexual Medicine of the Charité, 2013;
personality” which, being “fate rather than choice,” “cannot be Lehmiller, 2019).
changed by any treatment” (Institute for Sexology and Sexual Marshall (2008) assumes that pedophilia is not immutable
Medicine of the Charité, 2013, pp. 65–66). Going further, it because after treatment phallometric data show a reduction of
means “the recognition of reality [. . .], one’s way of easing the arousal to children and an increase of arousal to adults in
pain, when realizing that things are unchangeable [. . .] [, and] “quite deviant child molesters” (Marshall, 2008, p. 42). Marshall
abandonment of the wish to change given reality” (Institute for et al. (2009) describe behavioral procedures for modifying
Sexology and Sexual Medicine of the Charité, 2013, p. 67). Thus, sexual interests, i.e., different aversion and masturbatory-based
it goes along with an ego syntonic concept of sexual preference techniques, and the evidence base for these procedures. Marshall
which allows therapy right away to focus on aspects other than et al. (2011) opt for identifying “individualized appropriate (i.e.,
sexual preference (Institute for Sexology and Sexual Medicine of non-deviant) sexual scripts that can serve as both templates for
the Charité, 2013). actual sexual relations and as images for masturbation activities”
Some therapists and scientists argue that “acceptance” of and refer to the fact that “[U]unreinforced habitual behavior (in
pedophilic interest is needed for individuals with sexual interest this case, deviant sexual interests) has been shown to extinguish
in minors (SIM) for dealing in a responsible way with their sexual such habits” (Marshall et al., 2011, p. 152). Recent research on
interest and thus preventing sexual abuse (Ahlers et al., 2008; age of onset of sexual interest in children with individuals from
Institute for Sexology and Sexual Medicine of the Charité, 2013). different contexts (with and without treatment, explicitly and not
Other therapists and scientists, however, argue that – even in explicitly advocating against acting on sexual interest in children)
some persons, and maybe more often in men with non-exclusive shows broad ranges from 6 to 44 years and 7 to 66 years, with
pedophilia (Tozdan and Briken, 2019) – SIM might change over a mean value of 17 and 20 years (Tozdan and Briken, 2015b,
time and that “acceptance” might run counter to prevention goals 2019). It also shows that a later age of onset is associated with
because motivation to change as well the specific self-efficacy more perceived flexibility of sexual interest in children, that
for modifying sexual interest in children may be reduced by it more perceived flexibility is related to more motivation to change
(Tozdan and Briken, 2015a; Fedoroff, 2020). sexual interest in children (Tozdan and Briken, 2019), and that
Seto (2012) argues that sexual orientation is characterized an increasing specific self-efficacy for modifying a sexual interest
by an age of onset before the beginning of puberty and by in children is related to a decreasing sexual interest in children
stability over time, and that pedophilia is similar in these respects. in a considerable number of individuals (Tozdan et al., 2018b)1 .
He refers to findings on identified and self-identified child Researchers who assume that pedophilia is mutable usually
sexual abuse offenders with pedophilia of whom a considerable conclude that therapeutic interventions should target patients’
proportion report an age of onset of sexual interest in children specific self-efficacy (Briken et al., 2014; Tozdan and Briken,
before adulthood (e.g., Li, 1991; Marshall et al., 1991; Freund 2015a; Tozdan et al., 2018a), empower them (Fedoroff, 2018),
and Kuban, 1993). Furthermore, he cites studies which suggest and work on “relationship skills and healthy sexuality [. . .], self-
that pedophilia is predictive for future sexual behavior involving esteem, empathy, prosocial sexual attitudes, and coping skills”
children even one to three decades later (Hanson et al., 1993),
1
and that treatment-related decrease of sexual arousal related The findings mentioned were criticized as insufficient because of still unclear
to children does not reduce recidivism rates (Rice et al., definitions (e.g., defining pedohebephilic interests as a SIM who are at least
five years younger than oneself vs. defining pedohebephilic interest on the basis
1991), i.e., does not persist (Seto, 2012). Due to these findings, of a developmental approach involving biological, social, psychological, and
he makes the case for “conceptualizing pedophilia as a type experiential aspects that started before experiencing a SIM for the first time), not
of sexual orientation in males” (Seto, 2012, p. 231). Cantor enough consideration of developmental processes (e.g., transitional processes, the
fact that age of onset is maybe a part of ongoing developmental processes instead
(2018) also takes the position that pedophilia is immutable.
of a discrete event) and not sufficiently precise operationalization of age of onset
He claims that lines of indirect evidence referring to the (i.e., asking about age of first SIM might be misunderstood because almost every
following “all have much more parsimonious and mundane person is attracted to minors at some point in life) (McPhail, 2018).

(Marshall et al., 2011, p. 152). Briken et al. (2018) argue that, MATERIALS AND METHODS
under the assumption of mutability of deviant sexual interests
in at least a part of patients, therapists should clarify what their Participants
patients’ motivation and goal in treatment is. As many patients do This exploratory pilot study included 84 adult men with SIM
not have a desire for change, therapists should be open to adapt who underwent initial diagnostic procedures between autumn
treatment to assignment, risk of (re-)offending, and exclusiveness 2011 and autumn 2019, gave their informed consent, and started
of SIM (Briken et al., 2018). treatment at the Institute for Sex Research, Sexual Medicine
In summary, at the present stage of research, one can assume and Forensic Psychiatry in Hamburg. Thirteen men who had
that there are patients in which SIM might change over time or not given their informed consent were not included. The study
not, depending on, for example, flexibility, exclusivity, and age was approved by the Ethics Committee of the Chamber of
of onset of SIM (Tozdan and Briken, 2019). Hence, it can be Psychotherapists Hamburg (09/2019-PTK-HH, 02/2015-PTK-
expected that acceptance is a differentiated construct, too. HH). Data were prepared for analysis by two researchers (UL
and ST) working at the research unit of the institute. Their work
was independent of the processes of treatment indication and
STUDY AIM psychotherapeutic care.
All of the participants fulfilled the preconditions for receiving
We can safely assume that a therapist’s attitude concerning the treatment in the program offered by the Prevention Network
necessity of accepting SIM is considered rather relevant for the “Kein Täter werden” (see Table 1 one for demographic
therapeutic process. Indeed, it already has been shown that there characteristics of the participants). These are:
is a relation between therapists’ attitude toward the immutability
of SIM and their patients’ self-efficacy to change their SIM • not (yet) having offended and/or never having consumed
(Tozdan et al., 2018a). That means, the more therapists are child sexual abuse images, though fearing doing so, or
convinced that SIM is mutable, the more their patients believe • already having offended and/or having consumed child
they can change it and vice versa. It is lively debated if patients sexual abuse images, but not being known to the legal
should be told that SIM can change, if patients should be told system, or
that SIM is unchangeable (Cantor, 2018; Cantor and Fedoroff, • previously having been charged with and/or found guilty
2018; Fedoroff, 2018), or if this question can only be answered of relevant offenses and having fully served any sentence
in the course of the therapeutic process, because there are very received as a result, and fearing committing further offenses
different courses of SIM. (Netzwerk “Kein Täter werden”, 2018).
What we do not know is if the acceptance of SIM is
related to patients’ motivation to control sexual urges, to Treatment involved 90 min of group therapy led by two
change their SIM, or to behavioral treatment outcomes, such as group therapists weekly, or individual therapy sessions every
consumption of sexual imagery of minors and/or sexual abuse 1 or 2 weeks. Every individual had to participate in an initial
of minors. Furthermore, it has not yet been studied if there is diagnostic procedure that comprised diagnostic interviews, a risk
a difference between acceptance of SIM and behavior outcomes assessment, and a battery of self-report questionnaires. Hereafter,
that correspond to pedophilic or to hebephilic interests. We also every participant was introduced to the therapeutic team by
do not know if acceptance of SIM is related to dynamic risk the therapist who conducted the initial diagnostic procedure.
factors which are linked with sexual self-regulation (Hanson et al., Referral for group vs. individual therapy was debated and decided
2007). Moreover, it has not yet been clarified if acceptance of SIM within the whole team, i.e., medical doctors and psychologists
is associated with sexual desire. Furthermore, we do not know [see Lampalzer et al. (2020) for more detailed information on
if acceptance of SIM changes in the course of treatment, i.e., is indicators for group vs. individual treatment in this sample].
affected by treatment. The treatment program at the Institute for Sex Research,
Therefore, the purpose of this exploratory pilot study is to Sexual Medicine and Forensic Psychiatry is based on the
investigate the following research questions: risk-need-responsivity model (Andrews et al., 1990) since, in
addition to reducing possible distress from SIM, its main
(1) Is there a relationship between the acceptance of SIM and objective is to prevent sexual abuse of children and the
dynamic risk factors? use of abusive images. The risk principle determines therapy
(2) (a) Is there a relationship between the acceptance of intensity. According to the need principle therapy focuses
SIM and the frequency/intensity of the use of child on the three most important dynamic risk factors that are
abuse/exploitation material? (b) Is there a relationship related to the individual’s modifiable risk of (re-)offending,
between the acceptance of SIM and the frequency/intensity e.g., in the realm of intimacy deficits or poor self-regulation.
of the use of adolescent abuse/exploitation material? With regard to the responsivity principle referral to group vs.
(3) Is there a relationship between the acceptance of SIM and individual treatment, therapeutic technique, and decision for
the frequency/intensity of sexual desire/behavior toward psychiatric treatment or medication in addition to psychotherapy
minors? are considered important. In the initial phase of treatment
(4) Does the level of acceptance of SIM change in the course of motivation and aims are clarified and biography work is
treatment? done. In the intermediate phase risk factors and behavioral

TABLE 1 | Sample characteristics for the total sample (N = 79) when undergoing After their last treatment session participants normally, i.e.,
initial diagnostic procedure.
if they were willing to do so, underwent a final diagnostic
Variables Total (N = 79, 100%) procedure which consisted, except for some updates, of the same
questionnaires as the initial diagnostic procedure.
Na %b Five participants, who were included in the present study,
Education level had not filled in the Inventory of the Acceptance of Sexual
Less than 10 years 13 16.5 Inclination (IASI rev, Mundt et al., 2011). For this reason, they
More than 10 years 66 83.5 were excluded from the analysis. The final sample consisted of 79
Employed participants. Their age ranged from 19 to 61 years (M = 35.99,
Yes 63 79.7 SD = 11.25). One participant did not indicate his age. Twenty-
No 16 20.3 four (30%) participants were still in treatment, and 55 (70%)
Relationship status had partly or fully completed the treatment program. Of these
In a relationship 39 49.4 participants who had partly or fully completed the treatment
Currently single 40 50.6 program 35 (64%) had completed the final diagnostic procedure,
Living alone including IASI rev. Only these 35 men could be included into pre-
Yes 38 48.1 post comparison analysis. Treatment duration ranged from 7 to
No 41 51.9
67 months (M = 30.66, SD = 15.39).
Own children
Yes 17 21.5
No 62 78.5 Measures
Self-reported exclusiveness (Interest is . . .) Inventory of the Acceptance of Sexual Inclination
. . . exclusively in children 11 13.9 The IASI rev is an unpublished self-report questionnaire designed
. . . not exclusively in children 67 84.8 to assess the extent of acceptance or integration of a sexual
. . . not specified 1 1.3 inclination into the individual self-concept (Ahlers et al., 2008).
Self-reported age group attracted to Sexual inclination is understood as the third of three axes of
Prepubertal (pedophile) 1 1.3 sexual preference2 : “our sexual inclination toward a (i) preferred
Pubertal (hebephile) 3 3.8 specific type of sexual partner and (ii) a preferred specific mode
Prepubertal and pubertal (pedophile and hebephile) 8 10.1 of sexual activity” (Schaefer and Ahlers, 2018, p. 88). According
Prepubertal and adult (pedophile and teleiophile) 4 5.1 to Schaefer and Ahlers, it “resembles the current definition
Pubertal and adult (hebephile and teleiophile) 27 34.2
of paraphilias in the DSM-5” which refers to erotic activities
Prepubertal, pubertal and adult (pedophile, hebephile, 35 44.3
and teleiophile)
(modes) such as spanking and whipping, and erotic targets
Not specified 1 1.3
(types), such as children, corpses or inanimate objects (Schaefer
Self-reported sexual orientation and Ahlers, 2018, p. 89). The IASI rev is a short version of
Attracted to males 14 17.7 the Inventory of the Acceptance of Sexual Preference (“Inventar
Attracted to females 48 60.8 zur Akzeptanz der sexuellen Präferenz”; cf. Ahlers et al., 2008)
Attracted to both sexes 16 20.3 which has four subscales: Attitude (subjective attitude toward the
Not specified 1 1.3 acceptance of one’s own sexual preference, extent to which the
Self-reported prior lifetime sexual offensesc attitude can become relevant for behavior), Perceived Acceptance
Non-offending 6 7.6 (extent of real acceptance), Emotion (emotional processing of
Child sexual abuse only 7 8.9 one’s own sexual preference), and Fantasy and Control (handling
Child pornography use only 43 54.4 fantasies and needs that correspond to the sexual preference).
Mixed offenses 23 29.1 The IASI rev has 15 items that are answered on a 5-point
Previously known to justicec scale (e.g., “My sexual fantasies scare me.”; “I am aware of my
Child pornography offenses 11 13.9 sexual inclination.”) (see Appendix 1 for an English translation
Child sexual abuse offenses 5 6.3 including all items). Total scores range from 15 to 75, with higher
Child pornography and child sexual abuse offenses 2 2.5
values indicating a greater acceptance. The IASI rev has not been
Not previously known to justice 61 77.2
validated yet. For the present pilot study the IASI rev was found
a Absolute share in the sample.
b Percentage
to be highly reliable (15 items; Cronbach’s α = 0.88). The items
share in the sample.
c Status when entering the treatment program. of the IASI rev do not specifically refer to SIM. However, in
the context of the battery of questionnaires that was filled in by
the participants of the present study, it is rather unlikely that
change are focused, particularly sexual self-regulation, emotional sexual inclination was not interpreted as SIM. Because the battery
congruence with children, awareness and handling of risk is explicitly designated for individuals with SIM and all of the
situations, abuse related attitudes, hypersexuality and sexual
urges, increase in interpersonal abilities, improvement of coping 2
The other two axes are: “(a) our sexual orientation towards the preferred sex of
strategies, and empathy. The final phase concentrates on
the desired sexual partner [. . .], (b) our sexual alignment towards the preferred
preventing relapses, considering support groups, and developing stage of physical development of the sexual partner’s body” (Schaefer and Ahlers,
future plans (Briken et al., 2018). 2018, p. 88).

participants turned to the network site specifically due to their adults sexually interacting with adolescents is used, as well as
SIM. In this study, we used the IASI rev total score for analyzing a subset of items assessing the frequency of desire for sexual
“acceptance” as a multidimensional construct and Item 5 of IASI activities, and actual sexual activities with minors. In order to
rev (“My inclination is an integral part of my personality.”) for reduce the number of these items and generate six adequate items
focusing even more on the aspect of (im-)mutability of SIM. This for the analysis, we amalgamated the relevant items via taking the
item of IASI rev (hereinafter also referred to as “Integral Part score of the highest frequency of the relevant items. The six items
Item”) most closely corresponds to this aspect. are (see Appendix 2 for the specific questions of three different
batteries):
STABLE-2007
The STABLE-2007 (Hanson et al., 2007; Matthes and (1) Frequency of Use of Legal Imagery of Children;
Rettenberger, 2008) is a 13-item risk assessment tool to (2) Frequency of Use of Legal Imagery of Adolescents;
measure dynamic risk for contact sexual recidivism among adult (3) Frequency of Use of Illegal Child Sexual Abuse Imagery;
males who have been charged with a sexual offense. The 13 (4) Frequency of Use of Illegal Adolescent Sexual Abuse
risk factors are evaluated by third party. They have been shown Imagery;
to be associated with sexual recidivism and are systematized (5) Frequency of Desire for Sexual Activities with Minors;
into five sections: Significant Social Influences, Intimacy Deficits (6) Frequency of Sexual Activities with Minors.
(capacity for relationship stability, emotional identification with Via these six items common distinctions in the literature are
children, hostility toward women, general social rejection, lack represented: (1) legal vs. illegal imagery, (2) pedophilic interests
of concern for others), General Self-Regulation (impulsivity, vs. hebephilic interests, (3) use of child abuse/exploitation
poor problem-solving skills, negative emotionality), Sexual material (hands-off) vs. (drive to) child sexual abuse (hands-on).
Self-Regulation (sex drive and preoccupations, sex as coping, Frequency is rated on a 5-point-Likert scale, with the following
deviant sexual preference), and Cooperation with Supervision. answer options: 1 = “never,” 2 = “few times,” 3 = “monthly,”
All of the items are scored on a 3-point scale. Total scores range 4 = “weekly,” 5 = “daily.” For amalgamation, the score of the
from 0 to 26 (exception: emotional identification is not scored highest frequency of the relevant items was taken because this
for offenders not having a child as a victim so that for them total corresponds to the answer the participant would have given if he
scores only range to 24), with higher values indicating a higher had answered to the amalgamated item. The distinction between
dynamic risk of recidivism. The STABLE-2007 has proven very legal and illegal imagery follows the COPINE scale (Quayle,
good interrater reliability for the English version [ICC = 0.79 2008), with category 1–3 classified as legal imagery and category
(Hanson et al., 2007)] and for the German version [ICC = 0.90 for 4–10 as illegal imagery. Table 2 shows that inter-item correlations
a population of sex offenders with 50.9% being child molesters were significant and between 0.23 and 0.68, except for item 6 that
(Eher et al., 2012); ICC = 0.90 for a population of child molesters was not significantly correlated to the items 2, 3, and 4.
(Rettenberger et al., 2011)], too. It has also demonstrated good
predictive validity for recidivism [AUC = 0.67–0.71 for sexual, Statistical Analysis
violent and general recidivism (Eher et al., 2012)]. As mentioned As the battery of questionnaires was revised with regard to the
above, the STABLE-2007 is designed for contact sexual offenders current state of research in the course of data collection, not
who have been charged with a sexual offense. The dynamic risk all questionnaires used in the present study were filled in by all
factors of the STABLE-2007 are not validated for recidivism participants. As the STABLE-2007 was only added to the battery
relating to child/adolescent sexual abuse imagery and not for of questionnaires in 2014, it was filled in for many participants
individuals with SIM without or with undetected offences.
Nevertheless, the STABLE-2007 is used in this exploratory pilot
study because it measures stable dynamic risk factors and there TABLE 2 | Results of Pearson correlation tests for inter-item correlations of the
is no established tool available yet for this specific group with items for frequency of the use of child abuse/exploitation material and frequency
SIM. However, data are analyzed for the group of patients in of sexual desire/behavior toward children inter-item correlation.
this sample who reported child/adolescent sexual abuse in their Items 1 2 3 4 5 6
past as well as for the whole group. Because one of the tool’s
developers, Karl Hanson, explicitly recommended to not use it (1) Use of Legal Imagery of 1
Children
for internet offenders who only offended with indecent images of
(2) Use of Legal Imagery of 0.53** 1
children, which might apply to the majority of the sample, and
Adolescents
not with identifiable victims (Webb, 2018, p. 107).
(3) Use of Illegal Child Sexual 0.66** 0.42** 1
Abuse Imagery
Items Assessing the Frequency of the Use of Child (4) Use of Illegal Adolescent 0.33** 0.68** 0.62** 1
Abuse/Exploitation Material and Frequency of Sexual Sexual Abuse Imagery
Desire/Behavior Toward Children (5) Desire for Sexual Activities 0.45** 0.26* 0.45** 0.29* 1
A subset of items of questionnaires of the initial diagnostic with Minors
procedure which assess the frequency of consumption of (sexual) (6) Sexual Activities with Minors 0.23** −0.06 0.10 −0.12 0.32** 1
imagery of children, adolescents, adolescents sexually interacting *Correlation is significant at the 0.05 level.
with children, adults sexually interacting with children, and **Correlation is significant at the 0.01 level.

by the therapists at a later time during the course treatment. score (rs = 0.01, n = 16, p = 0.961), and no significant correlation
Only those participants (n = 45) were included in the statistical between Item 5 IASI rev score and STABLE-2007 total score
analysis of correlations regarding the STABLE-2007 for whom the (rs = –0.18, n = 16, p = 0.509), either (Table 3).
STABLE-2007 was completed not more than 6 months after the
participant had completed the self-report questionnaires of the Relationship Between Inventory of the
battery, assuming that no substantial change in dynamic risk of
Acceptance of Sexual Inclination and
sexual recidivism has taken place during this time. In a second
analysis of correlations regarding the STABLE-2007, only those Frequency/Intensity of the Use of
(n = 16) were included who reported child/adolescent sexual Material of Abuse/Exploitation of Minors
abuse in their past (see section “STABLE-2007” for explanation). Results of the Spearman’s correlation indicated that there was a
First, the relation between acceptance of SIM and dynamic medium3 , positive correlation between IASI rev total score and
risk for contact sexual recidivism before treatment, i.e., between Frequency of Use of Legal Imagery of Children score (rs = 0.41,
IASI rev total score/Item 5 IASI rev score and STABLE-2007 n = 72, p ≤ 0.001), and no significant correlation between
total score of the initial diagnostic procedure, was analyzed IASI rev total score and Frequency of Use of Legal Imagery of
using the Spearman’s correlation coefficient because variables Adolescents score (rs = 0.22, n = 72, p = 0.058). There was
were ordinally scaled (Upton and Cook, 2014). Second, the also a medium, positive correlation between Item 5 IASI rev
relation between acceptance of SIM and frequency/intensity of score and Frequency of Use of Legal Imagery of Children score
the use of material of abuse/exploitation of minors, i.e., IASI rev (rs = 0.32, n = 75, p = 0.005), and no significant correlation
total score/Item 5 IASI rev score and the items of the initial between Item 5 IASI rev score and Frequency of Use of Legal
diagnostic procedure assessing the frequency of the use of child Imagery of Adolescents score (rs = 0.08, n = 75, p = 0.474),
abuse/exploitation material, was analyzed using the Spearman’s either. The correlations remained statistically significant after
correlation coefficient because variables were ordinally scaled, Benjamini–Hochberg correction (Table 3).
too (Upton and Cook, 2014). Third, a Spearman’s correlation Moreover, results of the Spearman’s correlation indicated
was run to determine the relationship between acceptance of that there was a small, positive correlation between IASI rev
SIM and frequency/intensity of sexual desire/behavior toward total score and Frequency of Use of Illegal Child Sexual Abuse
minors, i.e., IASI rev total score/Item 5 IASI rev score and Imagery score (rs = 0.27, n = 70, p = 0.026), no significant
items assessing the frequency of sexual desire/behavior toward correlation between IASI rev total score and Frequency of Use
children, because variables were also ordinally scaled (Upton and of Illegal Adolescent Sexual Abuse Imagery score (rs = 0.11,
Cook, 2014). Fourth, a Wilcoxon signed-rank-test was performed n = 71, p = 0.367), no significant correlation between Item 5 IASI
to compare acceptance of SIM, i.e., IASI rev total score, between rev score and Frequency of Use of Illegal Child Sexual Abuse
initial and final diagnostic procedure because the data were not Imagery score (rs = 0.18, n = 74, p = 0.119), and no significant
normally distributed. Fifth, a paired-samples t-test was carried correlation between Item 5 IASI rev score and Frequency of Use
out to compare Item 5 IASI rev score between initial and final of Illegal Adolescent Sexual Abuse Imagery score (rs = –0.11,
diagnostic procedure because the data were normally distributed n = 75, p = 0.344). The correlation between IASI rev total score
(Kim, 2015). In the final diagnostic procedure, the IASI rev was and Frequency of Use of Illegal Child Sexual Abuse Imagery
completed by 35 participants. Significance was set at a value score was not statistically significant after Benjamini–Hochberg
less than 0.05. All statistical analyses were conducted using correction (Table 3).
SPSS (V 24) (IBM SPSS Statistics, IBM Corporation, Armonk,
NY, United States). Relationship Between Inventory of the
Acceptance of Sexual Inclination and
Frequency/Intensity of Sexual
RESULTS Tendencies/Behavior Toward Minors
Results of the Spearman’s correlation indicated that there was
Relationship Between Inventory of the a small, positive correlation between IASI rev total score and
Acceptance of Sexual Inclination and Frequency of Desire for Sexual Activities with Minors score
STABLE-2007 (rs = 0.23, n = 75, p = 0.044). Results of the Spearman correlation
For the whole group, results of the Spearman’s correlation also indicated that there was a medium, positive correlation
indicated that there was no significant correlation between IASI between IASI rev total score and Frequency of Sexual Activities
rev total score and STABLE-2007 total score (rs = –0.22, n = 43, with Minors score (rs = 0.30, n = 75, p = 0.008). There was
p = 0.166), and that there was no significant correlation between no significant correlation between Item 5 IASI rev score and
Item 5 IASI rev score and STABLE-2007 total score (rs = 0.07, Frequency of Desire for Sexual Activities with Minors score
n = 45, p = 0.642) (Table 3). (rs = 0.18, n = 79, p = 0.120), and a small, positive correlation
Similarly, for the group of patients who reported between Item 5 IASI rev score and Frequency of Sexual Activities
child/adolescent sexual abuse in their past, results of the
Spearman’s correlation indicated that there was no significant 3
The Spearman’s correlation coefficients are interpreted according to
correlation between IASI rev total score and STABLE-2007 total Cohen (1988, 1992).

TABLE 3 | Descriptive statistics and correlations between IASI rev (Item 5 IASI rev) and STABLE-2007, Frequency of Use of Legal Imagery of Children, Frequency of Use
of Legal Imagery of Adolescents, Frequency of Use of Illegal Child Sexual Abuse Imagery, Frequency of Use of Illegal Adolescent Sexual Abuse Imagery, Frequency of
Desire for Sexual Activities with Minors, and Frequency of Sexual Activities with Minors.
Variable M SD Mdn Range ra pa FDR adjusted na

pBH a,b
(1) IASI rev 51.20 11.03 51 28–73 – – –

(a) Item 5 IASI rev 3.30 1.21 3 1–5 – – –
(2) STABLE-2007
(a) Total sample 9.49 2.94 9 4–16 –0.215 0.166 0.266 43
0.071 0.642 0.685 45
(b) Participants with child sexual abuse in the past 10.75 2.72 11 5–15 0.013 0.961 0.961 16
–0.178 0.509 0.582 16
(3) Frequency of Use of Legal Imagery of Children 2.75 1.53 3 1–5 0.413* <0.001 <0.001 72
0.323* 0.005 0.040 75
(4) Frequency of Use of Legal Imagery of Adolescents 2.35 1.43 2 1–5 0.224 0.058 0.133 72
0.084 0.474 0.582 75
(5) Frequency of Use of Illegal Child Sexual Abuse Imagery 2.86 1.46 3 1–5 0.265 0.026 0.083 70
0.183 0.119 0.213 74
(6) Frequency of Use of Illegal Adolescent Sexual Abuse Imagery 2.68 1.42 3 1–5 0.109 0.367 0.489 71
–0.111 0.344 0.489 75
(7) Frequency of Desire for Sexual Activities with Minors 3.09 1.46 4 1–5 0.234 0.044 0.117 75
0.176 0.120 0.213 79
(8) Frequency of Sexual Activities with Minors 1.57 1.17 1 1–5 0.304* 0.008 0.043 75
0.264 0.019 0.076 79
*pBH < 0.05.

a The first line refers to IASI rev, the second line refers to Item 5 of IASI rev.
b Adjusted p-value using the Benjamini–Hochberg procedure (Benjamini and Hochberg, 1995; Hemmerich, 2016).
IASI, Inventory of the Acceptance of Sexual Inclination.
with Minors score (rs = 0.26, n = 79, p = 0.019). After Benjamini– justice) than forensic samples of contact sexual offenders with
Hochberg correction, only the correlation between IASI rev total a SIM who have been charged with a sexual offense. Maybe
score and Frequency of Sexual Activities with Minors score the STABLE-2007 is not an adequate instrument for measuring
remained statistically significant (Table 3). dynamic risk for contact sexual reoffending in this sample.
However, there was no correlation between IASI rev total score
Pre–post Comparison of Inventory of the or Integral Part Item score and STABLE-2007 score, either. This
Acceptance of Sexual Inclination might also be due to a lack of statistical power for finding
Results of the Wilcoxon signed-rank tests indicated no statistical small effects. Power analyses using G∗ Power (Faul et al., 2007)
difference between IASI rev total score before (Mdn = 54) and indicated that, with 80% power and α = 0.05, a sample size of
after (partial) completion of treatment (Mdn = 55), T = 261, 129 would be required to detect an effect of r = 0.215 that was
z = –0.26, p = 0.799, r = –0.03 (Table 4). A paired-samples t-test revealed for the whole sample in the present study, and a sample
indicated no statistical difference between the score of Item 5 size of 1,222 for an effect of r = 0.071 that was revealed for the
of IASI rev before beginning treatment (M = 3.51, SD = 1.07) subsample with child sexual abuse in the past.
and the score of Item 5 of IASI rev after (partial) completion Results of the current exploratory pilot study indicate a
of treatment (M = 3.57, SD = 1.12), t(34) = –0.26, p = 0.797, medium, positive correlation between acceptance of SIM and
d = –0.05, either (Table 4). the frequency of the use of legal imagery of children, and a
small, positive correlation between acceptance of SIM and the
frequency of the use of illegal child sexual abuse imagery. This
DISCUSSION means, the more participants report to accept their SIM, the
more they also report to use legal imagery of children as well
General Discussion as illegal child sexual abuse imagery, or vice versa. However, the
This exploratory pilot study investigated if acceptance of SIM findings reveal no correlation between acceptance of SIM and
is associated with pedophilia associated urges and behaviors. the frequency of use of legal imagery of adolescents, and no
Acceptance of SIM as measured by IASI rev total score and correlation between acceptance of SIM and the frequency of the
the score of the item “My inclination is an integral part of my use of illegal adolescent sexual abuse imagery. Furthermore, they
personality” were not related to dynamic risk factors for contact show a small, positive correlation between acceptance of SIM and
sexual reoffending as measured by STABLE-2007. It is possible the reported frequency of desire for sexual activities with minors,
that this is due to the sample which might be characterized by and a medium positive correlation between acceptance of SIM
other dynamic risk factors (77% were not previously known to and the reported frequency of sexual activities with minors.

TABLE 4 | Results of Wilcoxon signed-rank test and paired-samples t-test and descriptive statistics for IASI rev total score and Item 5 IASI rev score (n = 35).
Outcome Pre-test Post-test 95% CI for

mean
M SD Mdn Range n M SD Mdn Range n difference Z t p FDR adjusted Effect Effect
pBH a size r size d
IASI rev 52.50 10.44 54 28–73 34 52.43 10.32 55 25–69 35 –0.26 0.799 0.799 –0.03
Item 5 IASI 3.51 1.07 4 1–5 35 3.57 1.12 4 1–5 35 –0.51, –0.26 0.797 0.799 –0.05
rev 0.39
*pBH < 0.05.

a Adjusted p-value using the Benjamini–Hochberg procedure (Benjamini and Hochberg, 1995; Hemmerich, 2016).
IASI, Inventory of the Acceptance of Sexual Inclination.
The results may suggest that more acceptance of SIM might assumption of SIM as an integral part of one’s personality and
be associated with more awareness of a SIM, the desire might be the frequency of the use of legal child sexual abuse imagery, and
more present and associated with more use of legal imagery and the correlation between acceptance of SIM and the frequency of
even more proneness to sexual abuse. Less acceptance of SIM, sexual activities with minors remained statistically significant.
however, might be associated with more motivation to work on Post hoc power analyses that were conducted using the
diminishing pedophilia associated urges and behaviors and thus program G∗ Power (Faul et al., 2007) indicated that the statistical
make individuals reduce the focus on desire for sexual activities power for the correlation analyses of this study (assuming n = 70;
with children, use less legal and illegal material and prevent see Table 2) was 13% for detecting a small effect of r = 0.1, 80%
themselves from committing child sexual abuse. Of course, as we for detecting a medium effect of r = 0.32 and more than 99%
solely conducted correlation analyses, we are not able to make any for detecting a large effect of r = 0.5 (according to Cohen, 1988,
statement about causality. 1992), with α = 0.05. Thus, there was more than sufficient power
It is surprising that the correlations that were found for legal (i.e., 80%) at the large effect size level, quite enough power at the
and illegal imagery of children were not found for legal and medium effect size level and less than sufficient statistical power
illegal imagery of adolescents. Research suggests that pedophilic, at the effect size level of less than r = 0.32.
hebephilic and teleiophilic individuals are different from each The present study also investigated whether acceptance of SIM
other, but there is still a lack of research in this field (Sea changed in the course of treatment. Results indicate no difference
and Beauregard, 2018). In the present study, the majority of between before beginning treatment and (partial) completion of
participants (44%) were pedophilic, hebephilic and teleiophilic, treatment. Maybe for a change of acceptance of SIM it would
or hebephilic and teleiophilic (34%) at the same time. Obviously, have to be directly targeted in treatment. However, treatment
these sexual interests cannot be seen as totally distinct from has a particular focus on risk factors (cf. Institute for Sexology
each other. According to the findings of this exploratory pilot and Sexual Medicine of the Charité, 2013; Briken et al., 2018;
study, acceptance of SIM might be different in the context of a Netzwerk “Kein Täter werden”, 2018) and acceptance of SIM
pedophilic interest than in the context of a hebephilic interest. might not have appeared to be a relevant risk factor. Another
Maybe acceptance of a pedophilic interest has a deeper impact reason why acceptance of SIM did not change might be that the
on awareness and the frequency of corresponding sexual desire period under study was not sufficiently long enough to observe a
and sexual behavior than acceptance of a hebephilic interest (cf. change in acceptance of SIM. Furthermore, insufficient statistical
Prentky and Barbaree, 2011; Calkins Mercado and Beattey, 2012). power because of the small sample size in the present study
Additionally, results of the current exploratory pilot study (n = 35) may have played a role in limiting the significance of
indicate a medium, positive correlation between the assumption the pre-post comparison conducted. Effect sizes from a study
of SIM as an integral part of one’s personality and the frequency by Engel et al. (2018) with a sample of the Prevention network
of the use of legal imagery of children, and a small, positive “Kein Täter werden” were between d = –0.14 and d = –0.58 for
correlation between this assumption and the frequency of comparisons of treatment group (n = 35) before and after therapy
sexual activities with minors, but no correlation between this regarding several measures. This supports the expectation for
assumption and any of the other frequency measures used. This medium effects (according to Cohen, 1988, 1992). For the present
means, the more men regard their pedophilic interests as an study, a power analysis using the program G∗ Power (Faul et al.,
integral part of their personality, the more they sexualize non- 2007) indicated that a total sample of 35 participants would
sexual representations of children and commit hands-on child be needed to detect a medium effect (d = 0.5) with 80%
sexual abuse, or vice versa, the more men sexualize non-sexual power and α = 0.05. This is equivalent to the sample size of
representations of children and commit hands-on child sexual the present study.
abuse, the more they regard their pedophilic interests as an
integral part of their personality. Limitations and Future Studies
It has to be noted that after correction for multiple testing, only The generalizability of the present results is restricted due to
the correlation between acceptance of SIM and the frequency of the sample size of only 79 participants in the whole sample and
the use of legal imagery of children, the correlation between the only 35 participants in the sample for pre-post comparison. With

this sample size, especially with regard to pre-post comparison, analyses which allow the examination of causality (e.g., crossed-
this exploratory pilot study was underpowered for small effects. lagged panel analyses; Frees, 2004).
Therefore, the results of the present study need to be replicated The present study focused on the question if acceptance of
with studies that include larger samples that would guarantee a SIM is related to essential treatment outcomes. The outcome
sufficient statistical power. Generalizability is also limited because measures of the present study, except of “Frequency of Desire for
of the particular characteristics and institutional context of the Sexual Activities with Minors,” mainly focus on the frequency of
“Kein Täter Werden” network site in Hamburg. Furthermore, behavior. Further studies could also include outcome measures
13 men could not be included in the study because they that focus on the strength of SIM in itself (see Carvalho et al.,
had not provided informed consent. Therefore, these data, 2020 for different measures that are available), thus highlighting
and maybe specific characteristics, are not represented in the that not only behavioral changes are notable treatment goals but
findings of this study. also change of SIM in itself, in general or via a partial shift from
Concerning validity, our results are limited because almost pedophilic to teleiophilic interests, for example. Accordingly, it
only self-report measurements with forced-choice categories would be worth studying if the level of SIM is associated with
were used. Forced-choice categories may simplify answers and/or acceptance of SIM.
distort information because of the particular choice sets given. With larger samples, in the next stage of research, subgroups
The patients’ self-report was not validated by objective measures. of men with SIM should be studied in more detail with
Hence, it cannot be ruled out that an effect of social desirability regard to acceptance of SIM in order to be able to understand
distorted our data. Only the STABLE-2007 items are assessed their characteristics and personalities better, differentiate better
by third party. For a part of the patients whose STABLE-2007 between them, and adapt treatment accordingly. There might
date were analyzed the STABLE-2007 was only filled in for be differences between the subgroups of non-offenders, mixed
a period of a few weeks (for 25 patients more than 6 weeks offenders, hands-on sexual offenders, and offenders without
later) up to 6 months after completion of the initial diagnostic hands-on offenses but consumption of material depicting the
procedure, because the STABLE-2007 was not part of the battery sexual exploitation of minors. These subgroups were not
of questionnaires right from the beginning of data collection. It differentiated in this study. Previous research indicates that these
cannot be ruled out that this retrospective completion may have subgroups are, among other things, distinctive from each other
been biased. However, as mentioned above, we assume that no regarding indicators of antisociality (Babchishin et al., 2015),
substantial change in dynamic risk of sexual recidivism takes criminal history (Long et al., 2013), offense supportive attitudes
place during this time. Moreover, the self-report questionnaires (Jahnke et al., 2015), stability of life factors, and substance abuse
analyzed in this exploratory pilot study are not validated, yet. problems (Ly et al., 2018). Maybe they also differ from each
Thus, as mentioned before, the findings can only be seen as other regarding acceptance of SIM. Previous findings suggest
preliminary results. that offenders without hands-on offenses but consumption of
Until now, acceptance of SIM and its effects are not well material depicting the sexual exploitation of minors have greater
researched, yet. There is a need for a validation study of the levels of SIM than hands-on sexual offenders in the sense of
IASI rev. Furthermore, at this state of research, case studies greater sexual deviancy (even if going along with greater barriers
and qualitative research would help to understand this construct to hands-on offending), greater likelihood to have problems
and its impact on individuals better (see Jones et al., 2020 with sexual preoccupation and sexual self-regulation (Babchishin
for qualitative research giving some insights into acceptance et al., 2015), and a stronger diagnostic indicator of pedophilia
of pedophilia as a coping strategy of individuals who identify (Seto et al., 2006). Another subgroup worth studying are webcam
as pedophilic or hebephilic and do not offend). Quantitative child sexual abuse offenders (de Tribolet-Hardy et al., 2020).
research is needed in order to study how closely acceptance Differences in acceptance of SIM between individuals who are
of SIM is related to measures of other constructs like the attracted to children, individuals who are attracted to adolescents,
Unconditional Self-Acceptance Questionnaire (Chamberlain and those who are attracted to both, and those who are also attracted
Haaga, 2001), the Emotional Processing Scale (Baker et al., 2015), to adults also need to be investigated in greater detail. Last
and/or the Thought Control Questionnaire (Wells and Davies, but not least, prior research shows that social factors, such
1994). For a broader treatment plan, going beyond directly as social support and relationship status, affect self-acceptance
focusing on the main risk factors, research on the relation (Leavy and Adams, 1986; Vincke and Bolton, 1994; Huang
of acceptance of SIM to constructs associated with emotional et al., 2020). Thus, they might as well affect acceptance of SIM.
well-being should be conducted, e.g., on how closely it is Therefore, further research should also examine relationships
related to the context of personal suffering, including ambivalent between social factors, e.g., relationship status and living alone
identity experience (Blagden et al., 2018) and stigma-related vs. not living alone, acceptance of SIM and the effects on
stress (Jahnke et al., 2015; Wagner et al., 2016; Lievesley et al., treatment outcomes.
2020; Stelzmann et al., 2020).
Since we solely calculated correlation coefficients, we are not
able to make any statement about causal relations. This means, CONCLUSION
it is not clear whether the acceptance of SIM had an impact
on the other variables, or vice versa, or whether and how they The main question of this study was if acceptance of SIM is
interact with each other. A further study should include statistical related to essential treatment outcomes. The findings, on the

one hand, indicate a positive correlation between acceptance Psychotherapists Hamburg (09/2019-PTK-HH, 02/2015-PTK-
of SIM and the use of legal imagery of children. On the HH). The patients/participants provided their written informed
other hand, they suggest that acceptance of SIM might be consent to participate in this study.
positively correlated with illegal activities, such as the frequency
of sexual activities with minors. According to this, more
acceptance might reinforce or be reinforced by legal ways of AUTHOR CONTRIBUTIONS
dealing with pedophilic interest, but it might also pave the
way for or be increased by an enhancement of illegal activities. Under the supervision of PB, UL conceptualized the study and
Hence, acceptance of SIM should be further investigated was the primary writer of the manuscript. UL and ST prepared
before specific recommendations for treatment are made. With the data for analysis. All the authors had full access to all
our current knowledge, there is a need for individualized the data in the study and take responsibility for the integrity
treatment plans allowing for a modification of SIM in some of the data and the accuracy of data analysis, were involved
patients and, in other patients, for working on acceptance in developing, editing, reviewing, and providing feedback for
of SIM. this manuscript, and have given approval of the final version
to be published.
DATA AVAILABILITY STATEMENT ACKNOWLEDGMENTS

The datasets for this study are not publicly available due to patient We would like to thank all participants for giving their informed
confidentiality and participant privacy. The computer code is consent to analyze and publish their data. We would like to
available on request to the corresponding author. thank the therapists of the Institute for Sex Research, Sexual
Medicine and Forensic Psychiatry for collecting the data in their
outpatient treatment center. Moreover, we would like to thank
ETHICS STATEMENT Samantha-Insine Schröder, M.Sc., for English proofreading, and
Ingrid Mundt, Gerard A. Schaefer, and Christoph Joseph Ahlers
The studies involving human participants were reviewed for giving approval for publishing the items of the Inventory of
and approved by Ethics Committee of the Chamber of the Acceptance of Sexual Inclination.
REFERENCES Cantor, J. M., and Fedoroff, J. P. (2018). Can pedophiles change? Response to
opening arguments and conclusions. Curr. Sex. Health Rep. 10, 213–220. doi:
Ahlers, C. J., Neutze, J., Mundt, I. A., Hupp, E., Konrad, A., Beier, K. M., 10.1007/s11930-018-0167-0
et al. (2008). Erhebungsinstrumente in der klinischen Sexualforschung und Carvalho, J., Bradford, J., Murphy, L., Briken, P., and Fedoroff, P. (2020).
der sexualmedizinischen Praxis – Teil II [Assessment instruments in clinical Measuring pedophilic interests. J. Sex. Med. 17, 378–392. doi: 10.1016/j.jsxm.
sexology sexological research – part II]. Sexuologie 15, 82–103. 2019.12.008
Andrews, D. A., Bonta, J., and Hoge, R. D. (1990). Classification for effective Chamberlain, J. M., and Haaga, D. A. F. (2001). Unconditional self- acceptance and
rehabilitation. Crim. Justice Behav. 17, 19–52. doi: 10.1177/009385489001700 responses to negative feedback. J. Ration Emot. Cogn. Behav. Ther. 19, 177–189.
1004 doi: 10.1023/A:1011141500670
Babchishin, K. M., Hanson, R. K., and VanZuylen, H. (2015). Online child Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences, 2nd Edn.
pornography offenders are different: A meta-analysis of the characteristics of New York: Academic Press.
online and offline sex offenders against children. Arch. Sex. Behav. 44, 45–66. Cohen, J. (1992). A power primer. Psychol. Bull. 112, 155–159. doi: 10.1037/0033-
doi: 10.1007/s10508-014-0270-x 2909.112.1.155
Baker, R., Thomas, P., Thomas, S., Santonastaso, M., and Corrigan, E. (2015). de Tribolet-Hardy, F., Hill, A., and Habermeyer, E. (2020). „Webcam child sexual
Emotional Processing Scale. Oxford, UK: Hogrefe. abuse“. Eine neue Facette der Begutachtung von Internetsexualdelinquenz
Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery rate: A [“Webcam child sexual abuse”. A new facet of the assessment of online sexual
practical and powerful approach to multiple testing. J. R. Stat. Soc. B Stat. delinquency]. Forens Psychiat. Psychol. Kriminol. 14, 259–269. doi: 10.1007/
Methodol. 57, 289–300. doi: 10.2307/2346101 s11757-020-00619-8
Blagden, N. J., Mann, R., Webster, S., Lee, R., and Williams, F. (2018). “It’s not Dombert, B., Schmidt, A. F., Banse, R., Briken, P., Hoyer, J., Neutze, J., et al.
something I chose you know”: Making sense of pedophiles’ sexual interest in (2016). How common is men’s self-reported sexual interest in prepubescent
children and the impact on their psychosexual identity. Sex. Abuse 30, 728–754. children? J. Sex. Res. 53, 214–223. doi: 10.1080/00224499.2015.102
doi: 10.1177/1079063217697132 0108
Briken, P., Berner, W., Flöter, A., Jückstock, V., and von Franqué, F. (2018). Eher, R., Matthes, A., Schilling, F., Haubner-McLean, T., and Rettenberger,
Prävention sexuellen Kindesmissbrauchs im Dunkelfeld – das Hamburger M. (2012). Dynamic risk assessment in sexual offenders using STABLE-
Modell [Prevention of child sexual abuse out of the criminal law context – 2000 and the STABLE-2007: An investigation of predictive and
The Hamburg Model]. Psychother. Psychosom. Med. Psychol. 68, 142–161. doi: incremental validity. Sex. Abuse 24, 5–28. doi: 10.1177/107906321140
10.1055/s-0043-121680 3164
Briken, P., Fedoroff, J. P., and Bradford, J. W. (2014). Why can’t pedophilic disorder Engel, J., Körner, M., Schuhmann, P., Krüger, T. H. C., and Hartmann, U. (2018).
remit? Arch. Sex. Behav. 43, 1237–1239. doi: 10.1007/s10508-014-0323-1 Reduction of risk factors for pedophilic sexual offending. J. Sex. Med. 15,
Calkins Mercado, C., and Beattey, R. A. (2012). Is that ‘normal’ Behavior? Judicial 1629–1637. doi: 10.1016/j.jsxm.2018.09.001
Notebook 43. Available online at: https://www.apa.org/monitor/2012/12/jn Faul, F., Erdfelder, E., Lang, A.-G., and Buchner, A. (2007). G∗ Power 3: A flexible
Cantor, J. M. (2018). Can pedophiles change? Curr. Sex. Health Rep. 10, 203–206. statistical power analysis program for the social, behavioral, and biomedical
doi: 10.1007/s11930-018-0165-2 sciences. Behav. Res. Methods 39, 175–191.

Fedoroff, J. P. (2018). Can people with pedophilia change? Yes they can! Curr. Sex. Matthes, A., and Rettenberger, M. (2008). Die Deutsche Version des Stable-2007
Health Rep. 10, 207–212. doi: 10.1007/s11930-018-0166-1 [The German version of the STABLE-2007]. Wien: Institut für Gewaltforschung
Fedoroff, J. P. (2020). “Pedophilic disorder,” in The Paraphilias. Changing Suits in und Prävention.
the Evolution of Sexual Interest Paradigms, 1 Edn, ed. J. P. Fedoroff (New York: McPhail, I. V. (2018). More on age of onset of attractions and sexual development
Oxford University Press), 109–160. in pedohebephilic individuals: Response to Tozdan and Briken (2018). Arch.
Frees, E. W. (2004). Longitudinal and panel data: Analysis and applications in the Sex. Behav. 47, 2156–2168. doi: 10.1007/s10508-018-1302-8
social sciences. Cambridge, UK: Cambridge University Press. Mundt, I. A., Schaefer, G. A., and Ahlers, C. J. (2011). Inventar zur Akzeptanz
Freund, K., and Kuban, M. (1993). Toward a testable developmental model of der sexuellen Neigung (IASN) [Inventory of the Acceptance of Sexual
pedophilia: The development of erotic age preference. Child Abuse Negl. 17, Inclination (IASI)]. Berlin: Institute for Sexology and Sexual Medicine of the
315–324. doi: 10.1016/0145-2134(93)90051-6 Charité.
Hanson, G. T., Steffy, R. A., and Gauthier, R. (1993). Long-term recidivism of Netzwerk “Kein Täter werden” (2018). lieben sie kinder mehr, als ihnen lieb ist?
child molesters. J. Consult. Clin. Psychol. 61, 646–652. doi: 10.1037/0022-006X. [Do you love children more than you would like?]. Available online at: https:
61.4.646 //www.kein-taeter-werden.de/betroffene/
Hanson, R. K., Harris, A. J. R., Scott, T. L., and Helmus, L. (2007). Assessing The Prentky, R., and Barbaree, H. (2011). Commentary: Hebephilia – a would-be
Risk of Sexual Offenders on Community Supervision: The Dynamic Supervision paraphilia caught in the twilight zone between prepubescence and adulthood.
Project. Ottawa: Public Safety Canada. J. Am. Acad. Psychiat. Law 39, 506–510.
Hemmerich, W. (2016). StatistikGuru: Rechner zur Adjustierung des α-Niveaus Rettenberger, M., Matthes, A., Schilling, F., and Eher, R. (2011). Die Validität
[StatistikGuru: Calculator for alpha level adjustment]. Available online at: https: dynamisch-veränderbarer Risikofaktoren bei der Vorhersage einschlägiger
//statistikguru.de/rechner/adjustierungdes-alphaniveaus.html. Rückfälle pädosexueller Täter: Eine Studie über Stable-2000 und Stable-2007
Huang, Y., Ruibin, W., Wu, J., Yang, Q., Zheng, S., and Wu, K. (2020). [The validity of dynamic and changeable risk factors for the prediction of sexual
Psychological resilience, self-acceptance, perceived social support and their recidivism in a sample of pedosexual offenders: A study of Stable-2000 and
associations with mental health of incarcerated offenders in China. Asian J. Stable-2007]. Forens Psychiat. Psychol. Kriminol. 5, 45–53. doi: 10.1007/s11757-
Psychiat. 52:102166. doi: 10.1016/j.ajp.2020.102166 010-0086-z
Institute for Sexology and Sexual Medicine of the Charité (2013). BEDIT – The Rice, M. E., Quinsey, V. L., and Harris, G. T. (1991). Sexual recidivism among child
Berlin Dissexuality Therapy Program. Berlin: Institute for Sexology and Sexual molesters released from a maximum security psychiatric institution. J. Consult.
Medicine of the Charité. Clin. Psychol. 59, 381–386. doi: 10.1037/0022-006x.59.3.381
Jahnke, S., Schmidt, A. F., Geradt, M., and Hoyer, J. (2015). Stigma-related stress Quayle, E. (2008). The COPINE project. Ir. Probat. J. 5, 65–83.
and its correlates among men with pedophilic sexual interests. Arch. Sex. Behav. Schaefer, G. A., and Ahlers, C. J. (2018). “Sexual addiction and paraphilias,”
44, 2173–2187. doi: 10.1007/s10508-015-0503-7 in The Routledge International Handbook of Sexual Addiction, eds T.
Jones, S. J., Ó Ciardha, C., and Elliott, I. A. (2020). Identifying the coping strategies Birchard and J. Benfield (London, UK: Routledge/Taylor & Francis Group),
of nonoffending pedophilic and hebephilic individuals from their online forum 83–93.
posts. Sex. Abuse 33, 793–815. doi: 10.1177/1079063220965953 Sea, J., and Beauregard, E. (2018). The hebephiliac: pedophile or teleiophiliac?
Kim, T. K. (2015). T test as a parametric statistic. Korean J. Anesthesiol. 68, Int. J. Offender Ther. Comp. Criminol. 62, 2507–2526. doi: 10.1177/
540–546. 0306624X17723627
Lampalzer, U., Tozdan, S., von Franqué, F., and Briken, P. (2020). Referral for Seto, M. C. (2012). Is pedophilia a sexual orientation? Arch. Sex. Behav. 41,
group or individual treatment? Factors for consideration in the case of self- 231–236. doi: 10.1007/s10508-011-9882-6
referred individuals with a sexual interest in minors. Int. J. Impot. Res. 33, Seto, M. C. (2017). The puzzle of male chronophilias. Arch. Sex. Behav. 46, 3–22.
348–363. doi: 10.1038/s41443-020-0296-7 doi: 10.1007/s10508-016-0799-y
Leavy, R. L., and Adams, E. M. (1986). Feminism as a correlate of self-esteem, self- Seto, M. C., Cantor, J. M., and Blanchard, R. (2006). Child pornography offenses
acceptance, and social support among lesbians. Psychol. Women Q. 10, 321–326. are a valid diagnostic indicator of pedophilia. J. Abnorm. Psychol. 115, 610–615.
doi: 10.1111/j.1471-6402.1986.tb00757.x doi: 10.1037/0021-843X.115.3.610
Lehmiller, J. (2019). Pedophiles Can’t be ‘Cured,’ and It’s Dangerous to Suggest They Stelzmann, D., Jahnke, S., and Kuhle, L. F. (2020). Media coverage of pedophilia:
Can. Vice, 25/01/2019. Available online at: https://www.vice.com/en_us/article/ Benefits and risks from healthcare practitioners’ point of view. Int. J. Environ.
mbzj5q/pedophiles-cant-be-cured-and-its-dangerous-to-suggest-they-can Res. Public Health 17:5739. doi: 10.3390/ijerph17165739
Li, C. K. (1991). ‘The main thing is being wanted’: Some case studies Tozdan, S., and Briken, P. (2015a). ‘I believe I could, so I did’ – a theoretical
on adult sexual experience with children. J. Homosex. 20, 129–143. doi: approach on self-efficacy beliefs to positively influence men with a risk to
10.1300/J082v20n01_09 sexually abuse children. Aggress Violent Behav. 25, 104–112. doi: 10.1016/j.avb.
Lievesley, R., Harper, C. A., and Elliott, H. (2020). The internalization of social 2015.07.015
stigma among minor-attracted persons: Implications for treatment. Arch. Sex Tozdan, S., and Briken, P. (2015b). The earlier, the worse? Age of onset of sexual
Behav. 49, 1291–1304. doi: 10.1007/s10508-019-01569-x interest in children. J. Sex. Med. 12, 1602–1608. doi: 10.1111/jsm.12927
Long, M. L., Alison, L. A., and McManus, M. A. (2013). Child pornography Tozdan, S., and Briken, P. (2019). Age of onset and its correlates in men with
and likelihood of contact abuse: A comparison between contact child sexual sexual interest in children. Sex. Med. 7, 61–71. doi: 10.1016/j.esxm.2018.
offenders and noncontact offenders. Sex. Abuse 25, 370–395. doi: 10.1177/ 10.004
1079063212464398 Tozdan, S., Kalt, A., Keller, L. B., and Briken, P. (2018b). Keep faith in yourself! A
Ly, T., Dwyer, R. G., and Fedoroff, J. P. (2018). Characteristics and treatment pilot study on the relevance of specific self-efficacy for modifying sexual interest
of internet child pornography offenders. Behav. Sci. Law 36, 216–234. doi: in children among men with a risk to sexually abuse children. J. Sex. Marital
10.1002/bsl.2340 Ther. 44, 591–604. doi: 10.1080/0092623X.2018.1437488
Marshall, W. L. (2008). Are pedophiles treatable? Evidence from North American Tozdan, S., Kalt, A., Dekker, A., Keller, L. B., Thiel, S., Müller, J. L., et al. (2018a).
studies. Seksuol. Polska 6, 39–43. Why information matters: Examining the consequences of suggesting that
Marshall, W. L., Barbaree, H. E., and Eccles, A. (1991). Early onset and deviant pedophilia is immutable. Int. J. Offender. Ther. Comp. Criminol. 62, 1241–1261.
sexuality in child molesters. J. Interpers. Violence 6, 323–336. doi: 10.1177/ doi: 10.1177/0306624X16676547
088626091006003005 Upton, G., and Cook, I. (2014). A Dictionary of Statistics. Oxford, UK: Oxford
Marshall, W. L., Marshall, L. E., Serran, G. A., and O’Brien, M. D. (2011). University Press.
Rehabilitating sexual offenders. Washington, DC: American Psychological Vincke, J., and Bolton, R. (1994). Social support, depression, and self-
Association. acceptance among gay men. Hum. Relat. 47, 1049–1062. doi: 10.1177/
Marshall, W. L., O’Brien, M. D., and Marshall, L. E. (2009). “Modifying sexual 001872679404700902
preferences,” in Assessment and Treatment of Sex Offenders: A Handbook, eds Wagner, T., Jahnke, S., Beier, K. M., Hoyer, J., and Scherner, G. (2016).
A. R. Beech, L. A. Craig, and K. D. Browne (Chichester, UK: Wiley), 311–327. Pädophile Neigungen offenbaren: Stigma-Management bei Patienten

aus dem Berliner Präventionsprojekt Dunkelfeld [Disclosing pedophilic Publisher’s Note: All claims expressed in this article are solely those of the authors
preferences: Stigma management amongst patients from the Berlin prevention and do not necessarily represent those of their affiliated organizations, or those of
project dunkelfeld]. Z. Sex. Forsch. 29, 106–130. doi: 10.1055/s-0042-1 the publisher, the editors and the reviewers. Any product that may be evaluated in
08000 this article, or claim that may be made by its manufacturer, is not guaranteed or
Webb, R. (2018). A Review of the Risk Posed by Internet Offenders. Paisley, UK: Risk endorsed by the publisher.
Management Authority.
Wells, A., and Davies, M. I. (1994). The thought control questionnaire: A measure Copyright © 2021 Lampalzer, Tozdan, von Franqué and Briken. This is an open-
of individual differences in the control of unwanted thoughts. Behav. Res. Ther. access article distributed under the terms of the Creative Commons Attribution
32, 871–878. doi: 10.1016/0005-7967(94)90168-6 License (CC BY). The use, distribution or reproduction in other forums is permitted,
provided the original author(s) and the copyright owner(s) are credited and that the
Conflict of Interest: The authors declare that the research was conducted in the original publication in this journal is cited, in accordance with accepted academic
absence of any commercial or financial relationships that could be construed as a practice. No use, distribution or reproduction is permitted which does not comply
potential conflict of interest. with these terms.

APPENDIXES
APPENDIX 1 | IASI rev (Mundt et al., 2011).
Instruction: In the following you are asked to give some information on how you handle your sexual inclination.
This statement is true. . . Not at all Little Moderately Fairly Very much
(1) I forbid myself my sexual fantasies. 1 2 3 4 5

(2) I can enjoy my sexual fantasies without a bad conscience. 1 2 3 4 5
(3) I hate my sexual inclination. 1 2 3 4 5
(4) I cannot accept my sexual inclination. 1 2 3 4 5
(5) My inclination is an integral part of my personality. 1 2 3 4 5
(6) I reject my sexual inclination. 1 2 3 4 5
(7) One must resist disagreeable sexual fantasies. 1 2 3 4 5
(8) I allow myself my sexual fantasies. 1 2 3 4 5
(9) I reject myself because I have this sexual inclination. 1 2 3 4 5
(10) My sexual fantasies have nothing to do with me. 1 2 3 4 5
(11) My sexual fantasies scare me. 1 2 3 4 5
(12) I have no sexual fantasies. 1 2 3 4 5
(13) I am aware of my sexual inclination. 1 2 3 4 5
(14) I like to embellish my sexual fantasies. 1 2 3 4 5
(15) I resist my sexual fantasies. 1 2 3 4 5
APPENDIX 2 | The specific descriptions/questions that the six items represent which assess the frequency of the use of child abuse/exploitation material and frequency
of sexual desire/behavior towards children.
(1) Frequency of Use of Legal • How often in the last 6 months did you consume (1) films with, (2) images/photos of lightly dressed children (e.g., in underwear,
Imagery of Children gym shorts, swimming trunks, leotards, transparent clothing. . .) for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume films with normally dressed children (e.g., children’s films, feature films,
documentary films. . .) for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume images/photos/portraits of dressed children (e.g., in magazines, on postcards,
in illustrated books. . .) for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume imagery of (a) prepubescent, (b) early pubescent children – dressed in
non-sexual representations – for masturbation or stimulating pastime?
• How often in the last 6 months did you consume imagery of (a) prepubescent, (b) early pubescent children – lightly dressed
(e.g., in underwear) or naked in non-sexual representations – for masturbation and/or stimulating pastime?
• In the last 12 months I used media for sexual arousal which depict a child under 14 years of age with a (a) prepubescent, (b)
early pubescent, (c) late pubescent body scheme, dressed, lightly dressed (e.g., in underwear) or naked in real non-sexual
representations (or natural poses, respectively).
(2) Frequency of Use of Legal • How often in the last 6 months did you consume (1) films with, (2) images/photos of lightly dressed adolescents (e.g., in
Imagery of Adolescents underwear, gym shorts, swimming trunks, leotards, transparent clothing. . .) for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume films with normally dressed adolescents (e.g., children’s films, feature films,
documentary films. . .) for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume images/photos/portraits of dressed adolescents (e.g., in magazines, on
postcards, in illustrated books. . .) for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume imagery of late pubescent adolescents – dressed in non-sexual
representations – for masturbation and/or stimulating pastime?
• How often in the last 6 months did you consume imagery of late pubescent adolescents – lightly dressed (e.g., in underwear) or
naked in non-sexual representations – for masturbation and/or stimulating pastime?
• In the last 12 months I used media for sexual arousal which depict an adolescent between 14 and 17 years of age, dressed,
lightly dressed (e.g., in underwear) or naked in real non-sexual representations (or natural poses, respectively).
(3) Frequency of Use of Illegal • How often in the last 6 months did you consume (1) films with, (2) images/photos of naked children (e.g., on the beach, bathing,
Child Sexual Abuse Imagery medically examined, or posing in front of a camera) for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume images/photos of children who were dressed up or posed styled up (e.g.,
made up, in special clothing, school uniforms, sailor suits etc.) for (a) masturbation, (b) stimulating pastime?
(Continued)

APPENDIX 2 | (Continued)
• How often in the last 6 months did you consume (1) films, (2) images/photos with children in which sexual organs (e.g.,
buttocks, penis, vagina, breasts. . .) can be seen in detail and/or in which a child masturbates for (a) masturbation, (b)
stimulating pastime?
• How often in the last 6 months did you consume (1) films, (2) images photos with children in/on which sexual acts are
performed between an adult and a child for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume imagery of (a) prepubescent, (b) early pubescent children – dressed, lightly
dressed (e.g., in underwear), or naked in erotic/slinky/provocative poses – for masturbation and/or stimulating pastime?
• How often in the last 6 months did you consume imagery of (a) prepubescent, (b) early pubescent children which show genital
and anal area in detail for masturbation and/or stimulating pastime?
• How often in the last 6 months did you consume imagery of (a) prepubescent, (b) early pubescent children which show sexual
acts between pre-/early pubescent children (touching, masturbation, mutual stimulation, oral/vaginal/anal intercourse) for
masturbation and/or stimulating pastime?
• How often in the last 6 months did you consume imagery of (a) prepubescent, (b) early pubescent children which show sexual
acts with an adult (touching, masturbation, mutual stimulation, oral/vaginal/anal intercourse) for masturbation and/or stimulating
pastime?
• In the last 12 months I used media for sexual arousal which depict real children under 14 years of age with a (a) prepubescent,
(b) early pubescent, (c) late pubescent body scheme, dressed, lightly dressed (e.g., in underwear) or naked in
erotic/slinky/provocative poses.
(b) early pubescent, (c) late pubescent body scheme and show genital and/or anal area in detail or which show sexual acts
without adults being involved.
(b) early pubescent, (c) late pubescent body scheme and show sexual acts with adults being involved.
(4) Frequency of Use of Illegal • How often in the last 6 months did you consume (1) films with, (2) images/photos of naked adolescents (e.g., on the beach,
Adolescent Sexual Abuse bathing, medically examined, or posing in front of a camera) for (a) masturbation, (b) stimulating pastime?
Imagery • How often in the last 6 months did you consume images/photos of adolescents who were dressed up or posed styled up (e.g.,
made up, in special clothing, school uniforms, sailor suits etc.) for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume (1) films, (2) images/photos with adolescents in/on which sexual organs (e.g.,
buttocks, penis, vagina, breasts. . .) can be seen in detail and/or in which an adolescent masturbates for (a) masturbation, (b)
• How often in the last 6 months did you consume (1) films, (2) images/photos with children in/on which sexual acts are
performed between an adult and an adolescent for (a) masturbation, (b) stimulating pastime?
• How often in the last 6 months did you consume imagery of late pubescent adolescents – dressed, lightly dressed (e.g., in
underwear), or naked in erotic/slinky/provocative poses for masturbation and/or stimulating pastime?
• How often in the last 6 months did you consume imagery of late pubescent adolescents which show genital and anal area in
detail for masturbation and/or stimulating pastime?
• How often in the last 6 months did you consume imagery of late pubescent adolescents which show sexual acts between late
pubescent adolescents (touching, masturbation, mutual stimulation, oral/vaginal/anal intercourse) for masturbation and/or
• How often in the last 6 months did you consume imagery of late pubescent adolescents which show sexual acts with an adult
(touching, masturbation, mutual stimulation, oral/vaginal/anal intercourse) for masturbation and/or stimulating pastime?
• In the last 12 months I used media for sexual arousal which depict a real adolescent between 14 and 17 years of age, dressed,
lightly dressed (e.g., in underwear) or naked in erotic/slinky/provocative poses.
• In the last 12 months I used media for sexual arousal which depict real adolescents between 14 and 17 years of age and show
genital and/or anal area in detail or which show sexual acts without adults being involved.
• In the last 12 months I used media for sexual arousal which depict real adolescents between 14 and 17 years of age and show
sexual acts with adults being involved.
(5) Frequency of Desire for • How often do you have a desire for sex with intercourse with a child/adolescent?
Sexual Activities with Minors • How often do you have a desire for sex without intercourse with a child/adolescent?
• How often do you have a desire for intimate body contact with a child/adolescent?
• I had sexual desire (= a state of a strong longing or strong desire which arises in thought, triggered by, for example, certain
encounters, pictures, moods or feelings; not to be equated with masturbation or other sexual acts) with regard to persons with
a (a) prepubescent, (b) early pubescent, (c) late pubescent, (d) postpubescent body scheme, whether or not an act followed or
not.
(6) Frequency of Sexual • How often do you normally have intercourse with a child/adolescent?
Activities with Minors • How often do you normally have sex without intercourse with a child/adolescent?
• How often do you normally have intimate body contact with a child/adolescent?
• In the last 12 months I performed sexual acts without penetration with (a) prepubescent, (b) early pubescent children, (c) late
pubescent children under 14 years of age, (d) adolescents between 14 and 17 years of age (e.g.,
observing/photographing/filming in intimate situations or of charges, peeping, sexualized talk, watching porn together, invitation
to (mutual) masturbation, (deep) kiss, masturbation in front of a child/adolescent, rubbing, exposing oneself etc.)
• In the last 12 months I performed sexual acts with penetration with (a) prepubescent, (b) early pubescent children, (c) late
pubescent children under 14 years of age, (d) adolescents between 14 and 17 years of age (e.g., intercourse, i.e., oral
intercourse, anal intercourse, vaginal intercourse, penetration with finger and/or objects etc.).

Advantages 90
of publishing OPEN ACCESS

Articles are free to read
FAST PUBLICATION
Around 90 days
in Frontiers for greatest visibility

and readership
from submission
to decision
HIGH QUALITY PEER-REVIEW TRANSPARENT PEER-REVIEW

Rigorous, collaborative, Editors and reviewers
and constructive acknowledged by name
peer-review on published articles
Frontiers
Avenue du Tribunal-Fédéral 34
1005 Lausanne | Switzerland
Visit us: www.frontiersin.org

Contact us: frontiersin.org/about/contact REPRODUCIBILITY OF DIGITAL PUBLISHING
RESEARCH Articles designed
Support open data for optimal readership
and methods to enhance across devices
research reproducibility
FOLLOW US IMPACT METRICS EXTENSIVE PROMOTION LOOP RESEARCH NETWORK

@frontiersin Advanced article metrics Marketing Our network
track visibility across and promotion increases your
digital media of impactful research article’s readership

Child Sexual Abuse: Empirical Research On Understanding and Helping Victims and Offenders

Uploaded by

Copyright:

Available Formats

Child Sexual Abuse: Empirical Research On Understanding and Helping Victims and Offenders

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Child Sexual Abuse: Empirical Research On Understanding and Helping Victims and Offenders

Uploaded by

Copyright:

Available Formats

CHILD SEXUAL ABUSE: EMPIRICAL

RESEARCH ON UNDERSTANDING AND

EDITED BY : Noora Ellonen, Robert Horselenberg, Cristina Izura,

Frontiers in Psychology 1 March 2022 | Child Sexual Abuse: Empirical Research

Frontiers in Psychology 2 March 2022 | Child Sexual Abuse: Empirical Research

Frontiers in Psychology 3 March 2022 | Child Sexual Abuse: Empirical Research

Frontiers in Psychology 4 March 2022 | Child Sexual Abuse: Empirical Research

Editorial: Child Sexual Abuse:

Editorial on the Research Topic

Frontiers in Psychology | www.frontiersin.org 5

Frontiers in Psychology | www.frontiersin.org 6

Frontiers in Psychology | www.frontiersin.org 7

Beliefs About Children’s Memory and

Frontiers in Psychology | www.frontiersin.org 8

Frontiers in Psychology | www.frontiersin.org 9

TABLE 1 | Comparison of child protection services in several countries.

Country Organization(s) Tasks Collaborations

Frontiers in Psychology | www.frontiersin.org 10

Frontiers in Psychology | www.frontiersin.org 11

Statements Response alternatives

Statements Behavioral scientists Medical doctor Social workers Total (N = 154)

1. Our memories are permanently stored in our 44 (n = 19) 73 (n = 8) 30 (n = 30) 37 (n = 57)

Frontiers in Psychology | www.frontiersin.org 12

Statement Behavioral Social χ2 p-value

The bold value means the p value was considered significant.

Item 1 Item 2 Item 3 Item 4 Item 5 Item 6 Item 7 Item 8 Item 9

DISCUSSION include more detailed analysis of underlying traits to explain

Frontiers in Psychology | www.frontiersin.org 13

Frontiers in Psychology | www.frontiersin.org 14

Frontiers in Psychology | www.frontiersin.org 15

Frontiers in Psychology | www.frontiersin.org 16

Risk Factors for Sexual Offending in

Background: The risk of child sexual abuse among non-forensic, non-correctional

Frontiers in Psychology | www.frontiersin.org 17

INTRODUCTION seek treatment for PD without prior identified child sexual

Frontiers in Psychology | www.frontiersin.org 18

Frontiers in Psychology | www.frontiersin.org 19

Frontiers in Psychology | www.frontiersin.org 20

Bolded figures are significant at p < 0.05.

Frontiers in Psychology | www.frontiersin.org 21

Risk factor/domain Pedophilic Disorder Non-clinical control Cohen’s d (95% CI)

Bolded figures are significant at p < 0.05.

Frontiers in Psychology | www.frontiersin.org 22

Child-Related Sexually Motivated

Frontiers in Psychology | www.frontiersin.org 23

Frontiers in Psychology | www.frontiersin.org 24

Department of Clinical Neuroscience, Karolinska Institutet, SUPPLEMENTARY MATERIAL

Frontiers in Psychology | www.frontiersin.org 25

Frontiers in Psychology | www.frontiersin.org 26

“All of Me Is Completely Different”:

Frontiers in Psychology | www.frontiersin.org 27

Frontiers in Psychology | www.frontiersin.org 28

Frontiers in Psychology | www.frontiersin.org 29

Frontiers in Psychology | www.frontiersin.org 30

TABLE 1 | Characteristics and survey-scores across participants.

Anna 13 5 years 24 Numerous Penetration, Yes, i.a. by 2/6/4 Low (4/30)

Frontiers in Psychology | www.frontiersin.org 31

Themes Brief descriptions

From thrilling to abusive

EDITED BY : Noora Ellonen, Robert Horselenberg, Cristina Izura,