Infant Behavior & Development 23 (2000) 91–111

The tactile context of a mother’s caregiving: Implications

for attachment of low birth weight infants夞
Sandra J. Weissa,*, Peggy Wilsonb, Matthew J. Hertensteinc,
Rosemary Camposd
a
Professor, Department of Community Health Systems, University of California, San Francisco, Box 0608,
San Francisco, California 94143, USA
b
Clinical Nurse Specialist, San Francisco General Hospital, San Francisco, CA, USA
c
PhD Candidate, Department of Psychology, University of California, Berkeley, CA, USA
d
Assistant Research Specialist, Institute for Human Development, University of California, Berkeley, CA, USA

Received 8 March 2000; received in revised form 8 September 2000; accepted 8 September 2000

Abstract
This study examined the degree to which specific properties of maternal touch may be associated
with a low birth weight infant’s security of attachment at one year of age, considering the potential
modifying effects of maternal sensitivity and history of touch as well as infant gender and biological
vulnerability. One hundred and thirty one socioculturally diverse infants and their mothers were
evaluated for medical complications during the neonatal period. Videotapes were made of the
infant-mother dyads during an infant feeding at three months of age. Each videotape was analyzed for
a mother’s properties of touch and her sensitivity as well as for infant responsiveness. When the infant
was six months old, each mother completed a questionnaire to determine felt security regarding her
own history of touch as a child. Researchers completed the Attachment Q-Set for each infant at one
year of age.
Analysis of covariance indicated that sheer frequency of touch had no relationship to infant
attachment but use of nurturing touch by mothers was associated with security of attachment.
However, the degree of infant vulnerability (i.e., perinatal complications, birthweight, and respon-
siveness) moderated the effects of nurturing touch. Nurturing touch was associated with more secure
attachment for robust infants but with less secure attachment for highly vulnerable babies. Neither
maternal sensitivity nor gender appeared to moderate the relationship of touch to attachment or to have
any direct relationship to attachment outcomes. Infant vulnerability decreased the likelihood of a

夞This research was funded by the National Institutes of Health, NINR, # NR02698 –5.
* Corresponding author. Tel.: ⫹1-415-476-1504; fax: ⫹1-415-476-6042.
E-mail address: sweiss@itsa.ucsf.edu (S.J. Weiss).

0163-6383/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved.
PII: S 0 1 6 3 - 6 3 8 3 ( 0 0 ) 0 0 0 3 0 - 8
92 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

secure attachment while a mother’s felt security regarding her own tactile experience as a child
increased her infant’s chances of having a secure attachment. © 2000 Elsevier Science Inc. All rights
reserved.

Keywords: Parenting; Touch; Attachment; Low birth weight

1. Introduction

Attachment theorists have long regarded the quality of parent-infant physical contact as a
central feature of the responsive and available caregiving environment that is necessary in
fostering an infant’s sense of security (Ainsworth et al., 1978; Bowlby, 1973). In fact, Main
(1990) argues that physical contact with an attachment figure is “the ultimate signal” to the
infant that s/he is safe and secure, noting that Bowlby (1969) placed physical contact as a
central requirement of proximity and felt security. The thesis that ‘contact comfort’ creates
an experience of refuge or safety for a child was originally proposed by Harlow (1962;
Harlow & Zimmermann, 1959) whose research with rhesus monkeys provided evidence of
the importance of bodily contact between infant and mother as the basis of reassurance and
felt security for an infant. Subsequent human research also indicates that close bodily contact
with the mother may be superior to other modalities in helping an infant cope with distress
or discomfort (Grossman et al., 1985; Hunziker & Barr, 1986). Although much of the focus
on physical contact has been on its role in situations of stress or uncertainty, Ainsworth and
colleagues’ (1978) seminal work demonstrated that it was the quality of touch in day- to- day
interactions, not simply separations, which influenced the infant’s attachment expectations.
While the link between physical contact and attachment may be salient for all infants, it
has special importance for low birth weight (LBW) babies. Because of their fragile or
underdeveloped nervous systems, LBW infants have diminished ability to regulate their
responses to their tactile environments during the first few months of life. Numerous studies
have demonstrated potentially detrimental effects of touch on high risk infants’ neurobe-
havioral status and emotional distress (e.g., Als, 1986; Gunnar et al., 1987; Harrison et al.,
1990). Because of these findings, many hospitals have adopted ‘minimal handling’ policies
that require nurses and parents to touch babies infrequently, and primarily for essential
medical procedures or cleaning rituals. As a result, some parents learn to withhold their touch
to protect their baby and thus develop ingrained patterns of infrequent interaction which
continue over time. Other parents who are less sensitized to their babies’ vulnerabilities may
overstimulate their infants without recognizing their regulatory difficulties. LBW infants
may also be especially vulnerable to touch because of their history of invasive and painful
medical procedures which can distort their somatosensory perception and increase the
potential for touch to be aversive (DeMaio-Feldman, 1994).
In addition, there is evidence that certain types of neurobehavioral limitations (many of
which are experienced by LBW infants) may affect the attachment relationship adversely
(Belsky & Rovine, 1987; Crockenberg, 1981; Miyake, Chen & Campos, 1985; Vaughn &
Bost, 1999; Van den Boon, 1994). These limitations (e.g., difficulty orienting, weaker muscle
tone, irritability, tremors, lack of responsiveness) are associated with diminished CNS
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 93

integrity at birth, resulting from underdeveloped or medically compromised nervous systems.

In contrast to full term, robust infants who have every sensory system functioning on their
behalf at birth or soon after, many LBW babies are at risk because of such neurobehavioral
problems. They are less capable of dampening their own distress, depending substantially on
a caregiver during their first few months to regulate their neurobehavioral response to both
internal and external stimuli (Als, 1986). As a result, they are especially susceptible to the
effects of parental behavior. In addition, the infant’s ability to engage the mother via eye
contact, vocalization or facial expression predicts greater security of attachment (Grossman,
Grossman & Schwan, 1986), placing minimally responsive LBW babies at even greater risk
for attachment problems (van Beek, Hopkins & Hoeksma, 1994).

2. The impact of touch on attachment

Attachment theory emphasizes the role of touch in the context of maternal sensitivity and
responsiveness to distress. The literature highlights two major characteristics of touch: its
overall availability to the infant and the degree to which it is affectionate and tender versus
rejecting or abusive.

2.1. Availability or frequency of touch

The amount of touch provided to the child is often defined as an index of maternal
availability. Ainsworth and colleagues (1978) originally noted that mothers of secure infants
used touch more extensively than mothers of insecure-anxious infants. Significant correla-
tions between the attachment of infants and a number of maternal qualities of touch have also
been described in a review and synthesis article written by several scientists with related
programs of research (Waters, Kondo-Ikemura, Posada and Richters, 1990). These authors
detail the relationships of monkey infant attachment to a mother’s a) close holding of her
infant for extended periods, b) carrying of the infant when moving from place to place, and
c) tolerance of the infant’s touching of her body during play. They also present data
indicating that human infants classified as secure enjoy more playful physical contact with
their mothers than babies classified as insecure.
In one of the few experimental studies manipulating the amount of physical contact
between mothers and their infants, Anisfeld and colleagues (1990) compared attachment
outcomes for infants who were carried ventrally in soft baby carriers with those transported
in infant seats. They concluded that increased maternal contact from early carrying enhanced
the likelihood of secure attachments for the infants.
In contrast, other studies indicate that mothers of insecure-avoidant infants may have some
aversion to touch, appearing to dislike close physical contact with their infants (Egeland &
Farber, 1984; Main, 1981; Main & Stadtman, 1981). Some scholars speculate that this
discomfort with touch is transmitted from mother to infant because the babies of mothers
who have insecure attachments begin to show some aversion to and/or lack of enjoyment of
physical contact as they develop (Main, 1990; Pederson & Moran, 1995; Posada et al., 1995).
While results cited above imply that more extensive touch enhances attachment outcomes,
94 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

other findings suggest that too much touch can have a negative impact on attachment.
Ainsworth et al. (1978) noted that mothers of insecure-resistant infants used touch that was
interfering and at times inappropriate to the infant’s needs. Other investigators have found
that insecure-avoidant attachments are related to intrusive, excessively stimulating caregiv-
ing approaches (Lewis & Fiering, 1989; Smith & Pederson, 1988; Vondra et al., 1995). On
the other hand, ‘moderate stimulation’ has been associated with greater security of attach-
ment (Belsky, Rovine & Taylor, 1984; Feldstein et al., 1995). This distinction is clearly
salient for low birth weight infants who are susceptible to overstimulation from touch as a
result of their fragile nervous systems. The effects of frequent touch on infant security of
attachment may well be moderated by the infant’s degree of biological vulnerability.
It has been argued that sensitive parenting may be particularly important in assuring
secure attachments when infants present more challenging neurobehavioral or temperamental
profiles such as those often seen with LBW infants (Frodi & Thompson, 1985; Pederson &
Moran, 1995). A mother’s sensitivity to infant cues and responsiveness to distress have been
integral concepts within attachment theory, viewed as antecedents to later attachment
security (Bowlby; 1973; 1982). Based upon this literature, it seems likely that a mother’s
sensitivity would interact with her frequency of touch to determine its effects. When a mother
is more generally sensitive toward the infant, touch would more likely be used with attention
to the child’s unique cues and state of readiness. Less sensitive mothers may use touch
without consideration of these factors, leading to more negative effects on the infant’s felt
security.

2.2. Nurturing versus harsh touch

In addition to their more extensive use of touch, Ainsworth and colleagues (1978) found
that mothers of securely attached infants also used more affectionate and tender touch with
their babies than did mothers of insecure infants. Specifically, they caressed, hugged and
kissed their infants more frequently and were careful and gentle in their handling. Other
studies have supported these initial findings, indicating that warm, affectionate touch is more
prominent among mothers of secure infants (Bates et al., 1985; Leyendecker et al., 1997;
Network, 1997; O’Connor et al., 1992) but lacking in mothers of insecure-resistant infants
(Schneider-Rosen & Cicchetti, 1984).
In contrast to nurturing touch, research has identified rejecting, angry, abrupt touch as a
more typical pattern of mothers of both insecure-avoidant infants and infants with disorga-
nized attachments (Ainsworth et al., 1978; Carlson et al., 1989; Crittenden, 1988; Van der
Kolk & Fisler, 1994). The impact of such harsh touch has been described most extensively
for children who are physically abused. They, as well as children exposed to substantial
physical punishment, show a high incidence of insecure attachments (Main & Hesse, 1990;
Main & Solomon, 1986; van IJzendoorn et al., 1992). These studies lead to the hypothesis
that less use of harsh touch and greater use of nurturing touch should be associated with
secure attachment. However, there is some danger in assuming that tender, affectionate touch
is inherently beneficial. Gentle, loving touch could be intrusive if used at an undesirable
moment. Such a possibility suggests that a mother’s sensitivity may also moderate the effects
of nurturing or even harsh touch on the infant’s attachment to the mother.
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 95

Gender represents another variable that could influence the relationship between maternal
touch and a child’s attachment. Research suggests that there are differential effects of
nurturing touch on boys versus girls, with these effects being much stronger for boys (Weiss,
1990; 1999). Such findings may stem from differences that have been noted (in both human
and animal studies) in the ways in which males and females perceive and process the touch
they receive (Diamond, 1990; Hofer,1981; Weiss, 1993). These touch-specific gender dif-
ferences are given further significance by the fact that LBW boys are consistently more at
risk for worse socioemotional outcomes than LBW girls (Brooks-Gunn & Benasich, 1997;
Pharaoh et al., 1994).

2.3. Maternal history of touch

Research supports the notion that a mother’s history of attachment-related experiences as

a child will influence the nature of the attachment she cultivates with her own child. Evidence
of this intergenerational transmission of attachment behavior abounds (Benoit & Parker,
1994; George & Solomon, 1996; Steele, Steele & Fonagy, 1996; van IJzendoorn, 1995; Ward
& Carlson, 1995). Within the theoretical tradition of Bowlby and Ainsworth, an individual’s
physical contact experience with the attachment figure would be an integral feature of overall
attachment security. A mother’s state of mind with respect to the touch in her family of origin
would constitute part of her internal working model of attachment behavior (i.e., her
cognitive and emotional expectations of how attachment figures respond when they are
turned to for support or security). Negative tactile experience could foster a lack of felt
security with the attachment figure’s physical availability when the need for contact comfort
arises. It is such a history that would potentially compromise subsequent attachment-related
interactions. For example, Main (1990) found that a mother’s aversion to physical contact
with her child during a stressful situation was significantly related to the mother’s perceived
rejection of her own bids for contact as a child and to her infant’s insecure attachment
behavior (Main & Weston, 1981). Such theoretically-based findings suggest that a mother’s
felt security regarding her own tactile history may interact with her actual use of touch to
influence her infant’s security of attachment.

3. Research aims

Based upon the theory and research described above, our study had two major aims: 1) to
determine whether a mother’s frequency of touch or her use of nurturing versus harsh touch
are associated with the low birth weight infant’s security of attachment, and 2) to determine
whether selected moderating variables may interact with maternal touch to influence its
relationship to attachment: specifically, a) the mother’s sensitivity toward the infant, b) the
mother’s felt security regarding her own history of touch, c) the infant’s degree of biological
vulnerability, and d) the infant’s gender.
96 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

4. Methods

4.1. Sample

The sample included 131 infants and their mothers who were recruited into the study
during the infant’s first two weeks of life. These dyads were part of a larger research program
focused on early development of low birth weight infants. The sample was ethnically diverse,
including 46% Caucasian, 28.4% Hispanic, 19.3% African American, and 6.5% Asian or
Native American babies. Their weight ranged from 570 g to 2500 g with a mean birthweight
of 1487 g (SD ⫽ 547). Ninety-four percent of the babies had a gestational age of less than
37 weeks, with a sample mean of 32 weeks. Forty-four percent of the babies were girls and
56% were boys, with 22% known to have been drug exposed in utero. Mothers had a mean
age of 29 years (SD ⫽ 7), ranging from 16 to 44. They averaged 12.5 years of education,
with a spread from 1 to 20 years. Eighty-one percent of the mothers lived with their partner,
but not necessarily a spouse. Forty-five percent of the mothers were living below the poverty
level. Sixty-one percent were employed at least part-time, with 29% being the sole support
for their family.

4.2. Procedures

Families were recruited from the neonatal intensive care units of 3 major teaching
hospitals in northern California. Mothers were told that the overall purpose of the research
was to understand factors that might influence the health and development of babies who are
small at birth. Once an informed consent was acquired, the mother was given a demographic
questionnaire to complete regarding both herself and her baby. When the baby was one
month old, a research assistant (RA) with clinical expertise in care of neonates reviewed the
medical charts of mother and infant to identify the presence or absence of specific perinatal
risk factors that constituted the Perinatal Complications Scale. When the baby was 3 months
of age, another RA who was a clinical nurse specialist made a home visit to videotape the
mother and infant during a typical feeding. Mothers fed the infant in their normal feeding
mode, either breast or bottle. The videocamera was set up and turned on as soon as the RA
arrived in order to allow time for the mother to adjust to its presence. The camera was then
left on the mother and infant while the baby was being fed. The videotape of the feeding was
analyzed subsequently using the Tactile Interaction Index (to measure the mother’s use of
touch) and the Nursing Child Assessment Feeding Scale (NCAFS) to measure the infant’s
interactive responsiveness and the mother’s sensitivity. When the infant was 6 months of age,
the RA made another home visit at which time the mother completed a questionnaire
regarding her experience with and attitudes toward physical contact. Shortly after the infant’s
first birthday (corrected for gestational age), an appointment was made for a final home visit
to interview the mother and spend approximately 3 hrs observing the mother and child during
a playtime and feeding. The interview and observations were used as the major source of
input for the RA’s completion of the Attachment Q-Set.
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 97

4.3. Measures

4.3.1. Infant vulnerability. Vulnerability was measured with an index that combined scores
from 3 variables: infant birthweight, perinatal risk status, and interactive responsiveness of
the infant at 3 months of age. Birthweight and perinatal risk data were determined by the
Perinatal Complications Scale (Littman & Parmelee, 1978; Parmelee et al., 1976). The
measure assesses maternal obstetrical history, pregnancy events, perinatal measures of the
infant (e.g., gestational age and apgar), and hazardous postnatal events such as need for
ventilatory assistance, metabolic abnormality, or surgery. A summary score is based on the
number of risk factors incurred by a mother-infant pair, with higher scores indicating greater
risk. Risk data were acquired primarily from chart review during the first month postnatal,
with supplemental input from the infant’s primary nurse. Reliability and validity have been
demonstrated for these scales (Francis et al., 1987; Scott et al., 1997). The score for infant
responsiveness (which was combined with the perinatal risk score and the birth weight score
to yield an infant’s vulnerability index) was determined through the NCAFS measure
described next.

4.3.2. Maternal sensitivity and infant responsiveness. The NCAFS was used to measure
infant responsiveness to the mother as well as maternal sensitivity. The scale is an obser-
vational measure with established normative data. It shows evidence of total and subscale
internal consistency and test-retest reliability as well as construct, concurrent and predictive
validity (Barnard et al., 1989; Sumner & Spietz, 1994). While observations relevant to the
entire scale were made, only subscale scores pertinent to this study are described here: 1) the
infant subscale for ‘responsiveness to caregiver’ which was incorporated into the vulnera-
bility index mentioned above, and 2) two maternal subscales for ‘sensitivity to cues’ and
‘responsiveness to distress’. The subscales each consist of binary items which are summed;
larger scores indicate greater sensitivity or responsiveness. The measure also provides a
combined maternal score which is the sum of the ‘sensitivity to cues’ and ‘responsiveness to
distress’ scales; this is the score we used in our analysis.
An expert coder was trained and certified as reliable in use of the NCAFS by the
University of Washington NCAST office (i.e., met their criterion of at least 85% reliability
for each scale on a series of standardized case studies). This individual completed the coding
for all infant-mother dyads based upon review of the entire videotape of their feeding
situation. The coder was blind to any other information about the mother-infant pairs.

4.3.3. Maternal touch. The specific properties used by mothers in their touching of infants
were measured with the Tactile Interaction Index (TII) (Weiss, 1992). The TII includes both
a series of training videotapes differentiating various properties of touch as well as a coding
system to analyze these properties within a specified interaction. The coding system was
created specifically for microanalysis of videotaped or filmed interaction. While it has
indices measuring the intensity, location, action and frequency of a touch, only the action and
frequency indices were pertinent to the variables under study (i.e., nurturing touch, harsh
touch and frequency of touch).
The action index identifies 28 different gestures or movements that can be used in
98 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

touching (e.g., grab, push, rub, squeeze). From the raw data regarding specific tactile
gestures, 2 percentage scores were developed. The first score rated each mother on the degree
to which she used nurturing actions (e.g., stroking, kissing) and the second score identified
the degree of harsh actions (e.g., hitting, slapping). In contrast to the action index, the score
for frequency of touch was not a percentage score but rather the actual number of times a
mother touched her baby.
A standardized 5 minute segment of the infant feeding situation was used for analysis of
each mother’s touch. The initial few minutes of the feeding were not used in order to allow
mother and baby to adjust comfortably to their normal feeding routine. Two subsequent
minutes were then coded during active feeding of the baby. Three additional minutes were
also coded at the end of the feeding, when mothers were burping the baby and engaging in
other interaction with the infant unrelated to the feeding per se. Although we found no
significant differences within the sample for overall qualities of touch used by mothers who
breast and bottle fed, this latter three minutes of the coding period allowed for a variety of
interaction that was independent of the mother’s feeding mode. Standardizing the length of
time coded for each feeding limited some of the variance in overall frequency of touch.
However, it was necessary in light of a number of factors, including the length of some feeds,
periods where the baby was sleeping with no interaction taking place, and the time-intensive
nature of the tactile coding process. The natural, spontaneous quality of the feeding was
preserved by standardizing what was coded rather than rigidly structuring the feeding
situation itself.
The content, construct and predictive validity, as well as internal consistency, interrater
reliability and test-retest stability of the TII have been established previously (Weiss, 1990;
1992). A trained RA (who had achieved coding reliability of 0.92 or above for each index)
reviewed the five minute feeding tapes two separate times, coding the action and frequency
indices separately. This RA had no other contact with the families participating in the study,
nor any access to information about them.

4.3.4. Maternal history of touch. A mother’s history of touch was measured by the Physical
Contact Assessment Questionnaire. The 20 item Likert -type questionnaire was developed
specifically for the larger research program of which this study was a part. Although there
are 3 subscales within the entire measure (I. ‘Attitudes toward Touch’, II. ‘Felt Security
regarding Tactile Experience as a Child’, and III. ‘Felt Security regarding Current Tactile
Experience’), only the score from the subscale assessing tactile experience as a child was
used in the study described here. This second subscale was the one most relevant to the
mother’s internal working model of attachment and, in fact, the only subscale which showed
a relationship to infant attachment when a preliminary correlation matrix was computed. The
self-report items in this subscale measured the degree to which the touch received in the
mother’s family of origin was perceived as affectionate versus punitive and the degree to
which the amount and type of touch met the respondent’s needs for affection and comforting
as a child. The subscale’s alpha reliability within the sample was 0.83. We have found in the
larger research program (from data of 185 mothers of low birth weight infants) that the
subscale predicts the degree of a mother’s rejection of her own child, with less felt security
in her past tactile experience predicting greater rejection (r ⫽ ⫺0.24, p ⬍ .003). These data
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 99

also indicate that the subscale is correlated with a) the mother’s felt security in her current
tactile experience (r ⫽ 0.33, p ⬍ .000) and b) the perceived supportiveness of her current
relationships (r ⫽ 0.17, p ⬍ .04). Its predictive validity is further evidenced by correlations
with family functioning (r ⫽ 0.23, p ⬍ .005) and satisfaction with family interactions (r ⫽
0.32, p ⬍ .000).

4.3.5. Infant attachment security. Security of attachment was measured with the Attachment
Q-Set (Waters & Deane, 1985). This is a 90 item Q Sort using observations of secure based
behavior of infants and young children during their interactions with primary caregivers in
the home. The items are sorted into 9 categories in terms of whether they are characteristic
or uncharacteristic of the child being described. All RAs who were assigned to follow various
families longitudinally had been trained in the use of the Q-Set. Each RA completed the
Q-Set for a particular child with whom they had worked for a full year prior to the Q-Set
completion. As part of the larger research program, each RA had bimonthly contact with a
specific group of families for whose data collection they were responsible. This contact
involved home visits of about two hours in length when the infant was both 3 and 6 months
of age. Although these visits were focused on data collection unrelated to the attachment
measure, the RA was able to use this extended experience with a specific family as an
additional source of input for sorting the statements.
Specific to the attachment Q-Set, the RA spent approximately 3 hrs in the home shortly
after the child’s first birthday, observing the child and mother during a playtime and a
feeding. As part of our protocol for the observations, the mother was asked to leave the room
at one point without telling the child where she was going or when she would return. After
a period of observing the child’s response, the RA offered the baby a new toy and initiated
a play attempt. After about 3–5 minutes, the mother returned. During each of these separate
periods of time, the RA observed the infant’s level of distress specific to the brief separation
and reunion, the various responses to mother and RA, and the interaction of mother and
infant. While not attempting to replicate the carefully standardized features of the ‘Strange
Situation’ protocol, this semistructured aspect of the home observation did enhance the RA’s
ability to characterize the child on the attachment-related behaviors in the Q-Set. Finally, as
an additional attempt to create a comprehensive and valid data base from which to determine
the child’s characteristics, the RA interviewed the mother to acquire her input regarding the
behaviors in the Q-Set. For this interview, the 90 statements were converted into a respon-
dent-friendly interview format. Within 24 hrs after this home observation and interview, the
RA completed the Q-Set, integrating past experiences with the family, the three hour home
observation, and maternal interview data to make judgments regarding the relevance of each
statement to the child. Waters and Deane’s (1985) criterion set of items for secure attachment
was used as the measure of security, with the percentage of agreement for an infant’s ranking
on the criterion items being the final score for secure attachment.
Studies report interobserver reliability ranging from 0.72 to 0.95 for the Q-Set (Belsky &
Rovine, 1990; Solomon & George, 1999; Teti & McGourty, 1996). We found an interob-
server reliability of 0.89 for those trained to complete the Q-sorts in our study. Content,
construct and predictive validity have been supported (Kerns, 1994; LaFreniere et al., 1992;
Posada et al., 1995; Strayer et al., 1995).Concurrent validity is evidenced by the measure’s
100 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

ability to differentiate attachment classifications using the ‘Strange Situation’ in several but
not all studies (Belsky & Rovine, 1990; Mangelsdorf et al., 1996; Vaughn & Waters, 1990).

4.4. Data analysis

Analytic procedures involved three phases. First, we computed a correlation matrix to

identify independent and moderating variables with the potential for a significant relationship
to security of attachment. Independent variables were the mother’s frequency of touch,
nurturing touch, and harsh touch. Moderating variables included the infant’s biological
vulnerability and gender as well as the mother’s sensitivity toward the child and her felt
security regarding her own tactile experience as a child. Because gender was a discrete
variable, we performed t tests to determine any potential differences in security of attachment
between boys and girls.
In the second phase, we identified any interaction terms that might have a significant effect
on attachment. We tested the potential effects of the interaction between each independent
variable and each moderating variable through regression models, computing separate
equations for each independent variable.
Finally, because we were interested in differentiating any subtle effects of various degrees
of maternal touch on attachment, analysis of covariance (ANCOVA) procedures were
employed to address our research aims. We computed separate ANCOVA’s for each touch
variable as between subjects factors, grouping subjects into 3 categories based on their
extensive, moderate or minimal use of each type of touch. Any moderating variables which
were identified via preliminary analysis as having significant associations to attachment were
also included in the model as covariates or interaction terms using a nested design.
Because only 9 mothers used any harsh touch during the feeding, we could not perform
all of the above procedures to examine its link to attachment security. We did, however,
compute the correlation between the degree to which mothers used harsh touch and their
child’s security. t tests were also computed to compare the attachment security scores of
children whose mothers used any harsh touch (n ⫽ 9) with children whose mothers used
extensive nurturing touch, that is, at least 75% of their total touch (n ⫽ 12).

5. Results

5.1. Maternal touch descriptors

Within the standardized feeding period, mothers touched their babies an average of 213
times (SD⫽109), ranging from 41 to 714 touches. As an average, approximately half of a
mother’s touch was nurturing (47%), with a range from 0% to 91%. Seventeen percent of the
mothers used no nurturing touch at all during the feeding situation. For the 9 mothers who
used harsh touch (slapping or hitting), their mean for its use was 2% of their total touch, with
a range from 1% to 19%.
Maternal age, income and years of education had no relationship to the frequency of
maternal touch. However, a mother’s use of nurturing touch was positively associated with
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 101

Table 1
Pearson correlations for the association of infant attachment to other infant and maternal variables
Infant Maternal
V FT NT HT Hx S
Infant
Attachment Security ⫺.20* .12 .22** .61 .19* .02
Vulnerability (V) .07 .07 ⫺.41 .00 ⫺.13
Maternal
Frequency of Touch (FT) .06 .11 .05 .09
Nurturing Touch (NT) .82** ⫺.02 .02
Harsh Touch (HT) .29 ⫺.32
History of Touch (Hx) .09
Sensitivity (S)
Note: N ⫽ 131 except for all correlations involving harsh touch (n ⫽ 9).
*p ⬍ .05, **p ⬍ .01.

more adequate financial income (r ⫽ 0.19, p ⬍ .04), older age (r ⫽ 0.25, p ⬍ .004), and years
of education (r ⫽ 0.34, p ⬍ .000). Once we controlled for these socioeconomic variables, we
found no differences across ethnic groups in their patterns of touch.

5.2. Preliminary associations between attachment security and study variables

The preliminary coefficients for potential correlations among study variables are displayed
in Table 1. Only one of the independent variables, nurturing touch, showed any significant
relationship to infant security of attachment. The correlation between scores for a mother’s
use of harsh touch and attachment security was substantial and in a positive direction.
However, this relationship involved only 9 mothers and was not significant. Two of the
moderating variables, infant vulnerability and maternal history of touch, were also associated
with attachment security. The t tests to examine potential differences between boys and girls in
attachment security or for any other study variables identified no gender differences of any kind.
Only one significant relationship was noted among the independent and moderating
variables. Nurturing touch was significantly and positively related to harsh touch for the
small subsample of 9 mothers who used any harsh touch. In other words, mothers who used
harsh touch with their infants also used a substantial amount of nurturing touch.
In our preliminary regressions to identify any interaction terms which might be associated
with attachment security, only one relationship was significant or showed even a trend toward
significance. This was the interaction between infant vulnerability and a mother’s use of nurturing
touch. So we built this interaction into the final analysis of covariance, along with the other
variables which had shown significant relationships in the preliminary phases of analysis.

5.3. Final ANCOVA models for frequent and nurturing touch

The results of the model for nurturing touch are shown in Table 2. Nurturing touch, infant
vulnerability and the interaction between nurturing touch and vulnerability were all signif-
102 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

Table 2
Analysis of covariance for effects of nurturing touch on infant attachment security
Source df F
Nurturing Touch (N) 2 4.87*
Infant Vulnerability (V) 1 4.02*
Maternal Touch History 1 3.78
N⫻V 2 5.00*
Error 124 (1.45)
Note: The value in parentheses represents mean square error.
*p ⬍ .05.

icantly associated with an infant’s security of attachment. Although a mother’s felt security
regarding her own tactile experience as a child had been significantly associated with her
infant’s attachment in the preliminary correlations, it showed only a trend toward signifi-
cance in the final model.
Post hoc contrasts (using a modified Bonferroni procedure) among the three groups of
mothers who had used different degrees of nurturing touch showed a significant difference
between one of the groups and the other two. The group of children whose mothers used
extensive amounts of nurturing touch had significantly higher scores for their security of
attachment (M ⫽ 0.78) than either children whose mothers used a moderate amount of
nurturing touch (M ⫽ 0.67, p ⬍ .04) or a minimal amount of nurturing touch (M ⫽ 0.63, p ⬍
.03). However, cross breaks to identify specific effects within the interaction between infant
vulnerability and nurturing touch indicated significant differences in the effects of nurturing
touch on very vulnerable versus more robust LBW infants. While the effects were in a
positive direction for the larger group of more robust infants (r ⫽ 0.29, p ⬍ .05), nurturing
touch was negatively associated with security of attachment for the higher risk, more
vulnerable babies (r ⫽ ⫺0.40, p ⬍ .007).
As shown in Table 3, frequency of touch showed no relationship to an infant’s security of
attachment nor did its interaction with any of the moderating variables. However, it is
interesting to note that when no variance was accounted for by a mother’s actual touch of her
infant, her felt security regarding her own tactile history was significantly associated with her
infant’s security of attachment.

Table 3
Analysis of covariance for effects of frequent touch on infant attachment security
Source df F
Frequency Touching 2 .16
Infant Vulnerability 1 4.14*
Maternal Touch History 1 4.80*
Error 126 (4.75)
Note: The value in parentheses represents mean square error.
*p ⬍ .05.
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 103

5.4. Harsh touch

As noted earlier, we were unable to adequately test the relationship between harsh touch
and security of attachment. The mean score on security of attachment for children receiving
harsh touch (M ⫽ 0.69) was somewhat lower than the mean score for children receiving high
amounts of nurturing touch (M ⫽ 0.78) but the t test was not significant.

6. Discussion

6.1. Maternal touch and infant attachment

6.1.1. Nurturing touch. The results did indicate that a mother’s use of more nurturing touch
was related significantly to a secure attachment, but only for more robust LBW infants of
larger weight and minimal perinatal risk. For this group, infants of mothers who used
nurturing touch as 75% or more of their total touching of the infant were more likely to
develop secure attachments than infants whose mothers used either moderate or minimal
amounts of nurturing touch. Their attachment scores decreased proportional to the dimin-
ishing use of nurturing touch, although the use of moderate vs minimal nurturing touch did
not appear to make a significant difference to attachment outcomes. These findings do
support previous attachment research which indicates that mothers of securely attached
infants use more affectionate and tender touch with their babies than mothers of less secure
infants (e.g., Ainsworth et al., 1978; Leyendecker et al., 1997).
However, the results show an opposite effect of nurturing touch for the more vulnerable
infants in our sample. For these very low birth weight, minimally responsive babies with
many perinatal complications, the use of extensive nurturing touch was associated with less
secure attachment. These findings indicate that more vulnerable infants may be more
susceptible to overstimulation and distress from large amounts of kissing and caressing as a
result of their fragile and underdeveloped nervous systems. Previous research has shown that
high risk infants with more complicated courses exhibit both a diminished ability to modulate
their arousal (Aylward et al., 1984) and substantial avoidance of interaction (Bennett, 1997)
in an apparent attempt to manage their arousal. Because of their less sophisticated use of
regulatory skills such as gaze aversion or crying, very high risk infants have less ability to
communicate to their mothers any aversion or overstimulation they may experience during
the interaction. In addition, infants who experience more perinatal complications and longer
hospital stays may develop tactile aversion as a result of the often painful handling and
probing incurred as part of their medical care (DeMaio-Feldman, 1994). These previous
findings help to explain our results.
However, we expected that frequent touching rather than nurturing touch might show a
potentially detrimental effect for attachment of high risk infants. But, as we considered the
results, we noted that patting the infant’s back was the most frequently used tactile action by
mothers and was thus highly correlated with frequency of touch. In contrast to patting (which
in most cases involves minimal stimulation), actions such as stroking, hugging and kissing
make contact with numerous areas of the child’s body, involving more proprioceptive
104 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

stimulation, stronger intensity, and contact with areas of the infant’s body having many nerve
pathways such as the face, head and hands. It is this combination of factors (rather than
affectionate touch per se) which may be excessively arousing or intrusive for the infant,
engendering feelings of distress rather than comfort from extensive nurturing touch.
The direct association we found between vulnerability and attachment also suggests that
more vulnerable infants are at risk for less secure attachments, irrespective of the nature of
touch received. This outcome seems very important in light of previous studies which
indicate that prematurity does not adversely affect the attachment relationship (Frodi &
Thompson, 1985; Macey et al., 1987). Although the mean attachment scores for our sample
do not imply that the infants as a whole are insecurely attached and we did not look at
attachment classifications per se, our findings reinforce previous research which has shown
that “high risk” premature infants exhibit more insecure attachments than their full term
counterparts (Plunkett et al., 1986). These results speak to the need for individualized
attention to the low birth weight infant’s unique challenges so that intervention can be
tailored to assist the highest risk infants and their mothers.
In contrast to the significant effects of infant vulnerability, the child’s gender appeared to
have no association to attachment. Our findings did not support the prediction that LBW boys
are at substantially greater risk than girls for attachment problems as a result of the touch they
receive. It is interesting to note that the absence of any gender differences is in concert with
much of previous attachment research in which a paucity of sex differences has emerged
(Belsky & Cassidy, 1994; Hazan & Shaver, 1994).

6.1.2. Harsh touch. Our ability to address our research aim regarding harsh touch was
compromised by its lack of adequate distribution. However, infants of the small subsample
of mothers who did use harsh touch did not have significantly lower scores for secure
attachment than infants receiving extensive nurturing touch. Their scores were indeed lower
(M ⫽ 0.69 vs 0.78) but the small sample size in this analysis most likely precluded adequate
power to achieve any level of significance. It is surprising, however, that infants receiving
harsh touch had higher scores for secure attachment (M ⫽ 0.69) than infants receiving
moderate (M ⫽ 0.67) and minimal amounts (M ⫽ 0.63) of nurturing touch. In addition, there
was a positive and significant correlation between use of harsh touch and use of nurturing
touch (r ⫽ 0.82). This finding suggests that mothers who used harsh touch also were
providing fairly substantial amounts of affectionate tactile stimulation; that is, they were
tactually expressive mothers whose touch communicated both positive and negative affect.
Although the number of mothers using harsh touch seriously limits our ability to generalize
from their data, the provocative nature of these findings warrants further study. We need to
test the speculation that receiving harsh touch per se may have no major negative effects on
a child’s development of a secure attachment as long as the child is also receiving appropriate
nurturing touch. It is important to remember that mothers in our sample who used harsh touch
did so, on average, only 2% of their total touching. Still, to slap or hit a child while being
videotaped as part of a study would be more likely to occur if the mother used this type of
touch quite spontaneously as part of her typical repertoire of caregiving. In light of this, we
question whether it may be the absence of kissing, hugging and caressing (rather than the
presence of harsh touch) which places the child at greater risk for insecure attachment. As
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 105

long as nurturing touch is also available, a limited degree of harsh touch may not jeopardize
an infant’s general sense of being safe and secure.

6.1.3. Frequency of touch. The amount of touch received by a child appeared to have no
relationship to his/her security of attachment. In fact, the mean score for secure attachment
was identical for infants regardless of whether their mothers fell into the group which used
very minimal amounts of touch, moderate amounts, or a high frequency of touch. These
results differ from some research cited in the introduction to this article, indicating that both
too little touch and too much touch may have negative consequences for attachment (e.g.,
Main & Stadtman, 1981; Vondra et al., 1995). However, our study did not examine maternal
aversion to touch nor the intrusive or interfering quality of touch, characteristics which have
been associated with minimal touch or overstimulating touch respectively. In addition, our
measure of frequency did not differentiate amount of close holding or ventral contact from
other types of touch such as holding the child’s hand or patting her head. Greater use of
ventral contact and close holding are specific types of touch which have been found to
enhance attachment outcomes (Anisfeld et al., 1990; Waters et al., 1990) so the frequency of
these types of touch may have different implications for secure attachment. Some of the
variance associated with close holding and ventral contact may have been accounted for in
our measure of nurturing touch but this possibility would need to be explored further.
Accounting for such characteristics would speak to the “qualitative” nature of the touch used
rather than its strict “quantity.” In testing the hypothesis regarding frequent touch, our
purpose was to examine the overall amount of physical contact, regardless of its qualitative
nature.
There was no evidence that sensitivity played any moderating role between frequency of
touch (or nurturing touch) and attachment. Nor did sensitivity show any direct association to
attachment. Our study is not alone in finding an absence of any relationship between
sensitivity and attachment (e.g., Seifer et al., 1996). In addition, the strength of the relation-
ships which have been found is weak or modest at best, whether in more naturalistic,
home-based studies using the Attachment Q-Set or in the laboratory-based strange situation
(Solomon & George, 1999; Weinfeld et al., 1999). In fact, DeWolff and van IJzendoorn
(1997) concluded from their meta-analysis that sensitivity does not appear to be an exclusive
or predominant factor in development of secure attachment.
Still, the lack of any relationship between sensitivity and attachment is puzzling, espe-
cially for this population of infants who are at greater risk from the effects of caregiving as
a result of their low birth weight status. It is important to remember that maternal sensitivity
is a multidimensional concept of which our study measured only certain dimensions.
Although the aspects of sensitivity which we examined are commonly held as central
features of maternal sensitivity (i.e. sensitivity to infant cues and responsiveness to infant
distress), the NCAFS measure allows for a limited assessment of specific behaviors which
reflect these aspects of sensitivity. Some of the instruments that have been used to measure
sensitivity in other studies determine a larger scope of behaviors indicative of sensitivity. We
also question whether differences in the developmental point at which sensitivity is measured
may be salient. We assessed the mother’s sensitivity toward the infant at 3 months of age
because we were most interested in the potential effects of her behavior during the infant’s
106 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

early months when touch is a primary component of caregiving. But the role of sensitivity
may strengthen as the infant becomes a more interactive partner and is more aware of the
mother’s behavioral contribution to that partnership. For instance, her sensitivity toward the
infant at 12 months of age may have interacted significantly with maternal touch to influence
attachment. All of these differential features of maternal sensitivity need to be explored in
future research in order to better understand how sensitivity may or may not influence the
effects of touch on attachment outcomes across the developmental span.

6.2. Maternal tactile history

A mother’s state of mind regarding her own tactile history did not appear to modify the
effects of her touch but instead was directly associated with her infant’s security of
attachment. This finding lends credence to attachment theory’s implication that a caregiver’s
own tactile experience as a child constitutes part of her internal working model of attach-
ment. As noted by Ricks (1985), a parent’s own childhood experience appears to tap an
important source of variance in attachment outcomes for his/her child.
The strength of the association which we found between maternal tactile history and infant
attachment attenuated to a trend once the mother’s use of nurturing touch was accounted for
in the model. However, when we removed tactile history from the model, the strength of the
maternal touch variable was diminished. These complex connections between a mother’s use
of touch and her own tactile history suggest that they make some interrelated contribution to
the infant’s attachment. The importance of understanding the tactile antecedents in parent-
child relations (specifically, the parent’s own history of physical affection, discipline or
physical neglect) has been emphasized by other investigators (Egeland et al., 1988). Our
findings indicate that the mother’s felt security about her attachment-related tactile experi-
ence may represent a core component of her internal working model of attachment which is
transmitted to her child, potentially through some facet of her nurturing touch.

7. Conclusion

Maternal touch does appear to play a significant role for low birth weight children in their
development of a secure attachment to the mother. Kissing, hugging and caressing are
significantly associated with attachment outcomes. However, the degree of an infant’s
vulnerability appears to influence whether extensive nurturing touch may support or impede
the infant’s development of a secure attachment. There is some indication that harsh touch
per se may not necessarily place these children at risk for insecure attachment as long as
nurturing touch is also present in the caregiving. But the subsample using harsh touch was
so small that generalizability of their results is extremely limited. The overall frequency of
maternal touch appears to have no link to the ultimate quality of a child’s attachment, even
when the mother shows a high degree of sensitivity in her caregiving. In fact, maternal
sensitivity toward the infant at three months of age appeared to have no relationship to later
attachment in this sample of low birth weight children, neither a direct association nor a
modifying effect on the relationship between maternal touch and attachment. These results
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 107

raise some important questions regarding a) the potential for differential effects of sensitivity
at various points in the child’s development and b) whether specific dimensions or features
of sensitivity may be associated with attachment while others are not.
Infants whose mothers felt more secure about their own childhood tactile experience were
more likely to develop secure attachments. These results lend support to the idea of
intergenerational transmission of working models of attachment and implicate a potential
role for the tactile component of the attachment relationship in development of one’s internal
attachment model.
The results suggest that low birth weight infants who are more vulnerable based on
medical complications, diminished responsiveness and very low birth weight are at signif-
icantly greater risk for attachment problems in general, with the potential for exacerbation of
these problems from the use of extensive nurturing touch. Early identification of these babies
may be critical to their later well being, so that interventions can be offered to enhance the
mother’s use of touch and to assist her in dealing with any unresolved feelings about her own
tactile experience as a child.

References

Ainsworth, M. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: a psychological study
of the strange situation. Hillsdale, NJ: Erlbaum.
Als, H. (1986). A synactive model of neonatal behavioral organization: framework for the assessment of
neurobehavioral development in the premature infant and for support of infants and parents in the neonatal
intensive care environment. Physical & Occupational Therapy in Pediatrics, 6, 3–53.
Anisfeld, E., Casper, V., Nozyce, M., & Cunningham, N. (1990). Does infant carrying promote attachment? An
experimental study of the effects of increased physical contact on the development of attachment. Child
Development, 61, 1617–1627.
Aylward, G., Hatcher, R., Leavitt, L., et al. (1984). Factors affecting neurobehavioral responses of preterm infants
and term conceptional age. Child Development, 55, 1155–1165.
Barnard, K., Hammond, M., Booth, C., Bee, H., Mitchell, S., & Spiker, S. (1989). Measurement and meaning of
parent-infant interaction. In F. Morrison, C. Lord, & D. Keating (Eds.), Applied developmental psychology
(Vol. 3). New York: Academic Press.
Bates, J. E., Maslin, C. A., & Frankel, K. A. (1985). Attachment security, mother-child interaction, and
temperament as predictors of behavior-problem ratings at age three years. Monographs of the Society for
Research in Child Development, 50, 167–193.
Belsky, J., & Cassidy, J. (1994). Attachment and close relationships: an individual-difference perspective.
Psychological Inquiry, 5, 27–30.
Belsky, J., & Rovine, M. (1990). Q-Sort security and first-year nonmaternal care. New Directions for Child
Development, 49, 7–22.
Belsky, J., Rovine, M., & Taylor, D. G. (1984). The Pennsylvania Infant and Family Development Project: III.
The origins of individual differences in infant-mother attachment: maternal and infant contributions. Child
Development, 55, 718 –728.
Belsky, J., & Rovine, M. (1987). Temperament and attachment security in the strange situation: an empirical
rapprochement. Child Development, 58, 787–795.
Bennett, F. (1997). The low birth weight, premature infant. In R. Gross, D. Spiker, & C. Haynes (Eds.), Helping
low birth weight infants (Chapter 1, pp. 3–16). Stanford, CA: Stanford University Press.
Benoit, D., & Parker, K. C. H. (1994). Stability and transmission of attachment across three generations. Child
Development, 65, 1444 –1456.
108 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

Bowlby, J. (1969). Attachment and loss, vol. 1, Attachment. New York: Basic Books.
Bowlby, J. (1973). Attachment and loss, vol. 2, Separation. New York: Basic Books.
Carlson, V., Cicchetti, D., Barnett, D., & Braunwald, K. (1989). Disorganized/disoriented attachment relation-
ships in maltreated infants. Developmental Psychology, 25, 525–531.
Crittenden, P. (1988). Relationships at risk. In J. Belsky & T. Nezwarski (Eds.), Clinical implications of
attachment (pp. 136 –176). Hillsdale, NJ: Erlbaum.
Crockenberg, S. (1981). Infant irritability, mother responsiveness, and social support influences on the security
of infant -mother attachment. Child Development, 52, 857– 865.
De Wolff, M., & van IJzendoorn, M. H. (1997). Sensitivity and attachment: a meta-analysis on parental
antecedents of infant attachment. Child Development, 68, 571–591.
DeMaio-Feldman, D. (1994). Somatosensory processing abilities of very low-birth weight infants at school age.
American Journal of Occupational Therapy, 48, 639 – 645.
Diamond, M. (1990). Evidence for tactile stimulation improving CNS function. In K. Barnard & T. B. Brazelton
(Eds.), Touch: the foundation of experience (pp. 73–96). Madison, CT: International Universities Press.
Egeland, B., & Farber, E. A. (1984). Infant-mother attachment: factors related to its development and changes
over time. Child Development, 55, 753–771.
Egeland, B., Jacobvitz, D., & Sroufe, L. A. (1988). Breaking the cycle of abuse. Child Development, 59,
1080 –1088.
Feldstein, S., Crown, C., Beebe, B., & Jaffe, J. (1995, April). Temporal coordination and the prediction of
mother-infant attachment. Paper presented at the biennial meeting of the Society for Research in Child
Development, Indianapolis, IN.
Francis, P. L., Self, P. A., & Horowitz, F. D. (1987). The behavioral assessment of the neonate: an overview. In
J. D. Osofsky (Ed.), Handbook of infant development (2nd ed., pp. 723–779). New York: John Wiley & Sons.
Frodi, A., & Thompson, R. (1985). Infants’ affective responses in the strange situation: effects of prematurity and
of quality of attachment. Child Development, 56, 1280 –1290.
George, C., & Solomon, J. (1996). Representational models of relationships: links between caregiving and
attachment. Infant Mental Health Journal, 17, 198 –216.
Grossman, K., Grossman, K. E., Gottfried, S., Suess, G., & Unzner, L. (1985). Maternal sensitivity and newborns’
orientation responses as related to quality of attachment in northern Germany. Monographs of the Society for
Research in Child Development, 50, 233–256.
Grossman, K. Grossman, K., & Schwan, A. (1986). Capturing the wider view of attachment: a reanalysis of
Ainsworth’s strange situation. In C. Izard & P. Reed (Eds.), Measuring emotions in infants and children (vol.
2, pp. 124 –171). New York: Cambridge University Press.
Gunnar, M. R., Isensee, J., & Fust, L. S. (1987). Adrenocortical activity and the Brazelton Neonatal Assessment
Scale: moderating effects of the newborn’s biomedical status. Child Development, 58, 1448 –1458.
Harlow, H. (1962). Development of the second and third affectional systems in macaque monkeys. In T. T.
Tourlentes, S. L. Pollack, & H. E. Himwich (Eds.), Research approaches to psychiatric problems. New York:
Grune & Stratton.
Harlow, H. F., & Zimmermann, R. R. (1959). Affectional responses in the infant monkey. Science, 130, 421– 432.
Harrison, L. L., Leeper, J. D., & Yoon, M. (1990). Effects of early parent touch on preterm infants’ heart rates
and arterial oxygen saturation levels. Journal of Advanced Nursing, 15, 877– 885.
Hazan, C., & Shaver, P. (1994). Deeper into attachment theory: author’s response. Psychological Inquiry, 5 (1),
68 –79.
Hofer, M. (1981). The roots of human behavior. San Francisco: W.H. Freeman.
Hunziker, U. A., & Barr, R. G. (1986). Increased carrying reduces infant crying: a randomized controlled trial.
Pediatrics, 77, 641– 648.
Kerns, K. (1994). A longitudinal examination of links between mother-child attachment and children’s friend-
ships in early childhood. Journal of Social and Personal Relationships, 11, 379 –381.
La Freniere, P., Provost, M., & Dubeau, D. (1992). From an insecure base: parent-child relations and internalizing
behavior in the preschool. Early Development and Parenting, 1, 137–148.
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 109

Lewis, M., & Feiring, C. (1989). Infant, mother, and mother-infant interaction behavior and subsequent attach-
ment. Child Development, 60, 831– 837.
Leyendecker, B., Lamb, M. E., Fracasso, M. P., & Scholmerich, A. (1997). Playful interaction and the antecedents
of attachment: a longitudinal study of Central American and Euro-American mothers and infants. Merrill-
Palmer Quarterly, 43, 24 – 47.
Littman, B., & Parmelee, A. H. (1978). Medical correlates of infant development. Pediatrics, 61, 470 – 474.
Main, M. (1981). Avoidance in the service of attachment. In K. Immelman, G. Barlow, M. Main, & L.
Petrinovitch (Eds.), The Bielefeld Interdisciplinary Project (pp. 651– 693.) NY: Cambridge University Press.
Main, M. (1990). Parental aversion to infant-initiated contact is correlated with the parent’s own rejection during
childhood: the effects of experience on signals of security with respect to attachment. In K. E. Barnard & T. B.
Brazelton (Eds.), Touch: the foundation of experience: full revised and expanded proceedings of Johnson &
Johnson Pediatric Round Table X. Clinical infant reports. (pp. 461– 495). Madison, CT: International
Universities Press.
Main, M., & Hesse, E. (1990). Parents’ unresolved traumatic experiences are related to infant disorganized
attachment status: is frightened and/or frightening parental behavior the linking mechanism? In M. T.
Greenberg & D. Cicchetti (Eds.), Attachment in the preschool years: theory, research, and intervention (pp.
161–182). Chicago: University of Chicago.
Main, M., & Solomon, J. (1986). Discovery of an insecure-disorganized/disoriented attachment pattern. In T. B.
Brazelton & M. W. Yogman (Eds.), Affective development in infancy (pp. 95–124). Norwood, NJ: Ablex.
Main, M., & Stadtman, J. (1981). Infant response to rejection of physical contact by the mother. Journal of the
American Academy of Child Psychiatry, 20, 292–307.
Main, M., & Weston, D. R. (1981). The quality of the toddler’s relationship to mother and to father: related to
conflict behavior and the readiness to establish new relationships. Child Development, 52, 932–940.
Mangelsdorf, S., Plunkett, J., Dedrick, C., Berlin, M., Meisels, S., McHale, J., & Dichtellmiller, M. (1996).
Attachment security in very low birth weight babies. Developmental Psychology, 32, 914 –920.
Miyake, K., Chen, S., & Campos, J. J. (1985). Infant temperament, mother’s mode of interaction, and attachment
in Japan: an interim report. Monographs of the Society for Research in Child Development, 50, 276 –297.
Network, N. E. C. C. R. (1997). The effects of infant child care on infant-mother attachment security: results of
the NICHD study of early child care. Child Development, 68, 860 – 879.
O’Connor, M. J., Sigman, M. D., & Kasari, C. (1992). Attachment behavior of infants exposed prenatally to
alcohol: mediating effects of infant affect and mother-infant interaction. Development & Psychopathology, 4,
243–256.
Parmelee, A. H., Kopp, C. B., & Sigman, M. (1976). Selection of developmental assessment techniques for
infants at risk. Merrill-Palmer Quarterly, 22, 177–199.
Pederson, D. R., & Moran, G. (1995). A categorical description of infant-mother relationships in the home and
its relation to Q-sort measures of infant-mother interaction. Monographs of the Society for Research in Child
Development, 60, 111–132.
Plunkett, J., Meisels, S., Stiefel, G., & Pasick, P. (1986). Patterns of attachment among preterm infants of varying
biological risk. Journal of the American Academy of Child Psychiatry, 25 (6), 794 – 800.
Posada, G., Waters, E., Crowell, J. A., & Lay, K.-L. (1995). Is it easier to use a secure mother as a secure base?
Attachment Q-sort correlates of the adult attachment interview. Monographs of the Society for Research in
Child Development, 60, 133–145.
Ricks, M. (1985). Social transmission of parental behavior: attachment across generations. In I. Bretherton & E.
Waters (Eds.), Growing points of attachment theory and research. Monographs of the Society for Research in
Child Development, 50 (1–2, serial #9), 211–227.
Schneider-Rosen, K., & Cicchetti, D. (1984). The relationship between affect and cognition in maltreated infants:
quality of attachment and the development of visual self-recognition. Child Development, 55, 648 – 658.
Scott, D., Bauer, C., Draemer, H., & Tyson, J. (1997). The neonatal health index. In R. Gross, D. Spiker, & C.
Haynes (Eds.), Helping low birthweight premature babies (pp. 341–357). Stanford: Stanford University Press.
Seifer, R., & Schiller, M. (1995). The role of parenting sensitivity, infant temperament, and dyadic interaction in
110 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111

attachment theory and assessment. Monographs of the Society for Research in Child Development, 60,
146 –174.
Seifer, R., Schiller, M., Sameroff, A. J., Resnick, S., & Riordan, K. (1996). Attachment, maternal sensitivity, and
infant temperament during the first year of life. Developmental Psychology, 32, 12–25.
Smith, P. B., & Pederson, D. R. (1988). Maternal sensitivity and patterns of infant-mother attachment. Child
Development, 59, 1097–1101.
Solomon, J., & George, C. (1999). The measurement of attachment security in infancy and childhood. In J.
Cassidy & P. R. Shaver (Eds.), Handbook of attachment: theory, research, and clinical applications (pp.
287–316). New York: Guilford Press.
Steele, H., Steele, M., & Fonagy, P. (1996). Associations among attachment classifications of mothers, fathers,
and their infants. Child Development, 67, 541–555.
Strayer, F., Verissimo, M., Vaughn, B., & Howes, C. (1995). A quantitative approach to the description and
classification of primary social relationships. In E. Waters, B. Vaughn, G. Posada, & K. Kondo-Ikemura
(Eds.), Caregiving, cultural and cognitive perspectives on secure-base behavior and working models. Mono-
graphs of the Society for Research in Child Development, 60 (2–3, serial #244), 49 –70.
Sumner, G., & Spietz, A. (1994). NCAST caregiver/parent-child interaction feeding manual. Seattle: NCAST
Publications, University of Washington.
Teti, D., & McGourty, S. (1996). Using mothers versus trained observers in assessing childrens’ secure base
behavior: theoretical and methodological considerations. Child Development, 67, 597– 605.
Van Beek, Y., Hopkins, B., & Hoeksma, J. (1994). The development of communication in pre-term infant-mother
dyads. Behavior, 129 (1–2), 35– 61.
Van den Boom, O. (1994). The influence of temperament and mothering on attachment and exploration: an
experimental manipulation of sensitive responsiveness among lower class mothers with irritable infants. Child
Development, 65, 1457–1477.
Van der Kolk, B. A., & Fisler, R. E. (1994). Childhood abuse and neglect and loss of self-regulation. Bulletin of
the Menninger Clinic, 58, 145–168.
Van IJzendoorn, M. H., Goldberg, S., Kroonenberg, P. M., & Frenkel, O. J. (1992). The relative effects of
maternal and child problems on the quality of attachment: a meta-analysis of attachment in clinical samples.
Child Development, 63, 840 – 858.
Van IJzendoorn, M. H. (1995). Adult attachment representations, parental responsiveness, and infant attachment:
a meta-analysis on the predictive validity of the Adult Attachment Interview. Psychological Bulletin, 117,
387– 403.
Vaughn, B. E., & Bost, K. K. (1999). Attachment and temperament: redundant, independent, or interacting
influences on interpersonal adaptation and personality development? In J. Cassidy & P. R. Shaver (Eds.),
Handbook of attachment: theory, research, and clinical applications (pp. 198 –225). New York: Guilford
Press.
Vaughn, B. E., & Waters, E. (1990). Attachment behavior at home and in the laboratory: Q-sort observations and
Strange Situation classifications of one-year-olds. Child Development, 61, 1965–1973.
Vondra, J. I., Shaw, D. S., & Kevenides, M. C. (1995). Predicting infant attachment classification from multiple,
contemporaneous measures of maternal care. Infant Behavior & Development, 18, 415– 425.
Ward, M. J., & Carlson, E. A. (1995). Associations among adult attachment representations, maternal sensitivity,
and infant-mother attachment in a sample of adolescent mothers. Child Development, 66, 69 –79.
Waters, E., & Deane, K. E. (1985). Defining and assessing individual differences in attachment relationships:
Q-methodology and the organization of behavior in infancy and early childhood. Monographs of the Society
for Research in Child Development, 50, 41– 65.
Waters, E., Kondo-Ikemura, K., Posada, G., & Richters, J. E. (1990). Learning to love: mechanisms and
milestones. In M. Gunner & L. A. Sroufe (Eds.), Self Processes and Development Minnesota Symposia on
Child Psychology (Vol. 23, pp. 217–255). Hillsdale, NJ: Erlbaum.
Weinfield, N. S., Sroufe, L. A., Egeland, B., & Carlson, E. A. (1999). The nature of individual differences in
infant-caregiver attachment. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: theory, research,
and clinical applications (pp. 68 – 88). New York: Guilford Press.
 S.J. Weiss et al. / Infant Behavior & Development 23 (2000) 91–111 111

Weiss, S. J. (1990). Parental touching: correlates of a child’s body concept and body sentiment. In K. E. Barnard
& T. B. Brazelton (Eds.), Touch: the foundation of experience: full revised and expanded proceedings of
Johnson & Johnson Pediatric Round Table X. Clinical infant reports. (pp. 425– 459). Madison, CT: Inter-
national Universities Press.
Weiss, S. J. (1992). Measurement of the sensory qualities in tactile interaction. Nursing Research, 41, 82– 86.
Weiss, S. J. (1993). Familial touch and perceptual reactivity of children with severe mental illness. New
Directions in Mental Health Services, 58, 53– 64.
Weiss, S. J. (1999). The effect of maternal touch on physical health and cognitive development of low birth
weight infants. Presentation at the meeting of the Society for Research in Child Development. Albuquerque,
NM.