Ni Hms 423794
Ni Hms 423794
Ni Hms 423794
Author Manuscript
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Published in final edited form as:
Annu Rev Clin Psychol. 2011 ; 7: 269–295. doi:10.1146/annurev-clinpsy-032210-104540.
Abstract
Understanding the association between personality and depression has implications for elucidating
etiology and comorbidity, identifying at-risk individuals, and tailoring treatment. We discuss
seven major models that have been proposed to explain the relation between personality and
depression, and we review key methodological issues, including study design, the heterogeneity of
mood disorders, and the assessment of personality. We then selectively review the extensive
empirical literature on the role of personality traits in depression in adults and children. Current
$watermark-text
Keywords
traits; temperament; mood disorders; neuroticism; extraversion
INTRODUCTION
$watermark-text
The hypothesis that depression is linked to personality can be traced to antiquity, when
Hippocrates, and later Galen, argued that particular “humors” were responsible for specific
personality types and forms of psychopathology. In this article, we discuss the major
conceptual models that have been proposed to explain the association between personality
and depression, comment on some important methodological issues, and selectively review
the empirical literature. Due to space limitations, we limit our review to nonbipolar forms of
depression.
This literature has developed along several distinct lines: (a) early clinical psychiatrists’
descriptions of affective temperaments; (b) research on the structure and neurobiology of
personality; (c) psychoanalytic and cognitive-behavioral theory and observations; and (d)
developmental psychologists’ work on temperament. In recent years, there has been
substantial convergence between these lines of work, and it is increasingly possible to view
them within a single integrative framework. Understanding the associations between
personality and depression has a number of potentially important implications for research
and practice. First, personality traits associated with emotional experience, expression, and
regulation may be intermediate phenotypes that provide more tractable targets for genetic
and neurobiological research than depressive diagnoses (Canli 2008). Second, personality
may be useful in identifying more homogeneous subgroups of depressive disorders that
differ in developmental trajectories and etiological influences (e.g., Beck 1983). Third,
tracing the pathways between personality and depressive disorders can help elucidate more
proximal processes involved in the development of mood disorders (Compas et al. 2004,
Klein et al. 2008a, Lahey 2009). Fourth, personality may be useful in tailoring treatment
(Zinbarg et al. 2008) and predicting treatment response (Quilty et al. 2008a). Fifth,
temperament/personality may provide a means to identify at-risk individuals who could
$watermark-text
benefit from prevention and early intervention efforts (Kovacs & Lopez-Duran 2010).
Finally, there is substantial comorbidity between depressive disorders and other forms of
psychopathology. Some personality traits, such as neuroticism, are associated with multiple
psychiatric conditions. Thus, personality could help explain patterns of comorbidity and
point toward more etiologically relevant classification systems (Brown & Barlow 2009,
Kotov et al. 2007, Watson 2009).
Second, a variety of personality classifications have been proposed over the past century, but
$watermark-text
in the 1980s they were integrated in a consensus taxonomy, the Five-Factor Model (FFM).
The FFM recognized that personality is ordered hierarchically from a large number of
specific traits to five general characteristics (Digman 1994, Goldberg 1993, Markon et al.
2005). These “Big Five” traits are neuroticism, extraversion, conscientiousness,
agreeableness, and openness to experience. Importantly, the FFM can be further reduced to
three dimensions of negative emotionality, positive emotionality, and disinhibition versus
constraint that form the next level of the personality hierarchy (Clark & Watson 1999,
Markon et al. 2005). This “Big Three” model is used in studies of temperament as well as
personality, although disinhibition is often labeled as effortful control in the child literature
(Caspi & Shiner 2006, Rothbart & Bates 2006). The Big Five and Big Three schemes are
closely related, with neuroticism being essentially identical to negative emotionality and
extraversion corresponding to positive emotionality (Clark & Watson 1999, Markon et al.
2005); we refer to these two dimensions as neuroticism/negative emotionality (N/NE) and
extraversion/positive emotionality (E/PE), respectively. Disinhibition does not have an exact
counterpart in the FFM but instead reflects a combination of low conscientiousness and low
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 3
agreeableness. Finally, openness to experience is outside the territory covered by the Big
Three.
Third, there is increasing recognition that temperament and personality are not a fixed, static
set of characteristics, but rather are dynamic constructs that develop over the lifespan and
change in response to maturation and life circumstances (Fraley & Roberts 2005, Rothbart
& Bates 2006). For example, although the rank-order stability of most personality traits is in
the moderate range, it increases over the course of development (Roberts & DelVecchio
2000). In addition, mean levels of conscientiousness and some facets of E/PE increase, and
levels of N/NE decrease, over time, particularly in young adulthood (Roberts et al. 2006). A
number of processes contribute to stability and change of personality. For example, genes
$watermark-text
are a major influence on stability (Krueger & Johnson 2008, Kandler et al. 2010). In
addition, people often select, create, and construe environments in ways that reinforce and
maintain their initial trait dispositions (Caspi & Shiner 2006). However, life stressors and
major shifts in social roles and relationships can contribute to personality change (Fraley &
Roberts 2005, Kandler et al. 2010). We consider the implications of these processes for the
relation between personality and depression below.
(b) personality and depressive disorders form a continuous spectrum; (c) personality is a
precursor of depressive disorders; (d) personality predisposes to developing depressive
disorders; (e) personality has pathoplastic effects on depression; (f) personality features are
state-dependent concomitants of depressive episodes; and (g) personality features are
consequences (or scars) of depressive episodes. The distinctions between some of these
accounts are subtle (cf. Kendler & Neale 2010), and other models, as well as combinations
of these scenarios, are plausible. However, these seven models provide a useful conceptual
framework for approaching the issue.
These models can be divided into three groups. The first three models (common cause,
continuum/spectrum, and precursor) view personality and depression as having similar
causal influences but do not see one as having a causal influence on the other. The fourth
$watermark-text
and fifth models (predisposition and pathoplasticity) hold that personality has causal effects
on the onset or maintenance of depression. Finally, the sixth and seventh models
(concomitants and consequences) view depression as having a causal influence on
personality. These models, and their unique predictions, are summarized in Table 1.
The common cause model views personality and depressive disorders as distinct entities that
arise from the same, or at least an overlapping, set of etiological processes. From this
perspective, personality and depression are not directly related; rather, the association is due
to a shared third variable. The common cause model would be supported by evidence that
personality traits and depression have shared etiological influences.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 4
association between the trait and disorder should be fairly specific because they are on the
same continuum.1 Moreover, this association is expected to be nonlinear, so that almost
nobody below the definitional threshold on the trait has the diagnosis but nearly everyone
above the threshold meets the criteria. Thus, the continuum/spectrum model would be
supported by evidence that the trait and depression are associated with the same etiological
influences and that the trait-disorder relationship is fairly specific and nonlinear.
phenomenologic similarity between the relevant trait and depression. However, the
precursor model differs from both of these other models in that it assumes a particular
developmental sequence, with the personality traits being evident prior to the onset of
depressive disorder. In other words, both the common cause and continuum/spectrum
models assume a fixed clinical expression as traits or disorder, whereas the precursor model
implies escalation from traits to disorder within individuals over time. Support for the
precursor model would come from evidence that the trait and depression are associated with
the same etiological influences and that individuals with high levels of the trait are at
increased risk for developing the disorder over time.2
The common cause, continuum/spectrum, and precursor models do not posit causal relations
between personality and depression. In contrast, the predisposition model holds that
personality plays a causal role in the onset of depression. However, the predisposition model
$watermark-text
overlaps with the precursor model in that both propose that the relevant traits are evident
prior to the onset of depressive disorder. The major difference between these two accounts is
that the precursor model assumes that personality and depression derive from the same set of
etiological processes, but the predisposition model posits that the processes that underlie
personality differ from those that lead to depression. Thus, the predisposition account
implies a complex interplay among risk factors involving moderation and/or mediation, and
this is what distinguishes it from the precursor model.3 The most common example—the
diathesis-stress model—conceptualizes personality as the diathesis and stress as a moderator
that precipitates the onset of depressive disorder. Alternatively, stress may be a mediator, so
that personality vulnerability leads to negative experiences (e.g., interpersonal rejection, job
loss), which in turn increase the probability of a depressive episode. A second difference
between these models is that the predisposition model does not assume any
$watermark-text
1There may not be complete specificity owing to diagnostic heterogeneity. As discussed below, depression is a heterogeneous
disorder with multiple etiological pathways (equifinality). A personality trait may be part of only one of the pathways. In contrast,
multifinality, in which the trait is associated with multiple disorders, is less consistent with the continuum/spectrum view.
2Application of the continuum/spectrum and precursor models to depressive disorders is not straightforward. Personality traits are
relatively stable, whereas depression is often episodic. Existing formulations of the continuum/spectrum model have not explained
how stable trait characteristics manifest as an episodic illness. Similarly, the precursor model does not account for why a stable trait
would subsequently develop into a nonstable depressive state. Thus, the continuum/spectrum and precursor models may provide a
better explanation for chronic than episodic forms of depression.
3Although moderating and mediating variables play an explicit and central role in the predisposition model, it should be
acknowledged that they are not incompatible with the precursor account. That is, the escalation from personality traits to depressive
disorders in the precursor model implies that additional variables (e.g., maturational or environmental factors) must be involved to
precipitate the change.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 5
The pathoplasticity model is similar to the predisposition model in that it also views
personality as having a causal influence on depressive disorder. However, rather than
contributing to the onset of depression, the pathoplasticity model posits that personality
influences the expression of the disorder after onset. This influence can include the severity
or pattern of symptomatology, course, and response to treatment. The pathoplasticity model
would be supported by evidence that personality explains variation among depressed
individuals in their clinical presentation or outcome.
The final two models also assume that there is a causal relation between personality and
depression. However, these models reverse the direction of causality. In the concomitants
(or state-dependent) model, assessments of personality are colored, or distorted, by the
$watermark-text
individual’s mood state. This model implies that personality returns to its baseline form after
recovery from the episode. In contrast, the consequences (or scar) model holds that
depressive episodes have an enduring effect on personality, such that changes in personality
persist after recovery. These models would be supported by evidence that depression alters
levels of personality traits, either concurrently (concomitants model) or over the longer-term
(consequences model).
experiences (Ormel & de Jong 1999). In fact, personality generally tends to change in a
more adaptive direction with age (Roberts et al. 2006), although this pattern is not universal
(Johnson et al. 2007). This may help to explain why the probability of first-episode
depression peaks in adolescence, as trait deviance is more common at that age.
Similarly, the predisposition model can be expanded to recognize personality change. This
dynamic predisposition model (Ormel & de Jong, 1999, Ormel et al. 2001) acknowledges
transactions between personality and the environment and integrates them with the
environmental moderation and mediation mechanisms of the classic predisposition model. In
the environmental moderation version of this account, negative life experiences influence
not only depression onset but also levels of trait vulnerability (Middledorp et al. 2008). This
increase in personality liability may then lead to additional life stress. If this vicious cycle is
perpetuated unchecked, personality liability would continue to increase, and at some point, a
4This could also be called the dynamic continuum model because once the dynamic element is introduced, it becomes virtually
impossible to distinguish the continuum/spectrum and precursor models.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 6
negative life event could overwhelm coping capabilities and elicit a depressive disorder.
Importantly, and in contrast to the dynamic precursor model, in this account maladaptive
traits alone are not sufficient to cause depression, and an environmental trigger is necessary.
The vicious cycle of increasing trait vulnerability and stress exposure does not necessarily
indicate that personality per se influences depression onset. Indeed, certain traits may
increase stress exposure but have no effect on depression otherwise (e.g., it is possible that
low conscientiousness does not cause depression directly but leads to depressogenic
experiences, such as academic difficulties, job loss, and relationship problems; Roberts et al.
2007), consistent with the environmental-mediation pathway.
$watermark-text
Dynamic models offer richer and more complete accounts of the role of personality in the
onset of depression. Moreover, it is important to recognize that depressive disorders have
been linked to multiple traits (as reviewed below), and it is likely that different personality
characteristics contribute through different pathways.
METHODOLOGICAL ISSUES
A number of methodological issues must be considered in evaluating the relation between
personality and mood disorders, including (a) study design, (b) heterogeneity of depressive
disorders, and (c) assessment of personality.
Study Design
A number of research designs can be useful in studying the relation between personality and
$watermark-text
particular personality traits predict the later onset of depressive disorder. Although no single
design can distinguish among these four models, the combination of designs can bolster the
case for particular accounts. For example, finding substantial common genetic variance in
twin studies, but no evidence of developmental sequencing in longitudinal studies, would
support the common cause and continuum/spectrum models. In turn, these two models could
be compared by examining the specificity of the association between trait and disorder and
whether there is a nonlinear relation between trait level and probability of disorder. On the
other hand, if there were evidence of developmental sequencing in longitudinal studies as
well as substantial common genetic variance in twin studies (or overlap of other etiological
factors in other designs), it would support the precursor model (particularly if the trait was
also phenomenologically similar to depression). In contrast, developmental sequencing but
less shared genetic (or other etiological) variance would support the predisposition model.
Also crucial for the predisposition model is evidence from longitudinal studies
demonstrating that other variables (e.g., life stress) moderate or mediate the association
between personality and subsequent depression.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 7
The concomitants model can be tested through cross-sectional studies comparing persons
who are currently depressed, persons who have recovered from depressive episodes, and
healthy controls. An even better approach is to conduct longitudinal studies assessing
individuals when they are in a depressive episode and again after they have recovered. If
personality measures are abnormal during depressive episodes but not after recovery, it
would suggest that they are concomitants of the depressed state. Multilevel analyses can also
be used to separate personality variance into trait and state components and to test whether
state variance is associated with concurrent measures of depression (e.g., Duncan-Jones et
al. 1990).
The consequences (or scar) hypothesis can be evaluated by assessing persons before and
after a first depressive episode. If personality deviance is much greater after the episode has
$watermark-text
Testing dynamic theories requires longitudinal data with at least three assessment points.
These assessments should measure relevant contextual factors (e.g., life stress) in addition to
depression and personality to allow the examination of dynamic and transactional effects.
Multilevel modeling and structural equation modeling offer powerful approaches to
evaluating such effects with longitudinal data.
relation between personality and depression differ for different forms of depression. The
current classification system for depressive disorders is based on clinical features and is a
poor approximation of etiological distinctions. Nonetheless, it is important to consider
whether the role of personality varies as a function of the specific depressive diagnosis (e.g.,
major depressive disorder, dysthymic disorder), subtype (e.g., psychotic, melancholic,
atypical), and clinical characteristics such as age of onset, recurrence, and chronicity. Failure
to take heterogeneity into account may obscure important personality-depression
associations. Conversely, personality may provide a basis for identifying more
homogeneous subgroups within the depressive disorders. Unfortunately, few studies of
personality and depressive disorders have attempted to take this heterogeneity into account.
It is important to note, however, that associations between personality and specific subtypes
and clinical characteristics do not necessarily indicate etiological heterogeneity. Instead,
they could reflect pathoplasticity, in which personality influences symptom presentation
and/or course, but the primary etiological process is the same, or they could reflect
differential severity of subtypes that results in quantitative differences in their trait profiles.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 8
Assessment
Temperament/personality can be assessed using a variety of methods, including self-report
inventories, semistructured interviews, informants’ reports, and observations in naturalistic
settings and the laboratory. Unfortunately, most of the literature examining the association
between personality and depressive disorders has assessed personality via self-report. This is
potentially problematic because self-reports of personality can be complicated by current
mood state, limited insight, response styles, and the difficulty of distinguishing traits from
the effects of stable environmental contexts (Chmielewski & Watson 2009). In addition,
when the same individual provides information on both personality and depression, as has
been the case in almost all studies in this area, common method variance can inflate
associations. Hence, there is a need for greater use of informant report and observational
measures.
$watermark-text
A second issue concerns the overlap between some personality constructs and
psychopathology (Lahey 2004). For example, many items on N/NE scales are similar to
depressive symptoms (Ormel et al. 2004b). This can inflate associations between measures
of personality and depression. On the other hand, personality and symptom assessments
usually have different time frames, with trait scales reflecting long-standing patterns and
depression measures tapping more recent experiences (e.g., past week, past month). This
trait versus state distinction parallels that between personality and other related constructs.
For example, measures of N/NE and negative affect have nearly identical content but are
distinguished by their time frames (Watson 2000). Thus, the degree to which this content
overlap threatens the validity of personality-psychopathology research depends, at least in
part, on the duration/chronicity of the disorders of interest. The extent to which this is a
concern also depends on one’s model of personality-depression relations. From the
$watermark-text
AFFECTIVE TEMPERAMENTS
The classical European descriptive psychopathologists in the late-nineteenth and early-
twentieth centuries observed that many patients with mood disorders, as well as their
relatives, exhibited particular patterns of premorbid personalities that appeared to be
attenuated versions of their illnesses. For example, Kraepelin (1921) described four patterns
of personality that he considered the “fundamental states” underlying manic-depressive
illness: depressive, manic, irritable, and cyclothymic temperament. He believed that these
were precursors or “rudimentary forms” of the major mood disorders. Schneider (1958)
described similar types; however, he viewed them as personality disorders that were not
necessarily related to the mood disorders. Two variants of these types, cyclothymic disorder
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 9
and dysthymic disorder, are included as mood disorder diagnoses in the Diagnostic and
Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; Am. Psychiatr. Assoc.
1994). However, these disorders are defined as fairly severe conditions, with the criteria
emphasizing symptomatology rather than personality traits. As a result, these categories
appear to be limited to the more severe, symptomatic manifestations of the affective
temperaments described by Kraepelin and Schneider (Akiskal 1989).
On the basis of Kraepelin’s and Schneider’s descriptions, Akiskal (1989) proposed formal
criteria for the affective temperament types, and he and his colleagues developed interview
and self-report measures of these constructs that have been applied in a number of settings
and cultures (e.g., Akiskal et al. 2005). Akiskal’s work also provided the basis for including
$watermark-text
Data on the nature of the relation between depressive personality and depressive disorders
are consistent with most of the causal models described above, illustrating the complexity of
the associations between personality and depression. The strongest support for the common
$watermark-text
cause and continuum/spectrum models derives from twin and family studies. In a large twin
study, Ørstavik et al. (2007) found that depressive personality and major depressive disorder
(MDD) shared substantial genetic variance, although there was evidence for unique genetic
factors as well. Family studies indicate that individuals with depressive personality have an
increased rate of mood disorders in their first-degree relatives (e.g., Klein & Miller 1993). In
addition, patients with MDD, particularly those with chronic forms of depression, have
elevated levels of depressive personality traits in their first-degree relatives (Klein 1999).
studies, these findings provide compelling support for the precursor model. In addition, in
light of the conceptual issue regarding traits and states raised above for the continuum/
spectrum and precursor models, it is noteworthy that depressive personality is most closely
associated with chronic forms of depression at both the family and individual levels.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 10
Depressive personality also appears to have a pathoplastic effect on the course of depressive
disorders, predicting poorer outcomes and response to treatment (Laptook et al. 2006, Ryder
et al. 2010). Moreover, Rudolph & Klein (2009) recently reported preliminary support for
the consequences model, at least in youth. They found that in a sample of early adolescents,
higher levels of depressive symptoms predicted an increase in depressive personality traits
12 months later. Finally, the limited evidence available suggests that semi-structured
interview assessments of depressive personality traits are not influenced by a depressive
episode (Klein 1990), arguing against the concomitants model.
temperamental processes that originate in early childhood, as their defining features include
a number of developmentally complex cognitive and interpersonal characteristics. Instead,
these temperament types are more likely to be intermediate outcomes that reflect the
interaction of more basic temperament traits that are elaborated over development in
conjunction with early socialization and other environment influences.
by high levels of N/NE and low levels of E/PE. A large number of cross-sectional studies
have evaluated these relations as well as the links between depression and the other FFM
dimensions. Kotov et al. (2010) recently conducted a meta-analysis of this literature, which
revealed that MDD is associated with very high N/NE (Cohen’s d =1.33) and low
conscientiousness (d =−0.90). The link to low E/PE was more modest (d =−0.62) and
inconsistent, with some studies finding positive effects. The associations with the other two
traits were weak and unremarkable. The N/NE finding is consistent with expectations, but
the effect for E/PE was smaller and that for conscientiousness was larger than anticipated.
Dysthymic disorder exhibited a more extreme profile with remarkably strong and consistent
links to E/PE (d =−1.47), N/NE (d =1.93), and conscientiousness (d =−1.24). This is not
surprising as dysthymic disorder is thought to be more trait-like than MDD, and a greater
contribution from personality might be expected.
To determine whether the observed personality links are specific to depression, Kotov et al.
(2010) also examined personality profiles of anxiety disorders. They found that with the
exception of specific phobia, which had relatively weak associations with all five traits, all
anxiety disorders showed stronger effects on N/NE, E/PE, and conscientiousness (average d
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 11
=1.91, −1.05, −1.02, respectively) than did MDD. Several also scored above dysthymic
disorder on N/NE. Dysthymia had stronger associations than anxiety disorders on the other
two traits, but the differences were slight.
It is conceivable that more specific associations were not evident because these analyses
focused on broad personality dimensions. Narrow traits that comprise the general
dimensions may have stronger associations with depressive disorders. Indeed, self-harm—a
component of N/NE that reflects propensity to self-deprecation and self-injury—was found
to contribute to depression even after controlling for the broad traits, and this effect was
specific relative to other common mental disorders (Watson et al. 2006). With regard to E/
PE, evidence is emerging that the positive affectivity facet, but not the sociability/
$watermark-text
extraversion facet, is related to depression (Durbin et al. 2005, Naragon-Gainey et al. 2009).
This may explain the surprisingly modest association between MDD and E/PE, if this
general trait includes much variance not relevant to depression. Thus, facet-level research
promises to yield stronger and more specific evidence of personality-depression links.
found that individuals with MDD report higher levels of N/NE when they are depressed than
when they are not depressed (Hirschfeld et al. 1983b, Kendler et al. 1993, Ormel et al.
2004a). In contrast, the evidence for mood state effects on E/PE is weaker and less
consistent (de Fruyt et al. 2006, Kendler et al. 1993, Morey et al. 2010). However, the
influence of mood state on personality should not be overstated. Even though levels of N/NE
decline significantly after remission from a depressive episode (i.e., absolute stability),
individuals’ relative positions with respect to levels of N/NE (i.e., rank-order stability) tend
to be moderately well preserved (de Fruyt et al. 2006, Morey et al. 2010). Moreover, clinical
trials suggest that changes in depressive symptoms are not necessarily accompanied by
changes in personality (Quilty et al. 2008b, Tang et al. 2009).
studies used remission designs, comparing patients who had recovered from a depressive
episode to never-depressed controls or population norms on self-rated personality traits.
These studies found that E/PE is significantly lower in formerly depressed patients than in
healthy controls (Hirschfeld et al. 1983a, Reich et al. 1987), arguing against the
concomitants model and in favor of the precursor, predisposition, and/or consequences
models. However, the results for N/NE were less consistent (Hirschfeld et al. 1983a, Reich
et al. 1987). This inconsistency may be due to a number of factors, including insufficiently
stringent criteria for recovery, thereby possibly confounding personality and residual
symptoms; using normative data collected by other investigators, which may introduce
demographic and sociocultural differences between the formerly depressed and comparison
samples; and selection effects, as N/NE is associated with a poorer course (discussed below)
and thus samples of remitted depressives may include a disproportionate number with low
levels of this trait.
Personality before and after a depressive episode: Several studies have tested the
consequences (or scar) hypothesis by comparing personality measures in depressed
individuals before and after a MDD episode. The results of these studies have been
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 12
inconsistent. Kendler and colleagues reported increases in N/NE (but not E/PE) after a
depressive episode in two separate samples (Fanous et al. 2007, Kendler et al. 1993);
however, other studies have found that N/NE and E/PE do not change from before to after a
MDD episode (e.g., Ormel et al. 2004a, Shea et al. 1996). Importantly, the studies reporting
scarring used less stringent criteria for recovery and shorter follow-ups, suggesting that the
findings may be due to residual symptoms and/or that the scars dissipate over time.
results have been mixed, with some studies reporting higher N/NE and/or lower E/PE in the
never-depressed relatives of probands with mood disorders, and other studies reporting no
differences. However, interpretation of these studies is complicated by two factors. First,
personality traits may not play the same role in risk for depression among familial as
nonfamilial forms of depression. Second, there may be selection biases in samples using
well relatives who are already partly through the risk period for mood disorder. Thus, those
relatives with the strongest personality vulnerabilities may have already developed the
disorder and be excluded from the study.
Twin studies: As discussed above, a valuable approach to testing the common cause,
continuum/spectrum, and precursor models is through twin studies. These studies indicate
that there are substantial associations between the liabilities for N/NE and MDD, but only
$watermark-text
weak associations between the genetic liabilities for E/PE and MDD (Fanous et al. 2007;
Kendler et al. 1993, 2006).
Prospective longitudinal studies: The most direct approach to testing the precursor and
predisposition models is to conduct prospective studies of personality in never-depressed
participants to determine whether personality characteristics predict the subsequent onset of
depressive disorders. Several studies using large community samples have reported that
higher levels of N/NE predict the onset of first lifetime MDD episodes (de Graaf et al. 2002;
Fanous et al. 2007; Kendler et al. 1993, 2006; Ormel et al. 2004a). In addition, several
studies using measures of other traits that overlap with N/NE or its facets have reported
similar findings (Hirschfeld et al. 1989, Rorsman et al. 1993). Although there is some
evidence that E/PE predicts the first onset of MDD (Kendler et al. 2006, Rorsman et al.
1993), it is much weaker, and several studies have failed to find an association (Fanous et al.
$watermark-text
Personality and the subsequent course of depression: Finally, there is evidence that both
N/NE and E/PE have pathoplastic influences on the course of depression after the onset of
the disorder. For example, many studies have reported that higher N/NE and lower E/PE
predict a poorer course and response to treatment, although the findings regarding E/PE are
slightly less consistent (de Fruyt et al. 2006, Duggan et al. 1990, Morris et al. 2009, Quilty et
al. 2008a, Tang et al. 2009). As noted above, however, these findings are also consistent
with diagnostic heterogeneity, such that personality dysfunction is a marker for a more
severe or etiologically distinct group. Indeed, there is evidence that the nonmelancholic
subtype is characterized by more vulnerable personality styles than is melancholia and that
chronic depressions are associated with higher N/NE and lower E/PE than is nonchronic
MDD (Klein 2008, Kotov et al. 2010).
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 13
1999, van Os & Jones 1999). N/NE shows reciprocal relations with a range of significant
life experiences, such as initiation and break-up of a committed relationship, relationship
quality, occupational attainment, and financial security (Neyer & Lehnart 2007, Roberts et
al. 2003, Scollon & Diener 2006). Furthermore, N/NE has been repeatedly implicated in the
generation of stressful life events (Kercher et al. 2009, Lahey 2009, Middeldorp et al. 2008),
which suggests an environmentally mediated relationship between this trait and depression.
The environmentally moderated mechanism has also received support, as several studies
found that N/NE interacts with stressful life events to predict first onset of major depression
(Kendler et al. 2004, Ormel et al. 2001, van Os & Jones 1999).
E/PE has demonstrated bidirectional effects with many significant social and occupational
$watermark-text
experiences (Neyer & Lehnart 2007, Roberts et al. 2003, Scollon & Diener 2006). In
addition, a decrease in E/PE over time was found to predict future internalizing problems
(Van den Akker et al. 2010). However, little attention has been given to mechanisms
underlying the association between this trait and depression. Support for an environmentally
mediated effect is limited and mixed (Middeldorp et al. 2008, Wetter & Hankin 2009), and
the environmental moderation model is largely untested, although there is some evidence
that positive affect moderates the effects of daily stressors on depressive symptoms
(Wichers et al. 2007).
negative life events (Anderson & McLean 1997, Compas et al. 2004), but mediation and
moderation effects have not been tested.
Kendler et al. 2006, Verstraeten et al. 2009). The interaction between N/NE and
conscientiousness is also of interest, as the latter construct includes aspects of self-regulation
and effortful control (Rothbart & Bates 2006) and may therefore reflect the ability to
modulate one’s affective reactivity. Indeed, there is cross-sectional evidence that effortful
control moderates the association between N/NE and depressive symptoms in adolescents
(Verstraeten et al. 2009).
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 14
The nature of relations between these personality traits and depression is complex, and our
understanding is still limited. N/NE, the most widely studied personality trait in depression,
raises challenging conceptual and methodological issues due to the overlap between some of
its features and depressive symptoms (Ormel et al. 2004b). Nonetheless, this cannot
completely explain the association between these constructs (Tang et al. 2009). N/NE is
moderately influenced by clinical state (the concomitants model), shares common etiological
influences with MDD (common cause, continuum/spectrum, and precursor models), predicts
the subsequent onset of MDD (precursor and predisposition models), and influences the
course of depression (pathoplasticity model). In addition, N/NE appears to contribute to
subsequent stress and adversity and increases the risk of depression in the face of negative
life events (predisposition model). Finally, it may also be changed by experience of MDD
$watermark-text
episodes (consequences model), but the evidence for this is weaker and less consistent.
The role played by E/PE in depression is less clear. Its cross-sectional association with
dysthymia is substantial, but its relation to MDD is more modest. E/PE is not influenced by
clinical state or changed by the experience of depressive episodes. It appears to be
abnormally low even during remission, which is consistent with the continuation of trait
deviance from the pre-morbid stage (precursor or predisposition accounts). Moreover, low
E/PE tends to predict a poorer course of depression. However, the degree of shared
etiological influences between E/PE and MDD is low, and the evidence that E/PE predicts
the onset of MDD in prospective longitudinal studies is weak. As noted above, three
possible reasons for the weaker and less consistent findings regarding E/PE are (a)it plays a
greater role in some forms of depression than others (e.g., chronic depressions); (b) only
some facets of the broader trait (e.g., low positive affective and approach motivation) are
$watermark-text
related to depression; and (c) E/PE may make a greater contribution to depression by
moderating N/NE than as a main effect (Olino et al. 2010).
Psychobiological Models
Gray’s model—Gray’s (e.g., Gray 1994) influential theory proposes that there are two
major neurobehavioral systems that underlie behavior: the behavioral activation system
(BAS), which responds to signals of reward, and the behavioral inhibition system (BIS),
which is sensitive to cues for punishment. Although BAS and BIS differ conceptually and
empirically from E/PE and N/NE, their relations with depression are thought to be similar.
Thus, it has been hypothesized that depression is associated with reduced BAS and/or
heightened BIS sensitivity (Depue & Iacono 1989, Gray 1994). Although much of this work
has focused on bipolar disorder (e.g., Alloy et al. 2008, Johnson et al. 2008), several recent
studies have examined self-report measures of BAS and BIS sensitivity in MDD. Consistent
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 15
with Gray’s model, compared with healthy controls, currently depressed patients report
lower levels of BAS and higher levels of BIS, and patients with a past history of MDD
report lower levels of BAS (Pinto-Meza et al. 2006). In addition, lower BAS sensitivity, but
not higher BIS sensitivity, is associated with a poorer course of MDD (e.g., Kasch et al.
2002, McFarland et al. 2006), suggesting that BAS may have a pathoplastic effect on
depression.
to signals of social approval and attachment), and persistence. The character dimensions are
self-directedness (responsible, goal-directed), cooperativeness (helpful, empathic versus
hostile and alienated), and self-transcendence (imaginative, unconventional). Harm
avoidance is conceptually and empirically associated with BIS, and novelty seeking and
persistence are associated with BAS. Similarly, harm avoidance is positively correlated with
N/NE and negatively associated with E/PE, self-directedness is negatively correlated with
N/NE, and novelty seeking and persistence are associated with E/PE (e.g., de Fruyt et al.
2000).
A number of studies have reported that patients with MDD report higher levels of harm
avoidance and lower levels of self-directedness than do healthy controls (e.g., Celikel et al.
2009). Most of the traits in Cloninger’s system are influenced by the respondent’s mood
state (e.g., Farmer et al. 2003); however, abnormal levels of harm avoidance and self-
$watermark-text
directedness are present even after remission (e.g., Smith et al. 2005). Increased harm
avoidance and lower self-directedness are also characteristic of most anxiety disorders,
indicating that these effects are not specific to MDD (Öngür et al. 2005).
Few studies have explicitly tested the common cause, precursor, predisposition, and
consequences hypotheses for Cloninger’s model. Farmer et al. (2003) found that the never-
depressed siblings of patients with MDD reported significantly greater harm avoidance and
less self-directedness than did the never-depressed siblings of healthy controls. In addition,
Cloninger et al. (2006) reported that in a large community sample, high harm avoidance and
persistence and low self-directedness predicted an increase in self-reported depressive
symptoms 12 months later. A larger number of studies have addressed the pathoplasticity
hypothesis, albeit with mixed results. Low harm avoidance, self-directedness, and reward
$watermark-text
dependency have predicted a poorer response to treatment in some, but not all, studies; the
other dimensions have generally not been associated with course and treatment outcome
(Joyce et al. 2007, Kennedy et al. 2005, Morris et al. 2009).
Clinical Traits
Independent of the traditional personality field, clinical researchers have developed a
number of trait-like constructs to describe dispositions to depression. These clinical traits are
similar in scope to personality facets, and their stability is comparable to that of a typical
personality dimension (e.g., Kasch et al. 2001, Zuroff et al. 2004). Also, factor analytic
studies have shown that most of these clinical traits can be successfully incorporated in the
personality taxonomy as components of neuroticism (Watson et al. 2006). Next, we briefly
discuss three of the most studied constructs: ruminative response style, self-criticism, and
dependency.
Ruminative response style, a tendency to dwell on sad mood and thoughts (Nolen-Hoeksema
1991), is correlated with concurrent depressive symptoms and predicts future symptoms as
well as increases in symptoms over time (Rood et al. 2009). Also, one study reported that
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 16
Blatt’s (1974, 1991) theory of depression focuses on two trait vulnerabilities: self-criticism
(an inclination to feelings of guilt and failure stemming from unrealistically high
expectations for oneself) and dependency (a disposition to feelings of helplessness and fears
of abandonment resulting from a preoccupation with relationships). These constructs are
similar, although not identical, to Beck’s (1983) constructs of autonomy and sociotropy.
Studies indicate that the link between dependency and depressive disorders is relatively
weak and nonspecific, whereas self-criticism has been established as an important and
$watermark-text
specific factor in these conditions (Zuroff et al. 2004). Both traits have been conceptualized
as dynamic predispositions to depressive disorders, and there is some support for this view,
including evidence of transactions with life stress as well as environmental mediation and
moderation of personality effects (Zuroff et al. 2004). Self-criticism, and to a lesser extent
dependency, have also been found to predict future increases in depressive symptoms. In
addition, there is evidence that dependency predicts the subsequent onset of major
depression in older, but not younger, individuals (Hirschfeld et al. 1989, Rohde et al. 1990).
The concomitants and pathoplasty models have also received empirical support (Zuroff et al.
2004). Finally, there is some research indicating that dependency may increase as a function
of depressive episodes (consequences model) in youth but not adults (Rohde et al. 1990,
1994; Shea et al. 1996).
As noted above, all of these constructs are strongly linked to N/NE (Cox et al. 2001, Kasch
$watermark-text
et al. 2001), and some (particularly rumination and self-criticism) can be considered facets
of this broader trait (Watson et al. 2006). Lower-order facets can account for variance over
and above that of higher-order traits (Paunonen & Ashton 2001), and several cross-sectional
studies have supported the incremental validity of ruminative response style and self-
criticism (Cox et al. 2004, Muris et al. 2009b) in associations with depressive symptoms.
However, this issue requires more research, particularly using longitudinal designs.
CHILD TEMPERAMENT
Most of the literature on personality and depression has focused on adolescents and adults.
Research that is grounded in the child temperament literature in developmental psychology
has the potential to extend existing work on personality in depression by (a) providing the
strongest test of the precursor and predisposition models; (b) more precisely delineating the
$watermark-text
The early childhood temperament dimensions that have received the greatest attention with
respect to depression are N/NE, E/PE, and behavioral inhibition (BI). BI refers to wariness,
fear, and low exploration in novel situations (Kagan et al. 1987). It combines aspects of N/
NE (fear and anxiety), E/PE (low approach), and conscientiousness (constraint/ effortful
control) that do not have a direct analog in most models of adult personality.
Cross-sectional and longitudinal studies of older children and adolescents using self-report
measures have generally reported associations of low E/PE and high N/NE with depression
similar to those in the adult literature (e.g., Lonigan et al. 2003).5 Observational studies of
younger children of depressed mothers also indicate that these traits may be associated with
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 17
risk for depression (Kovacs & Lopez-Duran 2010). For example, in a community sample of
100 three-year-olds, Durbin et al. (2005) reported that children of mothers with a history of
mood disorder exhibited low PE in emotion-eliciting laboratory tasks. Importantly, this
effect was limited to the affective (positive affect) and motivational (approach/engagement),
rather than the interpersonal (sociability), components of PE. Furthermore, low PE at age 3
predicted depressotypic cognition and memory biases at age 7 (Hayden et al. 2006) and
parent-reported depressive symptoms at age 10 (Dougherty et al. 2010).
Subsequently, using a larger community sample (N = 543), Olino et al. (2010) found that
preschool-aged children of parents with a history of depression had higher levels of NE and
BI. However, both main effects were qualified by interactions with child PE. At high and
$watermark-text
moderate (but not low) levels of child PE, higher levels of NE and BI were each associated
with higher rates of parental depression. Conversely, at low (but not high and moderate)
levels of child NE, low PE was associated with higher rates of parental depression. Taken
together, these results suggest that children of depressed parents may exhibit diminished PE
or elevated NE and BI. In this latter sample, low PE was also associated with elevated levels
of cortisol shortly after awakening, an index of hypothalamic-pituitary-adrenal axis
dysregulation that has been shown to predict MDD in adolescents and adults (Dougherty et
al. 2009).
In both the Durbin et al. (2005) and Olino et al. (2010) studies, the child temperament–
parental psychopathology associations were specific to depression. However, other work
suggests that children of parents with anxiety disorders may also exhibit elevated BI. For
example, Rosenbaum et al. (2000) assessed BI using laboratory measures in 2-to 6-year-old
$watermark-text
children of parents with a history of MDD and/or panic disorder and parents with no history
of mood or anxiety disorders. Children of patients with both MDD and panic disorder
exhibited significantly greater BI than did children of parents with no history of mood or
anxiety disorder. Children of parents with panic disorder alone and children of parents with
MDD alone had intermediate levels of BI that did not differ significantly from children of
parents in the comorbid and no-psychopathology groups.
Finally, there is some direct evidence that personality traits assessed in childhood predict the
development of depressive disorders in adults. Caspi et al. (1996) reported that children who
were rated as socially reticent, inhibited, and easily upset at age 3 had elevated rates of
depressive (but not anxiety or substance use) disorders at age 21. Moreover, van Os et al.
(1997) found that physicians’ ratings of behavioral apathy at ages 6, 7, and 11 were
predictive of both adolescent mood disorder and chronic depression in middle adulthood.
$watermark-text
CLINICAL IMPLICATIONS
Prevention
Personality research has important implications for the prevention of depression. Meta-
analytic evidence indicates that existing preventive interventions can reduce the incidence of
depressive disorders by 25% (Cuijpers et al. 2008). However, the available strategies are a
mix of universal (intervention is administered to the entire population), selective (to a well-
defined at-risk group), and indicated (to those with subthreshold disorder) approaches.
Universal interventions are costly, lack a personalized focus, and require very large samples
to yield detectable effects, whereas indicated interventions may be better described as
5Few studies have examined the association of conscientiousness/effortful control with depression in youth, but analogous to the adult
literature, there is cross-sectional evidence that effortful control is negatively correlated with depression (Verstraeten et al. 2009).
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 18
treatment than prevention (Muñoz et al. 2010). In contrast, selective interventions are true
preventive measures that are cost effective and can be tailored to a specific mechanism of
risk. However, implementation of selective strategies requires knowledge of risk factors and
causal processes that lead from the vulnerability to the disorder.
The majority of established risk factors for depressive disorders are either immutable (e.g.,
demographic characteristics, family history) or predict onset only in the short term (e.g.,
stressful life events). In contrast, personality is at least somewhat malleable, especially in
youth, but may forecast the onset of depression years in advance, which makes traits a
potentially attractive means of identifying individuals at risk and informing selection of
interventions. Different trait-disorder pathways would point to different preventive
$watermark-text
Treatment Response
Personality also can inform treatment of depressive disorders post onset. In particular, traits
can predict response to treatment. Substantial evidence has accumulated that individuals
with lower N/NE have better treatment outcomes across modalities (Kennedy et al. 2005,
Mulder 2002, Tang et al. 2009). Other Big Five traits have been studied less and their role is
not yet certain. However, a recent large investigation of a combination intervention
(medication plus psychotherapy) found that low N/NE and high conscientiousness predicted
who would respond to treatment, and although high E/PE did not contribute directly, it
$watermark-text
amplified the effect of high conscientiousness (Quilty et al. 2008a). As discussed above,
investigations of Cloninger’s traits have produced inconsistent results (Joyce et al. 2007,
Kennedy et al. 2005, Mulder 2002). Few studies have examined personality facets, but
preliminary evidence suggests that lower-order traits can add substantially to the prediction
of treatment response (Bagby et al. 2008). Among clinical traits, self-criticism, but not
dependency, was found to forecast poor treatment outcomes (Blatt et al. 1995). Furthermore,
personality may be useful in matching patients to interventions. For instance, Bagby et al.
(2008) reported that patients high on N/NE or low on some agreeableness facets respond
better to antidepressant medication than to psychotherapy.
The processes underlying these predictive associations are not entirely clear. One hypothesis
is that personality change mediates the effect of treatment on depression. Indeed, there is a
fair amount of evidence that depression treatment reduces N/NE and increases E/PE
$watermark-text
(Zinbarg et al. 2008) and that this effect is not due to confounding by the depressive state
(Tang et al. 2009). Quilty et al. (2008b) tested a mediation model and found direct support
for this hypothesis. Other possibilities need to be ruled out, however, particularly the
hypotheses that traits predict poorer response because they indicate a more severe form of
depression or that they interfere with treatment compliance and the therapeutic relationship,
thus reducing the efficacy of the intervention.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 19
multiple outcomes depending on subsequent events in the causal pathway. On the other
hand, many of the disorders that are currently classified as distinct conditions are closely
related; hence, research on personality-psychopathology associations can provide important
information for revising our nosological system. Third, reports of some traits (e.g., N/NE
and harm avoidance) are influenced by clinical state, whereas other traits (e.g., E/PE) appear
to be independent of mood state. However, state effects cannot fully account for the
associations between personality and depression. Fourth, shared etiological factors (e.g.,
genes) account for a portion of the association between N/NE and depression. Fifth,
depressive personality and some traits, particularly N/NE, predict the subsequent onset of
depressive disorders. However, it is unclear at this point whether they are best
conceptualized as precursors or predispositions, as it is difficult to tease these models apart,
$watermark-text
and there is evidence supporting both accounts. In either case, there is growing evidence that
temperamental risk factors are evident at an early age, suggesting a promising approach to
identifying young children at risk for depression. Sixth, there is evidence suggesting that
other traits, such as low E/PE and low conscientiousness/effortful control, may moderate the
relationship between N/NE and depression. Seventh, it appears unlikely that depressive
episodes produce enduring changes in most personality traits. Finally, personality traits
predict, and may in fact influence, the course and treatment response of depression.
To make further progress in elucidating the relation between personality and mood
disorders, future studies should be guided by six broad considerations. First, most of the
literature on personality and depression has focused on the broad traits of N/NE and E/PE.
There is a need for further work on conscientiousness and on lower levels in the trait
hierarchy (i.e., facets). It is important to determine whether a more specific level of analysis
$watermark-text
will yield more powerful effects and increase the specificity of associations between
personality constructs and particular forms of psychopathology. Clinical traits, such as
ruminative response style and self-criticism, need to be included in these studies and
evaluated jointly with traditional personality dimensions. Finally, it is important to continue
to explore interactions between traits.
Second, there is a critical need for prospective, longitudinal studies. Most existing
longitudinal studies have begun in late adolescence or adulthood. However, a substantial
proportion of mood disorders have already developed by mid-adolescence. Therefore, in
order to further test the precursor and predisposition models, and to trace the developmental
pathways between personality and depression, it is necessary to conduct longitudinal studies
that start as early as possible in order to obtain a sufficient number of first-onset cases and
$watermark-text
avoid selection biases caused by excluding participants who already have a history of mood
disorder at initial assessment.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 20
Fifth, self-reports have borne the brunt of most research in this area and have made
important contributions. However, like all methods, they have limitations and cannot be
applied in all contexts (e.g., young children). Thus, there is a need for further work using
complementary methods such as informant reports and observations in naturalistic and
laboratory settings.
Finally, the role of personality/temperament may differ for different forms of depressive
disorder. Personality appears to play an especially important role in early-onset, chronic, and
recurrent depressive conditions (e.g., Klein 2008, Kotov et al. 2010, van Os et al. 1997).
Focusing on broad diagnostic categories such as MDD may obscure important associations
with particular forms of depression; hence, future studies need to give greater consideration
$watermark-text
Glossary
LITERATURE CITED
Akiskal, HS. Validating affective personality types. In: Robins, L.; Barrett, J., editors. The Validity of
Psychiatric Diagnosis. New York: Raven; 1989. p. 217-227.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 21
Akiskal HS, Akiskal KK, Haykal RF, Manning JS, Connor PD. TEMPS-A: progress towards
validation of a self-rated clinical version of the Temperament Evaluation of the Memphis, Pisa,
Paris, and San Diego Autoquestionnaire. J. Affect. Disord. 2005; 85:3–16. [PubMed: 15780671]
Akiskal HS, Hirschfeld RMA, Yerevanian BI. The relationship of personality to affective disorders: a
critical review. Arch. Gen. Psychiatry. 1983; 40:801–810. [PubMed: 6344834]
Alloy LB, Abramson LY, Walshaw PD, Cogswell A, Grandin LD, et al. Behavioral approach system
and behavioral inhibition system sensitivities and bipolar spectrum disorders: prospective prediction
of bipolar mood episodes. Bipolar Disord. 2008; 10:310–322. [PubMed: 18271911]
Am. Psychiatr. Assoc.. Diagnostic and Statistical Manual of Mental Disorders. 4th ed.. Arlington, VA:
Am. Psychiatr. Assoc.; 1994.
Anderson KW, McLean PD. Conscientiousness in depression: tendencies, predictive utility, and
$watermark-text
Collaborative Research Program revisited. J. Consult. Clin. Psychol. 1995; 63:125–132. [PubMed:
7896977]
Brown TA, Barlow DH. A proposal for a dimensional classification system based on the shared
features of the DSM-IV anxiety and mood disorders: implications for assessment and treatment.
Psychol. Assess. 2009; 21:256–271. [PubMed: 19719339]
Canli T. Toward a neurogenetic theory of neuroticism. Ann. NY. Acad. Sci. 2008; 1129:153–174.
[PubMed: 18591477]
Caspi A, Moffitt TE, Newman DL, Silva PA. Behavioral observations at age 3 years predict adult
psychiatric disorders. Arch. Gen. Psychiatry. 1996; 53:1033–1039. [PubMed: 8911226]
Prospective birth-cohort study of relation of temperament in preschoolers and subsequent
psychopathology in early adulthood.
Caspi, A.; Shiner, RL. Personality development. In: Damon, W.; Lerner, R.; Eisenberg, N., editors.
Handbook of Child Psychology: Social, Emotional, and Personality Development. 6th ed.. New
$watermark-text
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 22
Compas BE, Connor-Smith J, Jaser SS. Temperament, stress reactivity, and coping: implications for
depression in childhood and adolescence. J. Clin. Child. Adolesc. Psychol. 2004; 33:21–31.
[PubMed: 15028538]
Cox BJ, Enns MW, Walker JR, Kjernisted K, Pidlubny SR. Psychological vulnerabilities in patients
with major depression versus panic disorder. Behav. Res. Ther. 2001; 39:567–573. [PubMed:
11341253]
Cox BJ, McWilliams LA, Enns MW, Clara IP. Broad and specific personality dimensions associated
with major depression in a nationally representative sample. Compr. Psychiatry. 2004; 45:246–
253. [PubMed: 15224266]
Cuijpers P, van Straten A, Smit F, Mihalopoulos C, Beekman A. Preventing the onset of depressive
disorders: a meta-analytic review of psychological interventions. Am. J. Psychiatry. 2008;
$watermark-text
Hillsdale, NJ: Erlbaum; 1994. Child personality and temperament: Does the five-factor model
embrace both domains?; p. 323-338.
Dougherty LR, Klein DN, Durbin CE, Hayden EP, Olino TM. Temperamental positive and negative
emotionality and children’s depressive symptoms: A longitudinal prospective study from age three
to age ten. J. Soc. Clin. Psychol. 2010; 29:462–488.
Dougherty LR, Klein DN, Olino TM, Dyson M, Rose S. Increased waking salivary cortisol and
depression risk in preschoolers: the role of maternal history of melancholic depression and early
child temperament. J. Child Psychol. Psychiatry. 2009; 50:1495–1503. [PubMed: 19686335]
Duggan CF, Lee AS, Murray RM. Does personality predict long-term outcome in depression? Br. J.
Psychiatry. 1990; 157:19–24. [PubMed: 2397360]
Duncan-Jones P, Fergusson DM, Ormel J, Horwood LJ. A model of stability and change in minor
psychiatric symptoms: results from three longitudinal studies. Psychol. Med. Monogr. Suppl.
1990; 18:1–28. [PubMed: 2236383]
$watermark-text
Durbin CE, Klein DN, Hayden EP, Buckley ME, Moerk KC. Temperamental emotionality in
preschoolers and parental mood disorders. J. Abnorm. Psychol. 2005; 114:28–37. [PubMed:
15709809]
Fanous AH, Neale MC, Aggen SH, Kendler KS. Alongitudinal study of personality and major
depression in a population-based sample of male twins. Psychol. Med. 2007; 37:1163–1172.
[PubMed: 17407614]
Farmer A, Redman K, Harris T, Mahmood A, Sadler S, et al. Neuroticism, extraversion, life events
and depression: the Cardiff Depression Study. Br. J. Psychiatry. 2002; 181:118–122. [PubMed:
12151281]
Farmer A, Mahmood A, Redman K, Harris T, Sadler S, McGuffin P. A sib-pair study of the
temperament and character scales in major depression. Arch. Gen. Psychiatry. 2003; 60:490–496.
[PubMed: 12742870]
Fraley RC, Roberts BW. Patterns of continuity: a dynamic model for conceptualizing the stability of
individual differences in psychological constructs across the life course. Psychol. Rev. 2005;
112:60–74. [PubMed: 15631588]
Gershuny BS, Sher KJ. The relation between personality and anxiety: findings from a 3-year
prospective study. J. Abnorm. Psychol. 1998; 107:252–262. [PubMed: 9604554]
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 23
Goldberg LR. The structure of phenotypic personality traits. Am. Psychol. 1993; 48:26–34. [PubMed:
8427480]
Gray, JA. Framework for a taxonomy of psychiatric disorder. In: van Goozen, SHM.; Van de Poll,
NE., editors. Emotions: Essays on Emotion Theory. Hillsdale, NJ: Erlbaum; 1994. p. 29-59.
Griffith JW, Zinbarg RE, Craske MG, Mineka S, Rose RD, et al. Neuroticism as a common dimension
in the internalizing disorders. Psychol. Med. 2010; 40:1125–1136. [PubMed: 19903363]
Hayden EP, Klein DN, Durbin CE, Olino TM. Low positive emotionality at age three predicts
depressotypic cognitions in seven-year-old children. Dev. Psychopathol. 2006; 18:409–423.
[PubMed: 16600061]
Hecht H, van Calker D, Berger M, von Zerssen D. Personality in patients with affective disorders and
their relatives. J. Affect. Disord. 1998; 51:33–43. [PubMed: 9879801]
$watermark-text
Hirschfeld RMA, Klerman GL, Clayton PJ, Keller MB. Personality and depression: empirical findings.
Arch. Gen. Psychiatry. 1983a; 40:993–998. [PubMed: 6615162]
Hirschfeld RMA, Klerman GL, Clayton PJ, Keller MB, Mc Donald-Scott P, Larkin BH. Assessing
personality: effects of the depressive state on trait measurement. Am. J. Psychiatry. 1983b;
140:695–699. [PubMed: 6846626]
Hirschfeld RMA, Klerman GL, Lavori P, Keller MB, Griffith P, Coryell W. Premorbid personality
assessments of first onset of major depression. Arch. Gen. Psychiatry. 1989; 46:345–350.
[PubMed: 2649038]
Hirshfeld-Becker DR, Micco J, Henin A, Bloomfield A, Biederman J, Rosenbaum J. Behavioral
inhibition. Depress. Anxiety. 2008; 25:357–367. [PubMed: 18412062]
Huprich SK. Evaluating facet-level predictions and construct validity of depressive personality
disorder. J. Personal. Disord. 2003; 17:219–232.
Johnson SL, Cueller AK, Ruggero C, Winett-Perlman C, Goonick P, et al. Life events as predictors of
$watermark-text
mania and depression in bipolar I disorder. J. Abnorm. Psychol. 2008; 117:268–277. [PubMed:
18489203]
Johnson W, Hicks BM, McGue M, Iacono WG. Most of the girls are alright, but some aren’t:
personality trajectory groups from ages 14 to 24 and some associations with outcomes. J. Personal.
Soc. Psychol. 2007; 93:266–284.
Joiner TE, Lonigan CJ. Tripartite model of depression and anxiety in youth psychiatric inpatients:
relations with diagnostic status and future symptoms. J. Clin. Child Psychol. 2000; 29:372–382.
[PubMed: 10969421]
Jorm AF, Christensen H, Henderson AS, Jacomb PA, Korten AE, Rodgers B. Predicting anxiety and
depression from personality: Is there a synergistic effect of neuroticism and extraversion? J.
Abnorm. Psychol. 2000; 109:145–149. [PubMed: 10740946]
Joyce PR, McKenzie JM, Carter JD, Rae AM, Luty SE, et al. Temperament, character, and personality
disorders as predictors of response to interpersonal psychotherapy and cognitive-behavioural
$watermark-text
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 24
Kendler KS, Neale MC, Kessler RC, Heath AC, Eaves LJ. A longitudinal twin study of personality
and major depression in women. Arch. Gen. Psychiatry. 1993; 50:853–862. [PubMed: 8215811]
Kendler KS, Gatz M, Gardner CO, Pederson NL. Personality and major depression. Arch. Gen.
Psychiatry. 2006; 63:1113–1120. [PubMed: 17015813] Tests models of personality and MDD in a
prospective study of a large twin sample.
Kennedy, SH.; Farvolden, P.; Cohen, NL.; Bagby, RM.; Costa, PT. The impact of personality on the
pharmacological treatment of depression. In: Rosenbluth, M.; Kennedy, SH.; Bagby, RM., editors.
Depression and Personality. Arlington, VA: Am. Psychiatr. Publ.; 2005. p. 97-119.
Kercher AJ, Rapee RM, Schniering CA. Neuroticism, life events and negative thoughts in the
development of depression in adolescent girls. J. Abnorm. Child Psychol. 2009; 37:903–915.
[PubMed: 19437113]
$watermark-text
Klein DN. Depressive personality: reliability, validity, and relation to dysthymia. J. bnorm. Psychol.
1990; 99:412–421.
Klein DN. Depressive personality in the relatives of outpatients with dysthymic disorder and episodic
major depressive disorder and normal controls. J. Affect. Disord. 1999; 55:19–27. [PubMed:
10512602]
Klein, DN. Dysthymia and chronic depression. In: Craighead, WE.; Miklowitz, DJ.; Craighead, LW.,
editors. Psychopathology: History, Theory, and Diagnosis. Hoboken, NJ: Wiley; 2008. p. 329-365.
Klein, DN.; Dougherty, LR.; Laptook, RS.; Olino, TM. Temperament and risk for mood disorders in
adolescents. In: Allen, N.; Sheeber, L., editors. Adolescent Emotional Development and the
Emergence of Depressive Disorders. Cambridge, UK: Cambridge Univ. Press; 2008a. p. 238-261.
Klein DN, Miller GA. Depressive personality in nonclinical subjects. Am. J. Psychiatry. 1993;
150:1718–1724. [PubMed: 8214182]
Klein, DN.; Torpey, DC.; Bufferd, SJ.; Dyson, MW. Depressive disorders. In: Beauchaine, TP.;
$watermark-text
Hinshaw, SP., editors. Child and Adolescent Psychopathology. Hoboken, NJ: Wiley; 2008b. p.
477-509.
Klein, MH.; Wonderlich, S.; Shea, MT. Models of relationships between personality and depression:
toward a framework for theory and research. In: Klein, MH.; Kupfer, DJ.; Shea, TM., editors.
Personality and Depression: A Current View. New York: Guilford; 1993. p. 1-54.
Kotov R, Gamez W, Schmidt FL, Watson D. Linking “Big” personality traits to anxiety, depressive,
and substance use disorders: a meta-analysis. Psychol. Bull. 2010; 136:768–821. [PubMed:
20804236] Documents personality profiles of common mental disorders through a systematic
meta-analysis.
Kotov R, Watson D, Robles J, Schmidt NB. Personality traits and anxiety symptoms: the Multilevel
Trait Predictor Model. Behav. Res. Ther. 2007; 45:1485–1503. [PubMed: 17240351]
Kovacs M, Lopez-Duran N. Prodromal symptoms and atypical affectivity as predictors of major
depression in juveniles: implications for prevention. J. Child. Psychol. Psychiatry. 2010; 51:472–
$watermark-text
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 25
Laptook RS, Klein DN, Dougherty LR. Ten-year stability of depressive personality disorder in
depressed outpatients. Am. J. Psychiatry. 2006; 163:865–871. [PubMed: 16648328]
Lonigan CJ, Phillips BM, Hooe ES. Relations of positive and negative affectivity to anxiety and
depression in children: evidence from a latent-variable longitudinal study. J. Consult. Clin.
Psychol. 2003; 71:465–481. [PubMed: 12795571]
Markon KE, Krueger RF, Watson D. Delineating the structure of normal and abnormal personality: an
integrative hierarchical approach. J. Personal. Soc. Psychol. 2005; 88:139–157.
McFarland BR, Shankman SA, Tenke CE, Bruder GE, Klein DN. Behavioral activation system deficits
predict the six-month course of depression. J. Affect. Disord. 2006; 91:229–234. [PubMed:
16487598]
Middledorp CM, Cath DC, Beem AL, Willemsen G, Boomsma DI. Life events, anxious depression
$watermark-text
and personality: a prospective and genetic study. Psychol. Med. 2008; 38:1557–1565. [PubMed:
18294422]
Morey LC, Shea MT, Markowitz JC, Stout RL, Hopwood CJ, et al. State effects of major depression
on the assessment of personality and personality disorder. Am. J. Psychiatry. 2010; 167:528–535.
[PubMed: 20160004]
Morris BH, Bylsma LM, Rottenberg J. Does emotion predict the course of major depressive disorder?
A review of prospective studies. Br. J. Clin. Psychol. 2009; 48:255–273. [PubMed: 19187578]
Mulder RT. Personality pathology and treatment outcome in major depression: a review. Am. J.
Psychiatry. 2002; 159:359–371. [PubMed: 11869996]
Muñoz RF, Cuijpers P, Smit F, Barrera AZ, Leykin Y. Prevention of major depression. Annu. Rev.
Clin. Psychol. 2010; 6:181–212. [PubMed: 20192789]
Muris P, Bos AER, Mayer B, Verkade R, Thewissen V, Dell’Avvento V. Relations among behavioral
inhibition, Big Five personality factors, and anxiety disorder symptoms in nonclinical children.
$watermark-text
3:400–424.
Olino TM, Klein DN, Dyson MW, Rose SA, Durbin CE. Temperamental emotionality in preschool-
aged children and depressive disorders in parents: associations in a large community sample. J.
Abnorm. Psychol. 2010; 119:468–478. [PubMed: 20677836] Examines relation of parental
depression with observations of temperament in large sample of preschool-aged offspring.
Öngür D, Farabaugh A, Iosifescu DV, Perils R, Fava M. Tridimensional Personality Questionnaire
factors in major depressive disorder: relationship to anxiety disorder comorbidity and age of onset.
Psychother. Psychosom. 2005; 74:173–178. [PubMed: 15832068]
Ormel, J.; de Jong, A. On vulnerability to common mental disorders. An evidence-based plea for a
developmental perspective. In: Tansella, M.; Thornicroft, G., editors. Common Mental Disorders
in Primary Care: Essays in Honour of Professor Sir David Goldberg. London: Routledge; 1999.
p. 34-51.
Ormel J, Oldehinkel AJ, Brilman EI. The interplay and ecological continuity of neuroticism,
difficulties, and life events in the etiology of major and subsyndromal, first and recurrent
depressive episodes in later life. Am. J. Psychiatry. 2001; 158:885–891. [PubMed: 11384895]
Prospectively tested a dynamic predisposition model of neuroticism-depression relations in a
sample of older adults.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 26
Ormel J, Oldehinkel AJ, Vollebergh W. Vulnerability before, during, and after a major depressive
episode. Arch. Gen. Psychiatry. 2004a; 61:990–996. [PubMed: 15466672] Prospectively tested
competing models of personality-depression relations in a large community sample.
Ormel J, Rosmalen J, Farmer A. Neuroticism: a noninformative marker of vulnerability to
psychopathology. Soc. Psychiatry Psychiatr. Epidemiol. 2004b; 39:906–912. [PubMed:
15549243]
Ørstavik RE, Kendler KS, Czajkowski N, Tambs K, Reichborn-Kjennerud T. The relationship between
depressive personality disorder and major depressive disorder: a population-based twin study.
Am. J. Psychiatry. 2007; 164:1866–1872. [PubMed: 18056242]
Ouimette PC, Klein DN, Pepper CM. Personality traits in the first degree relatives of outpatients with
depressive disorders. J. Affect. Disord. 1996; 39:43–53. [PubMed: 8835653]
$watermark-text
Paunonen SV, Ashton MD. Big Five factors and facets and the prediction of behavior. J. Personal. Soc.
Psychol. 2001; 81:524–539.
Pinto-Meza A, Caseras X, Soler J, Puigdemont D, Pèrez V, Torrubia R. Behavioral inhibition and
behavioral activation systems in current and recovered major depression participants. Personal.
Individ. Differ. 2006; 40:215–226.
Quilty LC, De Fruyt F, Rolland JP, Kennedy SH, Rouillon PF, Bagby RM. Dimensional personality
traits and treatment outcome in patients with major depressive disorder. J. Affect. Disord. 2008a;
108:241–250. [PubMed: 18067975]
Quilty LC, Meusel LAC, Bagby RM. Neuroticism as a mediator of treatment response to SSRIs in
major depressive disorder. J. Affect. Disord. 2008b; 111:67–73. [PubMed: 18384882]
Reich J, Noyes R, Hirschfeld R, Coryell W, O’Gorman T. State and personality in depressed and panic
patients. Am. J. Psychiatry. 1987; 144:181–187. [PubMed: 3812784]
Roberts BW, Bogg T. A 30-year longitudinal study of the relationships between conscientiousness-
$watermark-text
related traits, and the family structure and health-behavior factors that affect health. J. Personal.
2004; 72:325–354.
Roberts BW, Caspi A, Moffitt T. Work experiences and personality development in young adulthood.
J. Personal. Soc. Psychol. 2003; 84:582–593.
Roberts BW, DelVecchio WF. The rank-order consistency of personality from childhood to old age: a
quantitative review of longitudinal studies. Psychol. Bull. 2000; 126:3–25. [PubMed: 10668348]
Roberts BW, Kuncel NR, Shiner R, Caspi A, Goldberg LR. The power of personality: the comparative
validity of personality traits, socioeconomic status, and cognitive ability for predicting important
life outcomes. Perspect. Psychol. Sci. 2007; 2:313–345.
Roberts BW, Walton KE, Viechtbauer W. Patterns of mean-level change in personality traits across
the life course: a meta-analysis of longitudinal studies. Psychol. Bull. 2006; 132:1–25. [PubMed:
16435954]
Rohde P, Lewinsohn PM, Seeley JR. Are people changed by the experience of having an episode of
$watermark-text
depression? A further test of the scar hypothesis. J. Abnorm. Psychol. 1990; 99:264–271.
Rohde P, Lewinsohn PM, Seeley JR. Are adolescents changed by an episode of major depression? J.
Am. Acad. Child Adolesc. Psychiatry. 1994; 33:1289–1298. [PubMed: 7995795]
Rood L, Roelofs J, Bögels SM, Nolen-Hoeksema S, Schouten E. The influence of emotion-focused
rumination and distraction on depressive symptoms in non-clinical youth: a meta-analytic review.
Clin. Psychol. Rev. 2009; 29:607–616. [PubMed: 19682781]
Rorsman B, Grasbeck A, Hagnell O, Isberg PE, Otterbeck L. Premorbid personality traits and
psychometric background factors in depression: the Lundby Study 1957–1972.
Neuropsychobiology. 1993; 27:72–79. [PubMed: 8515830]
Rosenbaum JF, Biederman J, Hirshfeld-Becker DR, Kagan J, Snidman N, et al. A controlled study of
behavioral inhibition in children of parents with panic disorder and depression. Am. J.
Psychiatry. 2000; 157:2002–2010. [PubMed: 11097967]
Rothbart, MK.; Bates, JE. Temperament. In: Eisenberg, N.; Damon, W.; Lerner, RM., editors.
Handbook of Child Psychology: Vol. 3. Social, Emotional, and Personality Development. 6th
ed.. Hoboken, NJ: Wiley; 2006. p. 99-166.
Rudolph KD, Klein DN. Exploring depressive personality traits in youth: origins, correlates, and
developmental consequences. Dev. Psychopathol. 2009; 21:1155–1180. [PubMed: 19825262]
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 27
Ryder AG, Quilty LC, Vachon DD, Bagby RM. Depressive personality and treatment outcome in
major depressive disorder. J. Personal. Disord. 2010; 24:392–404.
Schneider, K. Psychopathic Personalities. London: Cassell; 1958.
Scollon CN, Diener E. Love, work, and changes in extraversion and neuroticism over time. J. Personal.
Soc. Psychol. 2006; 91:1152–1165.
Shea MT, Leon AC, Mueller TI, Solomon DA, Warshaw MG, Keller MB. Does major depression
result in lasting personality change? Am. J. Psychiatry. 1996; 153:1404–1410. [PubMed:
8890672]
Smith DJ, Duffy L, Stewart ME, Muir WJ, Blackwood DHR. High harm avoidance and low self-
directedness in euthymic young adults with recurrent, early-onset depression. J. Affect. Disord.
2005; 87:83–89. [PubMed: 15967233]
$watermark-text
Smits DJM, Boeck PD. From BIS/BAS to the Big Five. Eur. J. Personal. 2006; 20:255–270.
Tang TZ, DeRubeis RJ, Hollon SD, Amsterdam J, Shelton R, Schalet B. Personality change during
depression treatment: a placebo-controlled trial. Arch. Gen. Psychiatry. 2009; 66:1322–1330.
[PubMed: 19996037] Finds that antidepressants influenced personality independent of their
effects on depression in a placebo-controlled trial.
Vachon DD, Sellbom M, Ryder AG, Miller JD, Bagby RM. A five-factor model description of
depressive personality disorder. J. Personal. Disord. 2009; 23:447–465.
Van denAkker A, Deković M, Prinzie P. Transitioning to adolescence: how changes in child
personality and overreactive parenting predict adolescent adjustment problems. Dev.
Psychopathol. 2010; 22:151–163. [PubMed: 20102653]
van Os J, Jones PB. Early risk factors and adult person–environment relationships in affective
disorder. Psychol. Med. 1999; 29:1055–1067. [PubMed: 10576298]
van Os J, Jones P, Lewis G, Wadsworth M, Murray R. Developmental precursors of affective illness in
$watermark-text
a general population birth cohort. Arch. Gen. Psychiatry. 1997; 54:625–631. [PubMed: 9236546]
Verstraeten K, Vasey MW, Raes F, Bijttebier P. Temperament and risk for depressive symptoms in
adolescence: mediation by rumination and moderation by effortful control. J. Abnorm. Child
Psychol. 2009; 37:349–361. [PubMed: 19107592]
Watson, D. Mood and Temperament. New York: Guilford; 2000.
Watson D. Differentiating the mood and anxiety disorders: a quadripartite model. Annu. Rev. Clin.
Psychol. 2009; 5:221–247. [PubMed: 19327030]
Watson, D.; Kotov, R.; Gamez, W. Basic dimensions of temperament in relation to personality and
psychopathology. In: Krueger, R.; Tackett, J., editors. Personality and Psychopathology. New
York: Guilford; 2006. p. 7-38.
Wetter EK, Hankin BL. Mediational pathways through which positive and negative emotionality
contribute to anhedonic symptoms of depression: a prospective study of adolescents. J. Abnorm.
Child Psychol. 2009; 37:507–520. [PubMed: 19184402] Examines mediators and moderators of
$watermark-text
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 28
SUMMARY POINTS
1. There are moderate-to-large cross-sectional associations between depression and
three general personality traits—N/NE, E/PE, and conscientiousness—as well as
with a variety of related traits (e.g., harm avoidance, rumination, and self-
criticism) and personality types (e.g., depressive personality).
2. Most of the personality traits associated with depression also are related to other
forms of psychopathology, particularly anxiety disorders. This may reflect the
phenomenon of multifinality, in which variables early in the causal chain lead to
multiple outcomes depending on subsequent events in the causal pathway. On
$watermark-text
the other hand, many of the disorders that are currently classified as distinct
conditions are closely related; hence, research on personality-psychopathology
associations can also provide important information for revising our nosological
system.
3. Reports of some traits (e.g., N/NE and harm avoidance) are influenced by
clinical state, whereas other traits (e.g., E/PE) appear to be independent of mood
state. However, state effects cannot fully account for the associations between
personality and depression.
4. Shared etiological factors (e.g., genes) account for a portion of the association
between N/NE and depression.
5. Depressive personality and some traits, particularly N/NE, predict the
$watermark-text
response of depression.
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.
Klein et al. Page 29
Table 1
Summary of key predictions of the classic models
$watermark-text
Predisposition Predicts depression onset; other variables mediate or moderate this link
Pathoplasticity Predicts variation in presentation or outcome of depression above and beyond other baseline characteristics
Concomitants Is altered during a depressive episode but returns to premorbid level after
Annu Rev Clin Psychol. Author manuscript; available in PMC 2012 December 10.