Predictors of The First Onset of A Major Depressive Episode and Changes in Depressive Symptoms Adolescence: Stress and Negative Cognitions

Journal of Abnormal Psychology
2011, Vol. 120, No. 4, 779 796
2011 American Psychological Association

0021-843X/11/$12.00 DOI: 10.1037/a0025441
Predictors of the First Onset of a Major Depressive Episode

and Changes in Depressive Symptoms Across Adolescence:
Stress and Negative Cognitions
Jocelyn Smith Carter
Judy Garber
DePaul University
Vanderbilt University
This 6-year longitudinal study examined stressors (e.g., interpersonal, achievement), negative cognitions
(self-worth, attributions), and their interactions in the prediction of (a) the first onset of a major
depressive episode (MDE), and (b) changes in depressive symptoms in adolescents who varied in risk for
depression. The sample included 240 adolescents who were first evaluated in Grade 6 (M 11.86 years
old; SD 0.57; 54.2% female) and then again annually through Grade 12. Stressful life events and
depressive diagnoses were assessed with interviews; negative cognitions and depressive symptoms were
assessed with self-report questionnaires. Discrete time hazard modeling revealed a significant interaction
between interpersonal stressors and negative cognitions, indicating that first onset of an MDE was
predicted by high negative cognitions in the context of low interpersonal stress, and by high levels of
interpersonal stressors at both high and low levels of negative cognitions. Analyses of achievement
stressors indicated significant main effects of stress, negative cognitions, and risk in the prediction of an
MDE, but no interactions. With regard to the prediction of depressive symptoms, multilevel modeling
revealed a significant interaction between interpersonal stressors and negative cognitions such that
among adolescents with more negative cognitions, higher levels of interpersonal stress predicted higher
levels of depressive symptoms, whereas at low levels of negative cognitions, the relation between
interpersonal stressors and depression was not significant. Risk (i.e., maternal depression history) and sex
did not further moderate these interactions. Implications for intervention are discussed.
Keywords: stress, depression, negative cognitions, adolescents
self-schema (e.g., I am worthless, I cant do anything right, I

am unlovable) that is activated in the face of stress will interpret
their subsequent experiences based on these negative selfperceptions, and hence are vulnerable to depression (Beck, 1991).
Similarly, the helplessness/hopelessness model (Abramson et al.,
1978, 1989) posits that when exposed to stress, individuals inferences about the implications of the events with regard to characteristics of the self, the future consequences of the events, and
attributions about the causes of the events contribute to hopelessness and depression.
Although the various cognitive models of depression emphasize
somewhat different cognitions, the basic cognitive vulnerability
hypothesis derived from these theories is similar: some individuals are more likely to develop depression in the face of stressful
life events because of their interpretations and expectations of the
meaning of these events for their lives. This cognitive-stress model
typically is tested by examining the statistical interaction between
cognitions and stress; that is, at high levels of stress, more negative
cognitions, particularly about the self and the causes of events, will
predict higher levels of depression (Monroe & Simons, 1991).
Evidence consistent with the cognitive-stress model in relation
to depressive symptoms has been found with children and adolescents across a variety of negative cognitions (see Abela & Hankin,
2008; Lakdawalla, Hankin, & Mermelstein, 2007, for reviews).
However, many of the earlier studies testing this model in children
used cross-sectional designs and, therefore, could not address
whether the diathesis-stress interaction temporally preceded de-
Cognitive theories of depression propose that individuals with

certain cognitive vulnerabilities are likely to develop depression
when confronted with stressful life events (Abramson, Metalsky,
& Alloy, 1989; Abramson, Seligman, & Teasdale, 1978; Beck,
1967; Brown & Harris, 1987). Several theories have implicated
low self-worth, which is the degree to which an individual devalues him/herself as a person, as a particularly important vulnerability to depression (e.g., Beck, 1967, 1976; Brown & Harris,
1978; Roberts & Monroe, 1994, 1999). For example, Beck (Beck,
1976, 1991) suggested that individuals with a latent negative
This article was published Online First September 19, 2011.

Jocelyn Smith Carter, Department of Psychology, DePaul University;
Judy Garber, Psychology and Human Development, Vanderbilt University.
This work was supported in part by National Institute of Mental Health
(NIMH) Grants R29 MH454580, K02 MH66249; National Institute of
Child Health and Human Development Grant P30HD15052; and a Faculty
Scholar Award (1214 88) and Grant (173096) from the William T. Grant
Foundation. Jocelyn Smith Carter was supported in part by a Ford Foundation Fellowship. We appreciate the cooperation of the Nashville Metropolitan School District, and we especially thank the parents and children
who participated in the project. Thanks also to Dr. Grayson Holmbeck for
his statistical advice, and the Action Editor, Dr. Sherryl Goodman, and the
anonymous reviewers for their helpful suggestions.
Correspondence concerning this article should be addressed to Jocelyn
Smith Carter, PhD, DePaul University, Department of Psychology, Byrne
Hall 531, 2219 North Kenmore Avenue, Chicago, IL 60614. E-mail:
jcarter9@depaul.edu
779
780
CARTER AND GARBER
pression over time (e.g., Hankin & Abramson, 2001; Kraaij et al.,
2003). More recent studies using short-term longitudinal designs
have found prospective evidence of the moderating effect of negative cognitions on stress in the prediction of change in depressive
symptoms (Abela & Hankin, 2008). In a review of 18 such studies,
Lakdawalla et al. (2007) reported that the follow-up lengths ranged
from 1 to 60 months (M 8.5 months; median 3.0 months;
mode 12 months), and sample sizes ranged from 34 to 1,507
(M 310.56; median 176; mode 79). The current longitudinal study builds on this literature in the following important
ways by: (a) following a moderate size sample (N 240) of youth
over a longer time intervalsix years, (b) testing the cognitive
vulnerability hypothesis to predict both the first onset of a major
depressive episode (MDE) and changes in depressive symptoms
during adolescence, (c) using interview-based measures of stress
and depressive disorders, (d) examining two types of stressors
(interpersonal, achievement), (e) testing a composite index of
negative cognitions, and (f) exploring the direct and moderating
role of sex and two important features of maternal depression
history and current level of depressive symptoms.
Predicting Onset of a Major Depressive Episode and

Changes in Depressive Symptoms
An early critique of cognitive models of depression was that
they were relevant to depressive symptoms but not depressive
diagnoses (e.g., Coyne & Gotlib, 1983). Indeed, some studies
predicting self-reported depressive symptoms have yielded different results from those predicting a depressive diagnosis (Avenevoli
& Steinberg, 2001; Compas, Ey, & Grant, 1993; Hankin, Fraley,
Lahey, & Waldman, 2005). Some evidence now exists that is
consistent with the cognitive diathesis-stress model in relation to
diagnosed depressive disorders in college student samples (e.g.,
Alloy et al., 2006; Hankin, Abramson, Miller, & Haeffel, 2004).
Few studies (Hammen, 1988; Lewinsohn, Joiner, & Rohde, 2001),
however, have tested whether the cognitive vulnerability model
predicts depressive diagnoses in children and adolescents, and
results of these studies have been inconsistent. Using a normative
sample of adolescents, Lewinsohn and colleagues (2001) found
that the cognitive-stress model predicted the onset of depressive
disorders, although not changes in depressive symptoms during the
same time period. In contrast, Hammen and colleagues (Hammen,
1988; Hammen, Adrian, & Hiroto, 1988) did not find that the
cognitive diathesis-stress model predicted the onset of depressive
disorders in a high-risk child sample.
Possible explanations for these different results are that: (a) the
relatively small sample (n 88) in the study by Hammen and
colleagues (Hammen, 1988; Hammen et al., 1988) may have
lacked power to detect significant interactions; (b) Hammen and
colleagues included children across a wide age range (i.e., 8 16
years old); the younger childrens cognitive styles might not have
been developed yet (Abela, 2001; Gibb & Alloy, 2006; NolenHoeksema, Girgus, & Seligman, 1992; Turner & Cole, 1994); (c)
the nature and extent of the relations among negative cognitions,
stress, and depression may differ in high-risk versus normative
samples; and (d) these two studies assessed stress differently and
used different measures of cognitive vulnerability. Whereas Hammen (1988; Hammen et al., 1988) employed a semistructured life
events interview, Lewinsohn et al. (2001) used an 11-item check-
list. Regarding cognitions, Hammen and colleagues measured selfconcept, self-schema, and attributional style, whereas Lewinsohn
et al., measured attributional style and dysfunctional attitudes.
Given the different findings in these earlier investigations, and the
paucity of studies testing the cognitive vulnerability model with
respect to clinical diagnoses in youth despite the frequent call for
such research (e.g., Abela, 2001; Abela & Veronneau-McArdle,
2002), a primary aim of the current prospective study was to test
the cognitive diathesis-stress model using a larger, high-risk sample of same-age youth to predict both the first onset of a major
depressive episode as well as changes in depressive symptoms.
Stressors
Not all types of stressors are related to depression in the same
way and not all stressors are similarly moderated by negative
cognitions to predict depression. Given the heterogeneity in how
stress has been defined and measured in the literature, inconsistent
findings across studies are not surprising (Monroe & Simons,
1991). Investigations of the cognitive diathesis-stress model typically have used a broad index of stress comprised of events in
several domains. A few studies have tested the model in youth with
regard to specific stressors such as peer rejection and victimization
(e.g., Panak & Garber, 1992; Prinstein & Aikins, 2004; Prinstein,
Cheah, & Guyer, 2005), trauma (e.g., Gibb & Alloy, 2006), and
academic failure (e.g., Hilsman & Garber, 1995). These studies
found evidence consistent with the cognitive-stress model with
regard to the particular stressors examined, although they did not
contrast different types of stressors within the same study (McMahon, Grant, Compas, Thurm, & Ey, 2003).
Stressors can be categorized as interpersonal and achievement
events. Interpersonal stressors involve interactions with another
person(s) such as conflict, rejection, and break-ups. Achievement
stressors typically involve failure or disappointment in relation to
a goal. The examination of interpersonal stressors is particularly
relevant to cognitive-interpersonal models of depression (e.g.,
Gotlib & Hammen, 1992; Joiner & Coyne, 1999). The integration
of cognitive and interpersonal perspectives takes into account the
social context in which cognitions develop and the special salience
of interpersonal relationships (Rudolph, Hammen, & Burge, 1997;
Rudolph et al., 2000). Indeed, interpersonal stressors have been
found to have a stronger relation to depression than noninterpersonal stressors (e.g., Rudolph & Hammen, 1999a; Rudolph et al.,
2000).
The most commonly used method to assess stressful life events
is with self-report checklists, which usually yield a total count of
events rather than separate scores for interpersonal or noninterpersonal events. Although some self-report checklists and contextual
threat interviews have been found to have overlapping items and to
be similarly associated with depression (Lewinsohn, Rohde, &
Gau, 2003), life events checklists typically do not adequately
measure severe life events or gather information about the timing
of the events (Duggal et al., 2000; Grant & McMahon, 2005;
McQuaid, Monroe, Roberts, Kupfer, & Frank, 2000). In contrast,
contextual threat interviews allow investigators to make objective
ratings of the impact of a stressful event and to date the onset and
offset of an event (Brown & Harris, 1978). These interviews
facilitate the acquisition of detailed information about contextual
factors surrounding the event, the objective impact of the event on
PREDICTING FIRST ONSET OF MAJOR DEPRESSION
the participant, the duration of the event, and the possible role of
the individual in generating the event.
Several stress interviews have been developed for use with
children and adolescents (e.g., Hammen, 1991; Hankin, Mermelstein, & Roesch, 2007; Williamson et al., 1998). Few studies,
however, have tested the cognitive-stress interaction in adolescents
using such life stress interviews (e.g., Hammen, 1988; Hammen et
al., 1988). Therefore, another aim of the present study was to
examine the cognitive vulnerability model using objective threat
ratings of interpersonal and achievement events in adolescents.
Cognitive Vulnerability
The present study focused on cognitions that cut across the main
cognitive theories of depression and are considered to be central to
the development of depression (Garber, 2007). In particular, we
operationalized the cognitive vulnerability in terms of perceived
self-worth and attributional style because these constructs are
fundamental to several of the leading psychological theories of
depression including Becks (1967) cognitive theory, Browns
self-esteem vulnerability model (Brown, Bifulco, & Harris, 1987;
Southall & Roberts, 2002), and the Helplessness (Abramson et al.,
1978) and Hopelessness theories (Abramson, Metalsky, & Alloy,
1989; Abramson et al., 1978). The cognitive vulnerability hypothesis common to all of these theories is that individuals with
certain maladaptive thinking patterns (e.g., low self-esteem; negative attributional style) are at increased risk for depression when
they experience negative life events because of how they interpret
and respond to those events.
Previous research has shown that various measures of negative
cognitions in children and adolescents are highly correlated
(Abela, Aydin, & Auerbach, 2006; Adams, Abela, & Hankin,
2004) suggesting that they may represent a common underlying
construct that forms a latent factor (Ginsburg et al., 2009; Gotlib,
Lewinsohn, Seeley, Rohde, & Redner, 1993; although see also
Abela, 2001; Conley, Haines, Hilt, & Metalsky, 2001). For example, in a study of several different cognitive inventories, measures
of self-esteem and attributional style, in particular, were found to
load onto the same factor (Adams et al., 2007). In the current
study, we examined the relations between the measures of selfworth and attributional style across multiple assessments, and
created a composite cognitive vulnerability index similar to that
used in other studies testing the cognitive-stress model in both
adults (Alloy et al., 2006) and children (Turner & Cole, 1994).
Whereas several studies (e.g., Abela, 2001; Hankin et al., 2004)
have explicitly compared different measures of the cognitive vulnerability as prescribed by the various cognitive theories (e.g.,
Becks cognitive model vs. the Hopelessness theory), the current
study used a composite index of the cognitive vulnerability to
address a different set of questions regarding specific stressors,
moderators, and measures of depression.
Risk: Maternal Depression

High-risk research designs involve the study of individuals who
do not currently have the disorder of interest (e.g., depression) but
who are hypothesized to be at high risk for developing the disorder
due to some risk factor (Alloy, Lipman, & Abramson, 1992;
Zuroff, Mongrain, & Santor, 2004). One of most potent risks for
781
depression in youth is having a parent with a mood disorder

(Beardslee, Versage, & Gladstone, 1998; Cummings, Keller, &
Davies, 2005; Goodman & Gotlib, 1999). Children of depressed
parents are exposed to higher levels of stress (Cummings et al.,
2005; Hammen, Shih, Altman, & Brennan, 2003; Hammen, Shih,
& Brennan, 2004) and have more negative cognitions (Garber &
Robinson, 1997; Goodman, Adamson, Riniti, & Cole, 1994;
Jaenicke et al., 1987) than children of nondepressed parents. Moreover, the relation between stress and depressive symptoms has
been found to be significant in offspring of depressed mothers
(e.g., Jaser, 2005; Langrock, Compas, Keller, Merchant, & Copeland, 2002), and some evidence consonant with the cognitivestress model has been reported for at-risk children with high levels
of dysfunctional attitudes or low self-esteem in relation to depressive symptoms (Abela & McGirr, 2007; Abela & Skitch, 2007).
The sample in the present study consisted of adolescents who
varied in risk for depression as a function of their mothers history
of mood disorders. Such a high-risk design provides greater variability in the constructs of interest (i.e., stressful life events,
negative cognitions, and depression), thereby reducing problems
associated with restriction of range, which is particularly important
when testing statistical interactions (McClelland & Judd, 1993).
That is, sampling across a wide range of observations that include
more extreme scores can facilitate the detection of statistically
reliable interactions by reducing standard errors without biasing
parameter estimates (McClelland & Judd, 1993). Thus, using a
sample of high- and low-risk participants might increase power to
detect moderation.
The risk factor of maternal depression can vary with regard to
such characteristics as episode chronicity and recurrence, as well
as the severity of current symptoms (Foster et al., 2008; Tompson
et al., 2009). The present study explored the independent contributions of maternal depression history and current level of maternal depressive symptoms to the prediction of depression in their
children. Additionally, we examined whether the relations among
stress, negative cognitions, and depression in adolescents varied as
a function of maternal depression history (i.e., risk).
Sex Differences
Studies examining sex differences in cognitive vulnerability and
in the diathesis-stress interaction have yielded inconsistent results.
Whereas some investigations have reported no sex differences in
attributional style (Gladstone, Kaslow, Seeley, & Lewinsohn,
1997; Hankin, Abramson, & Siler, 2001; Thompson, Kaslow,
Weiss, & Nolen-Hoeksema, 1998), other studies have shown that
girls have more negative attributional and inferential styles (Hankin & Abramson, 2002), more negative automatic thoughts (Calvete & Cardenoso, 2005), and lower self-esteem (Allgood-Merten,
Lewinsohn, & Hops, 1990) than boys. One study reported a more
negative attributional style for boys (Gladstone et al., 1997).
Several researchers (e.g., Hankin & Abramson, 2001; Hyde,
Mezulis, & Abramson, 2008; Nolen-Hoeksema & Girgus, 1994;
Rudolph et al., 2000) have suggested that the rise in depression
rates among adolescent girls might be partially due to their experiencing higher levels of stress (Compas, Slavin, Wagner, & Vannatta, 1986; Rudolph, 2002; Rudolph & Hammen, 1999a) or their
being more reactive to stress than boys (Little & Garber, 2004;
Rudolph, 2002; Rudolph & Hammen, 1999a; Seiffge-Krenke &
CARTER AND GARBER
782
Stemmler, 2002). Given the increasing rate of depression during

adolescence, particularly among females, and the mixed findings
with respect to sex differences in negative cognitions and stress,
some studies have tested whether sex moderates the cognitive
diathesis-stress interaction (e.g., Abela, 2001; Abela & Payne,
2003; Prinstein & Aikins, 2004; Prinstein et al., 2005). Results of
these investigations, however, also have been inconsistent. Significant cognitive vulnerability by stress interactions have been found
for girls, but not for boys (Abela, 2001; Abela & McGirr, 2007;
Chambers et al., 1985; Prinstein & Aikins, 2004), or for boys only
(Abela & Payne, 2003; Prinstein et al., 2005). The present prospective study further tested whether the cognitive-stress model of
depression differed by sex in a sample of adolescents who had a
range of stressors, negative cognitions, and depressive outcomes.
The Current Study

The current study tested the cognitive diathesis-stress model
with regard to two outcomes (i.e., depressive disorder and symptoms) using two different data analytic approaches. To examine
predictors of the first onset of a major depressive episode (MDE)
from early to middle adolescence, we used discrete-time hazard
modeling (DTHM). DTHM takes into account the number of time
periods during which an individual could experience a target event
(e.g., a depressive episode) and calculates the probability of an
individual experiencing the event within the particular time period
given that s/he has not yet had the event. DTHM determines both
whether and when particular events (e.g., onset of a first depressive
episode) occur and can include predictors of the occurrence and
timing of these events (Singer & Willett, 2003). Second, we used
multilevel modeling (MLM) to examine changes in depressive
symptoms across six years. We utilized an idiographic approach
(e.g., Abela & McGirr, 2007; Abela & Skitch, 2007), which takes
into account individual differences in stress, negative cognitions,
and depressive symptoms over time. For example, the determination of an individuals level of stress was made based on whether
the person scored above his or her own mean level of stress over
time and not whether the person scored above the group mean for
stress levels (Abela et al., 2006; Abela, Skitch, Adams, & Hankin,
2006).
To date, few studies have used MLM to examine the cognitive
diathesis-stress model of depression in children. In an investigation
of the interaction between stress defined as mothers level of
depressive symptoms, and childrens initial cognitive vulnerability
to predict fluctuations in childrens depressed mood, Abela,
Skitch, and colleagues (2006) showed that this cognitive-stress
interaction could be modeled using MLM methods. The present
study expanded on this work of Abela and colleagues (e.g., Abela
& McGirr, 2007; Abela & Skitch, 2007) by using MLM methods
to test a broader range of stressors across a longer time period,
using interview-based measures of two different stressor types, and
exploring sex and risk as possible moderators of the cognitivestress relation in youth. The following specific questions were
addressed: (1) Does the interaction between each specific type of
stressor (i.e., interpersonal or achievement) and negative cognitions predict (a) the first onset of a major depressive episode, and
(b) change in depressive symptoms. (2) Do sex, risk (i.e., maternal
depression history), and current maternal depressive symptoms
further moderate the relation between stress and cognitions to

predict a MDE and change in depressive symptoms in adolescents?
Method
Participants
Participants were 240 mothers and children first assessed when
they were in 6th grade (mean age 11.86, SD .57). The sample
was 54.2% female, 82% Caucasian, 14.7% African American, and
3.3% other (Hispanic, Asian, Native American, or mixed ethnic
background). Families were predominantly working (e.g., nurses
aid, sales clerk) to middle class (e.g., store manager, teacher) with
a mean socioeconomic status (Hollingshead, 1975) of 41.84 (SD
13.25).
Procedures
Parents of 5th grade children from metropolitan public schools
were invited to participate in a study about parents and children. A
brief health history questionnaire comprised of 24 medical conditions (e.g., diabetes, heart disease, depression) and 34 medications
(e.g., Prozac, Elavil, Valium) was sent with a letter describing the
study to over 3,500 families. Of the 1,495 mothers who indicated
an interest in participating, the 587 who endorsed either a history
of depressive symptoms, use of antidepressants, or no history of
psychopathology were interviewed further by telephone. The remaining families were excluded because the mother did not indicate depression or indicated other kinds of psychiatric problems
without depression, or had a serious medical illness (e.g., cancer,
multiple sclerosis). Of the 587 families screened, 238 were excluded because they did not indicate sufficient symptoms to meet
criteria for a depressive disorder (38%), had other psychiatric
disorders that did not also include a depressive disorder (19%), the
mother or the target child had a serious medical condition (14%),
the family no longer was interested (21%), the target child was in
the wrong grade (6%), or the family had moved out of the area
(2%). The remaining 349 mothers who indicated that they had a
history of depression or had no psychiatric problems were interviewed in person using the Structured Clinical Interview for DSM
diagnoses (SCID; Spitzer, Williams, Gibbon, & First, 1990).
Based on the SCID, 149 families then were excluded because the
mother indicated a history of a psychiatric diagnosis that did not
also include a mood disorder or reported a serious medical condition, or the child had a serious and/or chronic medical illness or a
pervasive developmental disorder.
The final sample of 240 families consisted of 185 mothers who
had had a depressive disorder (147 mothers had had diagnoses of
Major Depressive Disorder (MDD); the remaining 38 mothers had
had diagnoses of Dysthymia, Depression NOS, or Adjustment
Disorder with Depressed Mood); 55 mothers were life-time free of
psychopathology (low risk). Among the depressed mothers, the
average number of depressive episodes they had had was 2.05
(SD 1.19); 26.7% (n 65) also had had alcohol and/or drug
abuse or dependence, and 27.5% (n 66) had had an anxiety
disorder.
Mothers depressive symptoms were assessed annually with the
Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock,
& Erlbaugh, 1961), which measures affective, cognitive, behav-
ioral, and somatic symptoms. For each of the 21 items, respondents

chose one of four statements that best described how they had been
feeling during the past two weeks. Items were scored on a 0 to 3
point scale and were summed to create a total score (range 0 to 63).
Coefficient alphas for mothers BDI scores in this sample ranged
from .89 to .92 across all assessments. At Waves 1 through 5,
mothers BDI scores [mean (standard deviation)] for the whole
sample were, respectively, 6.94 (7.79), 7.20 (8.05), 5.85 (6.57),
6.41 (7.10), 7.19 (7.92).
Mothers and children completed a battery of questionnaires and
were interviewed separately about the childs psychiatric history
by study personnel who were unaware of the mothers psychiatric
history. The present study reports the results of the baseline (Grade
6) and annual follow-up assessments of the adolescents during
Grades 7, 8, 9, 11, and 12.1 Only those measures relevant to the
current study are described here. Cognitive measures assessed in
Grades 6, 7, 8, 9, and 11, and depressive symptoms assessed in
Grades 7, 8, 9, 11, and 12 were used in the present analyses. The
life events interviews were in reference to the time since the last
evaluation. Thus, Wave 1 refers to cognitions measured in Grade
6, stressors that had occurred between Grades 6 and 7, and depression measured in Grade 7. Wave 2 Grade 7 cognitions,
stressors between Grades 7 and 8, and depressive symptoms in
Grade 8; Wave 3 Grade 8 cognitions, stressors between Grades
8 and 9, and depressive symptoms in Grade 9; Wave 4 Grade 9
cognitions, stressors between Grades 9 and 11, and depressive
symptoms in Grade 11; Wave 5 cognitions in Grade 11, stressors in Grades 11 to 12 and depressive symptoms in Grade 12.
Measures
Adolescent depression diagnosis and symptoms.
The
Schedule for Affective Disorders and Schizophrenia for SchoolAge ChildrenEpidemiological versionPresent and Lifetime (KSADS-PL; Kaufman et al., 1997) is a semistructured clinical
interview from which diagnoses of depressive disorders can be
made. At the first assessment (Grade 6), the K-SADS-PL was
administered to mothers and children. Follow-up interviews were
conducted annually using the Longitudinal Interval Follow-up
Evaluation (LIFE; Keller et al., 1987), which parallels the
K-SADS and assesses disorders since the previous interview.
Mothers and children reported separately about the youths depressive symptoms in the K-SADS-PL and LIFE interviews, and were
combined according to the standard procedures described in the
instructions of these interviews (Ambrosini, 2000). The LIFE
yields a depression score on a six-point scale from 1 (no or one
depressive symptom and no impairment) to 6 (meets criteria for a
major depressive episode with marked impairment). A score of 5
indicates a definite MDE. All interviews were audio-taped. A
second rater who was unaware of the scoring of the primary
interviewer reviewed a random 25% of the interview audiotapes.
Kappa (Cohen, 1960) was 0.81 for depressive disorders.
The Childrens Depression Inventory (CDI; Kovacs, 1981,
1985) contains 27 items measuring cognitive, affective, and behavioral symptoms of depression. Each item lists three statements,
scored 0 through 2, in order of increasing symptom severity. The
CDI has adequate internal consistency, testretest reliability, and
convergent validity with other self-report measures (e.g., Abela,
2001; Cole, Hoffman, Tram, & Maxwell, 2000; Smucker, Craig-
783
head, Craighead, & Green, 1986). Internal consistency of the CDI

in this sample was equal to or greater than .83 at all time points.
CDI scores from Grades 6, 7, 8, 9, 11, and 12 were used in the
present analyses.
Stressors. Life events were assessed annually with regard to
events that had occurred in the adolescents life since the previous
evaluation. Mothers and adolescents were interviewed separately
with the Life Events Interview for Adolescents (LEIA; Garber,
Keiley, & Martin, 2002), which was based on the Life Events and
Difficulties Schedule (Brown & Harris, 1978, 1989), and the Life
Stress Interview developed by Hammen et al. (1987). The LEIA is
a semistructured interview that allows for precise dating of events
and the assessment of objective consequences of events, given the
particular context in which they occurred. In separate interviews,
mothers and adolescents were asked to describe the context of the
event, what happened, who was involved, when it occurred, and
what changed for the adolescent as a result of the event. Events in
the adolescents life that were reported by either the youth or the
mother were included (see Morris, Ciesla, & Garber, 2010 for
further details).
Interviewers presented to a group of trained raters all information about each adolescents life events as reported by the adolescent and/or the mother; the group rated the events using a 7-point
severity scale with regard to the degree of objective threat the
event had for the adolescent given the context, ranging from 1
none to 7 severe. Raters were unaware of any information about
the mothers or adolescents psychopathology. Interrater reliability
of the objective stress ratings was obtained by having interviewers
present the information about the events at the same time to two
different groups who made independent ratings of the events.
Based on 202 events, agreement among raters was 89.6%, with a
Kappa of .79. A total event count and a total level of stress rating
were derived from the interview for each subject. Because the two
stress variables were highly correlated (r .92), analyses were
conducted using only one indicator of stress, the mean level of
stress that occurred each year. As in other studies of the relation
between stress and depression (e.g., Barnett & Gotlib, 1988;
Lewinsohn et al., 2001), life events were assessed at Wave N 1
regarding the previous year; that is, events that occurred between
Wave N and Wave N 1. In addition, monthly stress scores were
calculated and indicated the mean level of stress that occurred each
month. Annual stress scores were available for Grades 7 through
12.
Based on their content, events also were categorized as either
interpersonal or achievement focused. Interpersonal stressors were
defined as those that involved adolescents relationships with
another person (e.g., family or peer conflict), whereas achievement
stressors were those having to do with performance in academic or
job domains (e.g., failing a final exam). Two raters independently
read the written narratives recorded by the interviewers and coded
each event type as primarily interpersonal or achievement. Interrater reliability was good ( .86, p .001).
Cognitions. The global self-worth scale of the Harter SelfPerception Profile (GSW; Harter, 1985) contains six items mea1
An abbreviated data collection was conducted when participants were

in Grade 10 due to a temporary lapse in funding.
784
CARTER AND GARBER
suring the degree to which children are satisfied with themselves

and their lives and think the way things are is fine. Responses are
scored on a 4-point scale with lower scores representing lower
perceived self-worth. Coefficient alpha ranged from .81 to .84
across the time points.
The Childrens Attributional Style Questionnaire (CASQ;
Seligman et al., 1984) measures attribution dimensions derived
from the reformulated learned helplessness model (Abramson et
al., 1978). The revised CASQ (Thompson et al., 1998) containing 12 positive and 12 negative items, was used. Each item
varies one causal dimension (locus, stability, globality) while
holding the other two dimensions constant. A mean positive
score was created by dividing the number of external, unstable,
and specific responses to good events by the total number of
positive events; a mean negative score was created by dividing the number of internal, stable, and global responses to all
bad events by the total number of negative events. The total
score was derived by subtracting the negative from the positive
scores, as is typical when using the CASQ (Gladstone &
Kaslow, 1995). Lower total scores reflect a more negative
attributional style. Coefficient alpha for the total score ranged
from .63 to .76 across waves, which is consistent with what has
been found elsewhere (Gladstone & Kaslow, 1995; Robins &
Hinkley, 1989).
At each wave, a composite index of cognitive vulnerability was
created by standardizing and then averaging scores on the global
self-worth and attributional style measures; items were recoded so
that higher scores represented higher levels of negative cognitions
(i.e., poorer self-worth and a more negative attributional style). In
the current study, the autocorrelations for global self-worth
(.50.61) and attributional style (.50.65) were significant and
indicated moderate stability over time. The correlations between
global self-worth and attributional style were significant at each
wave (Wave 1 .51, Wave 2 .46, Wave 3 .55, Wave 4 .39,
Wave 5 .53; all ps .05); the composite measure demonstrated
a high degree of reliability (ryys .81; ryys .81; Nunnally &
Bernstein, 1994).
Missing Data
Data were available across Waves as follows: 207 participants
had complete data on all study measures at the assessment conducted at Wave 1, 181 at Wave 2, 158 at Wave 3, 156 at Wave 4,
and 146 at Wave 5. In the DTHM analyses, participants were
censored when they dropped out. MLM analyses allow for variability in how many times participants are measured. In the current
study, all available data from each participant were used to calculate the growth trajectories. Participants with and without missing
data were compared on study variables using ANOVAs on continuous variables and chi-square analyses for categorical variables.
Those who did not complete the study had significantly higher
levels of achievement stress at Wave 1, F 9.96, p .01 and
were more likely to be male than those who completed the study
[2(1) 11.75, p .01]. No other significant differences were
found at any other time point or on any other measure or demographic characteristic. Participants available data were used in
each of the analyses.
Results
Descriptive Statistics
Descriptive statistics and bivariate correlations for all study
variables were computed. Means and standard deviations for the
entire sample for each study variable are presented in Table 1. Data
are also presented separately for males and females and for lowand high-risk participants. One-way ANOVAs were used to test
for sex and risk differences in study variables.
Compared to low-risk youth, high-risk adolescents had significantly more interpersonal stressors, more negative cognitions,
higher depression scores at each wave, and more achievement
stressors at Wave 6. One-way ANOVAs testing for risk group
differences on monthly stress scores showed that high-risk youth
had higher interpersonal stress levels during 43 out of 70 (61%)
months.2 For achievement stressors, significant differences between high- and low-risk adolescents were found for 2 out of 70
months (2%).
Correlations among the study variables are presented in Table 2.
Interpersonal stressors were moderately correlated over time;
achievement stressors showed low stability across time. The negative cognitions composite variable was moderately stable over
time with correlations ranging from .32 to .66; depressive symptoms also were moderately stable across waves (rs .48 to .60).
At most time points, a more negative cognitive style was significantly correlated with higher levels of depressive symptoms.
Predicting Onset of a Major Depressive Episode:

Discrete-Time Hazard Models
Data analytic plan. Discrete-time hazard model (DTHM)
analyses, conducted with PASW 17 LOGIC, were used to test
predictors of the first onset of an MDE. First, models for the
specification of time were compared to each other. The completely
general model that included each month as a predictor of the
hazard function was compared to more parsimonious models for
time. The model with the fewest parameters for time that best
explained the hazard function was used as the baseline model
(Singer & Willett, 2003). Predictors then were added to the baseline model with the most parsimonious time specification. Sex and
risk were included as control variables and also were tested as
possible moderators.
Time-varying measures of stress and depression assessed
monthly were used; mean monthly stress ratings for each stressor
type (interpersonal or achievement) were tested as predictors in
separate analyses. Stressors occurring during the prior month were
used to predict depression in the following month. That is, stressors were lagged so that they predicted the onset of depressive
disorders during the month following the stressor onset. This time
frame is consistent with previous studies of both adolescents and
adults showing that major life events precede the onset of a
depressive disorder by four to nine weeks (e.g., Kendler,
Karkowski, & Prescott, 1999; Sandberg, Rutter, Pickles, McGuinness, & Angold, 2001). The time metric for DTHM analyses was
month; thus 70 variables were dummy-coded to model months.
2
Results available from authors upon request.
785
Table 1
Means, Standard Deviations for the Total Sample, and as a Function of Sex and Risk
Sex
Total Sample
Variable
Wave 1
Interpersonal stress
Achievement stress
Negative Cognitionsa
Depressive symptomsb
Wave 2
Achievement stress
Negative cognitionsa
Wave 3
Interperpersonal stress
Achievement stress
Negative cognitionsa Cognitionsa
Wave 4
Achievement stress
Wave 5
Achievement stress
a
Females
Risk
Males
Low
ANOVA
High
Sex
Risk
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
2.18
.18
.00
4.76
2.69
.60
.87
4.45
2.10
.11
.08
4.62
2.76
.35
.82
4.60
2.27
.25
.10
4.93
2.61
.80
.92
4.27
2.24
.07
.32
3.08
3.27
.23
.78
3.10
2.19
.21
.09
5.28
2.50
.67
.88
4.67
.22
3.39
2.49
.25
.01
2.04
9.96
9.72
1.75
.16
.00
4.68
2.21
.54
.86
5.06
1.58
.12
.06
4.43
2.09
.53
.88
4.62
1.95
.21
.10
4.98
2.34
.55
.92
5.56
1.53
.24
.26
2.88
1.88
.86
.82
3.14
1.81
.14
.08
5.27
2.30
.40
.85
5.43
1.60
1.75
1.08
.58
.69
1.64
6.08
8.55
1.71
.16
.00
4.88
2.20
.39
.88
5.22
1.75
.15
.10
5.17
2.27
.41
.89
5.81
1.67
.16
.12
4.51
2.12
.37
.87
4.35
1.84
.18
.30
3.57
2.40
.44
.81
3.64
1.67
.15
.09
5.31
2.14
.38
.89
5.58
.08
.02
2.98
.73
.24
.34
6.91
3.94
1.47
.10
.00
5.17
2.08
.26
.83
5.40
1.56
.13
.04
5.58
2.34
.31
.87
5.64
1.35
.07
.05
4.66
1.73
.17
.79
5.08
1.38
.11
.34
3.10
2.01
.33
.79
3.95
1.49
.10
.11
5.88
2.11
.23
.82
5.65
.57
3.38
.46
1.32
.12
.10
9.80
9.92
1.21
.09
.01
5.72
1.72
.33
.88
6.39
1.21
.09
.06
6.32
1.85
.25
.88
6.81
1.20
.08
.07
4.83
1.55
.41
.88
5.65
1.40
.08
.28
4.15
1.67
.23
.69
5.58
1.15
.09
.08
6.31
1.73
.36
.92
6.60
.00
.06
.98
2.44
.84
.03
5.69
4.23
Negative cognitions Cognitive composite.

p .05. p .01.
Depressive symptoms measured with the Childrens Depression Inventory.
Adolescents negative cognitions composite scores at study entry were tested as a time-invariant predictor in the DTHM.3 Untransformed variables were used in the DTHM analyses because a
person-centered score could not be calculated based on the single
measurement of the cognitive variable at Wave 1. Moreover,
person-centered scores are not appropriate for use with the depression diagnostic outcome variable as individuals are removed from
the analysis once they experience an onset of an event. Finally,
separate models were tested for each stressor type (interpersonal,
achievement). Model-trimming was performed as follows: threeway interactions with sex or risk were examined first, then twoway interactions, and then main effects.
Post hoc probing of interaction effects followed the guidelines
recommended by Aiken, West, and Reno (1991) and clarified by
Holmbeck (2002). For each significant interaction effect, new
variables were created at one standard deviation (SD) above the
mean and one SD below the mean. These new variables were used
to create new conditional moderator variables, and analyses were
rerun using the conditional moderator variable to obtain simple
slopes of the moderation effects. These simple slopes then were
used in the graphical representations of the interactions.
Results.
The hazard function modeled the probability of
participants experiencing a first onset of a depressive episode
[Depression Rating Scale (DSR) score of 5] during the six years
of the study. Seventeen participants were excluded from the
DTHM analyses because they had a previous or current depressive
episode at study entry. Given the large number of waves in the

current study, several alternative specifications for time were compared with the goal of finding a parsimonious time specification
that fit the data (Singer & Willett, 2003). The completely general
model that included all 70 month variables was compared to a
constant model that constrained hazard to be constant over time,
the linear model that allowed hazard to increase over time, and
then the quadratic, cubic, fourth order, and fifth order polynomials.
Parameters, deviance statistics, and Akaike Information Criterion
(AIC; Akaike, 1973) statistics for each time specification are
presented in Table 3. The quadratic specification was chosen based
on a combination of the chi-square difference test and examination
of AIC, following the guidelines that a more parsimonious time
specification is preferred as long as it works about as well as the
general specification and not any worse than a more complex
model (Singer & Willett, 2003). Bolded items indicate when the
chi-square difference test showed a significant drop in chi-square
fit between consecutive models (see Table 3). Whereas the quadratic model fit less well than the linear model, the cubic model
was not significantly better than the quadratic model and the AIC
3
Monthly measures of cognitions were not available, and adolescents
annual cognitions scores could not be divided into discrete monthly periods.
W1 Interpersonal Stress
W1 Achievement Stress
.03
W1 Negative Cognitions
.03
W1 Depressive Symptoms
DeepDepression
.15
.36
.08
.15
.10
.18
.05
.15
.10
.07
.08
.11
.03
.08
.02
.07
.16
.01
.08
.14
.02
.02
.11
.02
.08
.04
.06
.05
.14
.20
.07
.05
.12
.17
.05
.52
.18
.20
.07
.06
.58
.48
.33
.60
.14
.16
.05
.07
.55
.46
.26
.44
.00
.01
.07
.06
.44
.34
.22
.42
.01
.03
.12
.09
.32
.28
.23
.49
.01
.12
.00
.30
.12
.03
.02
.06
.07
.08
.08
.10
.01
.02
.03
.07
.00
.05
.16
.05
.08
.13
.01
.00
.03
.15
.07
.07
.01
.55
.10
.13
.05
.07
.66
.52
.48
.48
.02
.10
.07
.02
.54
.43
.40
.46
.09
.08
.02
.01
.43
.44
.41
.49
Note. W1 Wave 1; W2 Wave 2; W3 Wave 3; W4 Wave 4; W5 Wave 5.
p .05. p .01. p .001.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
1.
2.
3.
4.
Variable
Table 2
Correlations Among Study Variables
.01
.13
.05
.21
.17
.10
.03
.21
.08
.08
.12
.01
.04
.03
.01
.08
.01
.02
.07
.08
.00
10
12
13
14
.46
.01
.01
.00
.05
.07
.11
.62 .03
.18
.61 .58
.06
.08
.28
.57
.51
.07
.04
.09
.52
.04
.03
.36
.04
.10
.02
.03
.04
.11
11
.49
.12
.04
.61
.52
15
17
18
19
20
.06
.00
.14
.46 .04
.04
.56 .02 .02 .42
16
786
CARTER AND GARBER
787
Table 3
Discrete Time Hazard Model Time Specification Results
Deviance difference in comparison to
TIME
n parameters
Deviance
Previous model
General model
AIC
General
Constant
Linear
Quadratic
Cubic
Fourth order
Fifth order
70
1
2
3
4
5
6
890.65
859.53
858.30
850.46
850.39
843.66
842.33
31.12
1.23
7.84
0.07
6.73
1.33
31.12
32.35
40.19
40.26
46.99
48.32
1044.66
861.53
862.30
856.46
858.39
853.66
854.33
Note.
AIC Akaike Information Criterion. Deviance statistics in bold show significant changes in chi-square.
for the quadratic model was lower than the cubic model. Thus, the
most parsimonious specification for time was the quadratic model,
which included the month and the squared month variables.
The first series of models tested the three-way interactions to
address whether sex or risk moderated the effects of stress and
negative cognitions on the onset of a depressive episode. The
hazard function then was modeled as a function of a set of control
and substantive predictors. Similar analyses were run separately
for each stressor type. The main effects of sex, Wave 1 cognitions,
and mean monthly stress ratings were included as were the twoway interactions between stress and cognitions, stress and sex, and
cognitions and sex, and the three-way interaction among stress, cognitions, and sex. Parallel analyses were conducted to test risk as a
moderator of the cognitive-stress interaction, and included the threeway and two-way interactions among stressors, cognitions, and risk.
The model with both the three- and two-way interactions was
not significant for interpersonal stressors and did not converge for
achievement stressors. The next series of models included the
two-way interactions between stressor type and the negative cognitions composite. The interaction between interpersonal stressors
and negative cognitions was significant (see top of Table 4), over
and above the significant effect of risk. Simple slope analyses of
the interaction (see Figure 1) indicated that the first onset of a
major depressive episode (MDE) was predicted by high negative
cognitions in the context of low interpersonal stress (Odds 1.14,
p .01, CI 1.04 1.25, Hazard .53), as well as by high levels
of interpersonal stressors at both high (Odds .51, p .05, CI
.27.97, Hazard .35) and low levels of negative cognitions
(Odds .53, p .05, CI .30.97, Hazard .34).
The interaction of negative cognitions with achievement stressors was not significant. Therefore, we next tested the most parsimonious model for achievement stressors and found significant
main effects of time, risk, stressors, and negative cognitions in
predicting the hazard function (see bottom of Table 4). High-risk
youth had significantly increased odds of developing depression
over the course of the study. Additionally, the negative cognitive
composite and achievement stressors each significantly predicted
depression onset.4 For every one unit increase in the mean monthly
achievement stress score, the odds of developing a depressive
disorder increased by 1.24. Sex was not a significant predictor or
moderator of the onset of an MDE in either the interpersonal or
achievement stressor models.
Predicting Change in Depressive Symptoms Over

Time: Multi-Level Modeling
Data analytic plan. To test whether the interaction between
negative cognitions and stress predicted the dimensional measure
of depressive symptoms over time within individuals and whether
the cognitive-stress interaction varied by sex or risk, multilevel
modeling (MLM) in the PASW MIXED program was used. MLM
tests for fixed and random effects at multiple levels. In the current
study, data were analyzed at two levels: individual (level-2) and
repeated measures within individuals (level-1). Time, stress, negative cognitions, and maternal depressive symptoms were modeled
as time-varying (level-1) predictors; individuals mean level of
stress, individuals mean level of negative cognitions, sex, and risk
were modeled as time-invariant (level-2) predictors. In addition,
the prior years depression scores (CDI) were used as control
variables to test for increases in depression, which yielded a more
conservative test of study hypotheses (Shih, 2006).
The two stressor types (i.e., interpersonal, achievement), negative cognitions, and depressive symptoms were centered within
each person at each time-point to provide an idiographic approach
to the analyses (e.g., Abela & Skitch, 2007). This centering approach allowed us to examine whether changes in an individuals
level of stress and/or negative cognitions were associated with
changes in that persons depressive symptoms over and above their
previous levels of depressive symptoms. For each variable, the
mean of each persons scores on that variable across the five waves
was subtracted from the persons scores on the variable at each
wave. These person-centered scores thus represented fluctuations
4
Analyses of the three- and two-way interactions conducted separately
for the two cognitive measures (i.e., global self-worth, attributional style)
yielded similar results as those found with the cognitions composite index.
The model testing the two-way interaction between interpersonal stressors
and global self-worth was significant (OR .69, CI .51.93, hazard
.67, p .05) such that the odds of having a depressive episode were
increased by 2.37 for every one unit increase in interpersonal stress and by
.89 for every one unit increase in negative global self-worth. The interaction between achievement stressors and global self-worth was not significant (OR 3.55, CI .6419.72, hazard .78, p .15). The interactions between attributional style and interpersonal stressors (OR 1.00,
CI .891.11, hazard .73, p .95) and achievement stressors (OR
1.23, CI .722.10, hazard .77, p .72) were not significant.
CARTER AND GARBER
788
Table 4
Discrete Time Hazard Model Results: Two-Way Interactions of Stress and Negative Cognitions, and Main Effects of Stress and
Negative Cognitions, Controlling for Risk
Two-way interactions
Interpersonal
Achievement
Stressor type
Odds
CI (95%)
Wald
Hazard
Odds
CI (95%)
Wald
Hazard
Intercept
Risk
Month
Month Month
Stress
Negative Cognitionsa
Stress Neg Cognitions
4.66
3.66
.75
1.00
.65
.04
1.03
1.1012.22
.70.82
1.001.00
.28.97
.971.12
1.001.06
36.82
4.44
47.04
42.41
4.29
1.30
3.88
.79
.43
.50
.34
.51
.51
5.17
2.82
.74
1.00
2.51
1.08
.94
.849.43
.68.81
1.001.00
.867.30
1.021.15
.861.04
48.53
2.83
46.54
41.27
2.83
6.28
1.37
.74
.43
.50
.71
.52
.49
1.0011.82
.68.81
1.001.00
1.071.45
1.011.15
33.87
3.88
46.10
42.18
7.95
5.19
.77
.43
.50
.55
.52
Main Effects: Achievement Stress
Intercept
Risk
Month
Month Month
Stress
Negative Cognitions
.01
3.45
.74
1.00
1.24
1.08
Note. Haz Hazard Function; Neg Cog Negative Cognitions.

Because the model that included the interaction of interpersonal stress and cognitions was significant, the main effect model for interpersonal stressors
is not presented.
p .10. p .05. p .01. p .001.

a
from the persons mean level of that variable; positive scores

represented higher than average levels of the variable and negative
scores represented lower than average levels of the variable. In
addition, individuals mean levels of stressors and negative cognitions were included at level-2 for a more complete comparison of
the variance in depressive symptoms associated with yearly fluctuations in negative cognitions and stressors versus mean levels of
these variables.
A series of models that included the substantive predictors was
evaluated. Models containing the three-way interactions among
stress, cognitions, and the potential moderators (sex, risk, current
maternal depressive symptoms) were tested first. If no three-way
interactions were found, then models containing the two-way
interactions between stressors and cognitions were examined. For
significant interactions, we conducted tests of simple slopes to
60
50
High Negative Cognitions
40
Low Negative Cognitions
Depressive
Episode Hazard 30
Function
20
10
0
Low
High
Interpersonal Stressors
Figure 1. The interaction between interpersonal stressors and negative

cognitions was significant. The first onset of a major depressive episode
(MDE) was predicted by high negative cognitions in the context of low
interpersonal stress, and also by high levels of interpersonal stressors at
both high and low levels of negative cognitions.
determine whether the slopes were significantly different from

zero and graphed model results using prototypical values of scores
one standard deviation below and one standard deviation above the
mean. Control variables were sex, risk, and the prior years depression scores.
Results.
First, the series of three-way interactions among
stressor type, negative cognitions, and sex, risk, or current maternal depressive symptoms were not significant. Next, two models
were tested in which the two-way interactions between each type
of stressor and the negative cognitions composite were examined
(see Table 5). Depressive symptom scores from the prior year were
a significant negative predictor (B .19, p .001), which is
consistent with regression to the mean. Significant main effects
were found for interpersonal stress (B .19, p .05) and negative
cognitions (B .92, p .01); the interaction between interpersonal stress and negative cognitions also was significant (B
.31, SE .14, p .05). This interaction was plotted using one
standard deviation above and below the mean on component
variables to represent high and low scores (see Figure 2).5 Simple
slopes analyses revealed that the relation between interpersonal
stressors and depressive symptoms was significant at high levels of
cognitive vulnerability, t 3.18, p .01, but not at low levels of
cognitive vulnerability (t .28, ns). That is, among adolescents
with more negative cognitions, higher levels of interpersonal stress
predicted higher levels of depressive symptoms; at low levels of
negative cognitions, the relation between stress and depression was
The Y-axis includes negative values because the scores were centered
around each individuals mean scores over the course of the study.
789
Table 5
Multi-Level Model Results of the 2-Way Interaction Between Stressor Type and Negative
Cognitions Predicting Depressive Symptoms
Two-way interactions
Interpersonal
Stressor type
Initial status
Level 1
Age
Prior year Depressive Symptoms (CDI)
Stress
Negative Cognitions
Stress Negative Cognitions
Level 2
Risk
Person-Mean Stress
Person-Mean Negative Cognitions
Achievement
t
2.46
.01
.01
.10
2.36
1.81
1.91
4.38
2.48
3.46
2.19
.15
.19
.02
.84
1.06
1.43
4.38
.05
3.14
1.34
.05
.10
.50
.02
.40
.07
1.88
.16
.19
.19
.92
.31
.06
.58
.38
Note. CDI Childrens Depression Inventory.

p .10. p .05. p .01. p .001.
not significant.6 With regard to achievement stress, analyses revealed significant main effects for previous years depression
scores (B .19, p .001) and for negative cognitions (B .84,
p .01); the interaction between achievement stress and negative
cognitions was not significant (B 1.06, SE .79, ns), however.
To allow for more directs comparisons to other studies in the
literature (e.g., Abela & McGirr, 2007; Shih, 2006), we conducted
additional analyses using alternative methods of specifying the
time-varying predictors. Because the choice of centering methods
potentially provides different information about the effects of
time-varying predictors, we also tested grand-mean and personmean centered variables, Time 1-centered variables, and variables
that were not centered at all (Singer & Willett, 2003). The method
followed by Abela and Skitch (2007) in which only stressors, but
not cognitions or depressive symptoms were person-centered also
was tested. Using these alternative methods of variable specification, however, did not yield significant interaction effects.
1.5
Depressive
Symptoms
High Neg Cognitions
0.5
Low Neg Cognitions
-0.5
-1
-1.5
Finally, we conducted a supplemental analysis to explore

whether using the same index of cognitive vulnerability (i.e., Wave
1 negative cognitions composite) as was used in the DTHM
analyses would produce similar results in the MLM analyses. No
significant three- or two-way interactions were found in these
MLM analyses, but Wave 1 negative cognitions significantly predicted depressive symptoms (B .34, p .05) in the achievement
stressor model, which is consistent with what was found in the
DTHM analyses predicting an MDE onset.
Discussion
Several important findings emerged from this 6-year longitudinal study of predictors of both the first onset of a major depressive
episode (MDE) and changes in depressive symptoms during adolescence. First, evidence consistent with the cognitive-stress interaction model (Abramson et al., 1978, 1989; Beck, 1967; Brown &
Harris, 1978) was found using two different data analytic approaches (i.e., DTHM and MLM) and two different outcomes (i.e.,
onset of an MDE and changes in depressive symptoms). That is,
the relation between interpersonal stressors and depression varied
significantly as a function of level of negative cognitions. The first
onset of an MDE was predicted by either high negative cognitions
in the context of low interpersonal stress or high interpersonal
stress regardless of level of negative cognitions. Under conditions
of both low interpersonal stress and low negative cognitions,
6
Low
High
Interpersonal Stress
Figure 2. The interaction between interpersonal stressors and negative

cognitions significantly predicted depressive symptoms; among adolescents with more negative cognitions, higher levels of interpersonal stress
predicted higher levels of depressive symptoms; at low levels of negative
cognitions, the relation between stress and depression was not significant.
Analyses of the individual cognitive measures (i.e., self-worth or

attributional style) revealed no significant three-way interactions with
stress, sex, risk, or maternal depressive symptoms. Significant two-way
interactions between global self-worth or attributional style and interpersonal, but not achievement stressors were found. The nature of these
interactions paralleled those with the negative cognitive composite; among
youth with more negative self-worth (B .63, SE .17, p .001) or more
depressive attributional style (B .46, SE .22, p .05), higher levels
of interpersonal stressors were associated with higher levels of depressive
symptoms.
790
CARTER AND GARBER
however, the odds of an onset of a depressive episode were not

significant. These results can be placed within a titration framework (Abramson, Alloy, & Metalsky, 1995, p. 118), such that
when one risk factor (e.g., negative cognitions) is high, the outcome may occur at lower levels of another risk factor (e.g.,
interpersonal stress). Thus, to be truly protected from having an
episode of depression, low levels of both interpersonal stress and
negative cognitions may be needed.
The cognitive-stress interaction also was significant when predicting depressive symptoms across six years of adolescence.
Among youth with more negative cognitions, higher levels of
interpersonal stress predicted higher levels of depressive symptoms; at low levels of negative cognitions, the relation between
stress and depression was not significant. Although the pattern of
the relations among interpersonal stressors, cognitions, and depression differed somewhat depending on the outcome (i.e., onset of an
MDE or change in depressive symptoms), overall these findings
are consistent with other studies that have provided support for the
cognitive-stress model in children and adolescents in relation to
depressive diagnoses (Lewinsohn et al., 2001) and depressive
symptoms (cf., Abela & Hankin, 2008).
The few prior studies (Hammen, 1988; Lewinsohn et al., 2001)
that specifically tested the cognitive vulnerability model in the
prediction of the diagnosis of a depressive episode in youth have
yielded mixed results. Whereas Hammen (1988; Hammen et al.,
1988) did not find that the cognitive-stress model predicted the
onset of a depressive disorder in a small high-risk sample of
children, Lewinsohn et al. (2001) showed that the cognitive-stress
interaction predicted the onset of depressive disorders, but not
depressive symptoms in a large community sample of adolescents.
As noted earlier, these studies differed in several ways including
sample size, age of the participants (i.e., adolescents or mixed
children and adolescents), type of samples (i.e., high risk vs.
community), and in their measures of stressful life events and
cognitive vulnerability. To address some of the limitations of these
earlier studies, the current investigation used a larger high-risk
sample of same-age youth followed over six years and measured
stress with a contextual threat interview. It also is possible that the
different approaches used to model longitudinal data contributed to
the discrepant findings. The current study found a significant
diathesis-stress interaction when using an idiographic approach for
all three primary variablesstressors, negative cognitions, and
depressive symptoms, whereas the other studies used betweenperson comparisons of these variables.
The results of this study provide the clearest evidence to date in
an adolescent sample that the cognitive vulnerability-stress model
predicts the first onset of a major depressive episode as well as
increases in depressive symptoms. These findings add to evidence
from studies with adults that the cognitive-stress interaction predicts depressive diagnoses (Alloy et al., 2006; Hankin et al., 2004;
Haeffel et al., 2003). Interestingly, studies of children have tended
not to find support for the cognitive-stress interaction model until
around early adolescence (Nolen-Hoeksema et al., 1992; Turner &
Cole, 1994). In the present study, the age of participants and the
duration of the follow-up period were selected specifically to
capture the rise in the rates of depression from early to middle
adolescence (e.g., Hankin et al., 1998). Thus, this study highlights
the importance of targeting the appropriate developmental window
in which to test the cognitive-stress model.
The findings that negative cognitions significantly interacted

with interpersonal stressors to predict both depressive disorder and
symptoms may be relevant to the debate about the continuity of
depression (e.g., Flett, Vredenburg, & Krames, 1997). According
to the continuity hypothesis, the same risk factors and processes
that contribute to mild levels of depressive symptoms also may
underlie clinical depression (Hankin, Fraley, Lahey, & Waldman,
2005; Ruscio & Ruscio, 2000). Results of the current study are
consistent with this continuity perspective, although it is possible
that other potential etiologic processes not examined here (e.g.,
genes, neurobiology) might show a more discontinuous pattern
across severity levels.
Stressors
The current investigation tested the cognitive-stress model separately for interpersonal and achievement stressors because not all
types of life events are similarly related to depression (Monroe &
Simons, 1991). Both theory (e.g., Gotlib & Hammen, 1992; Joiner
& Coyne, 1999) and empirical studies have highlighted the special
salience of social stressors (e.g., romantic break-ups) with regard
to depressive disorders (Rudolph et al., 2000) and depressive
symptoms (Rudolph & Hammen, 1999b). Consistent with this
literature, we found that interpersonal stressors, in particular, interacted with negative cognitions to predict both the first onset of
an MDE and changes in depressive symptoms. Interpersonal stressors are more common during adolescence and are associated with
earlier pubertal timing in both males and females (Rudolph, 2008).
Thus, one explanation for the rise in depressive symptoms and
diagnoses in adolescence may be an increase in the experience of
interpersonal stressors during this developmental period, particularly for girls (Natsuaki et al., 2009; Rudolph, 2002). In contrast,
there was a significant main effect of achievement stressors predicting the first onset of an MDE, but no interaction with negative
cognitions. The smaller number of achievement as compared to
interpersonal stressors reported by participants quite possibly resulted in a more restricted range of achievement events, which
could have reduced our power to detect the interaction.
Moreover, stronger support for the cognitive-stress model might
have been found if we had tested the specific vulnerability hypothesis (e.g., Beck, 1983; Coyne & Whiffen, 1995). According to this
perspective, when there is a match between the content of an
individuals cognitive style and the domain of a stressful event
(e.g., interpersonal, achievement), depression is especially likely to
occur. For example, individuals who believe that they have to be
perfect likely will become very distressed when they fail to accomplish a desired goal; persons who highly value being liked and
accepted by others are at increased risk for depression if they
encounter an interpersonal stressor such as rejection.
In addition, certain events that were objectively classified by
independent raters as being interpersonal or achievement-focused
may have been experienced quite differently by some participants.
For example, receiving a bad grade on an exam typically would be
classified as an achievement stressor, but for some youth a bad
grade might be an interpersonal stressor because for them, the
salient event is their parents disapproval. Without assessing the
subjective meaning of events for participants, we cannot be certain
that objectively rated life events reflect participants actual experiences.
Finally, in this study we categorized life events as interpersonal

or achievement, but other distinctions are possible. For example,
stressors also can be classified as dependent versus independent,
controllable versus uncontrollable, and chronic versus acute
(Grant, Compas, Thurm, McMahon, & Gipson, 2004; McMahon et
al., 2003; Rudolph et al., 2000). Thus, future tests of the cognitivestress model of depression in youth should explore other stressor
categories, match specific cognitions to stressors, and assess participants subjective experience of events as being interpersonal or
achievement-focused.
791
attributional style, higher levels of interpersonal stressors were

associated with higher levels of depressive symptoms. Thus, in
general the results for global self-worth and attributional style
examined separately were similar to those reported with the negative cognitions composite variable. This does not mean, however,
that a composite index of negative cognitions necessarily should
be used in all future studies. Rather, the composite cognitive
measure used in the current study was appropriate given the
particular questions being addressed.
Risk: Maternal Depression

Cognitive Vulnerability
In the present study, we used a composite index of the cognitive
vulnerability for several reasons. First, evidence of considerable
overlap of the cognitive vulnerability-stress components in HT
[Hopelessness Theory] and BT [Becks Theory] has been reported
(Hankin et al., 2004, pp. 321, 330). Various measures of negative
cognitions have been found to be highly correlated (e.g., Abela et
al., 2006; Adams, Abela, & Hankin, 2007) and may represent a
common underlying construct that forms a latent factor (Ginsburg
et al., 2009; Gotlib et al., 1993). In the current study, at each time
point the correlations between self-worth and attributional styles
were moderate and significant, and the composite index had good
internal consistency. Thus, the composite variable was a more
reliable measure of negative cognitions that reduced error found in
either single measure alone.
Second, perceived self-worth and attributional style represent
constructs that are fundamental to several of the leading theories of
depression (Abramson et al., 1978, 1989; Beck,1967; Brown et al.,
1987; Southall & Roberts, 2002). Third, other studies in both
adults (e.g., Alloy et al., 2006) and children (e.g., Turner & Cole,
1994) have created a composite cognitive vulnerability index
similar to the one used here. Finally, whereas others (e.g., Abela,
2001; Haeffel et al., 2003; Hankin et al., 2004; Lewinsohn et al.,
2001) have explicitly compared various cognitions as prescribed
by the different cognitive models, the aim of the current study was
to address a different set of questions regarding specific stressors,
moderators, and measures of depression. Given the various analyses required to address these questions, it made sense to reduce
the number of constructs by creating this composite cognitive
index.
Nevertheless, to check whether the cognitive composite yielded
similar results as its individual components, we conducted separate
exploratory analyses of global self-worth and attributional style
that paralleled the analyses done with the cognitive composite
variable. For the prediction of the onset of an MDE, the interaction
between interpersonal stressors and global self-worth was significant; the interaction with attributional style was not significant,
however, possibly due to the lower internal consistency of the
measure of attributional style used here. Perhaps more reliable
measures such as the Childrens Attributional Style Interview
(Conley et al., 2001) or the Adolescent Cognitive Style Questionnaire (Hankin & Abramson, 2002) would yield stronger results
with respect to attributional or inferential style.
For the prediction of depressive symptoms, significant interactions were found between interpersonal stressors and both global
self-worth and attributional style when analyzed separately. That
is, among youth with either negative self-worth or a depressive
Results of this study were consistent with the extant literature on

children of depressed mothers (e.g., Beardslee et al., 1998; Goodman & Gotlib, 1999). Offspring of depressed mothers (high-risk)
had significantly greater odds of having a first onset of a depressive episode during the six years of the study as compared to
children of nondepressed mothers (low-risk). High-risk adolescents also had significantly more negative cognitions and higher
depression scores at each wave, and higher levels of interpersonal
stress in over 60% of the months across the six years. Thus, the
high- and low-risk groups differed significantly in their levels of
negative cognitions, stress, and depressive disorders and symptoms.
Interestingly, levels of cognitive vulnerability and stressors contributed to the prediction of an MDE onset, over and above risk.
Despite level differences in negative cognitions and stress, however, the relations among these variables were not moderated by
maternal depression history (i.e., risk) or mothers current level of
depressive symptoms. That is, the three-way interactions of risk,
stress, and cognitions were not significant when predicting the first
onset of an MDE or changes in depressive symptoms during
adolescence. Similarly, mothers current level of depressive symptoms did not interact with stress and negative cognitions to predict
adolescents depressive symptoms. Thus, the cognitive-stress interaction predicted depression in adolescents and did not vary as a
function of mothers prior or current depression.
The cognitive vulnerability model of depression may be robust
for adolescents at various levels of risk associated with past or
current maternal depression. On the other hand, the absence of a
significant interaction could have been the result of other factors
such as insufficient power to detect an effect due to a small sample
size, particularly in the low-risk group. Other studies of the cognitive vulnerability-stress interaction in high-risk samples have not
explicitly tested for the moderating effects of maternal depression
(Abela & Skitch, 2007; Hammen, 1988). Additionally, other characteristics of maternal depression (e.g., episode duration, chronicity, recurrence) could be explored as potential moderators of the
cognitive-stress relation when predicting childrens depression, or
as possible mediators of the association between maternal depression history and adolescents depressive symptoms (e.g., Halligan,
Murray, Martins, & Cooper, 2007; Pawlby, Hay, Sharp, Waters, &
OKeane, 2009).
Sex Differences
Sex differences in levels of study variables or in the relations
among them were not observed in the current study. These results
are consistent with other studies that have not found sex differ-
792
CARTER AND GARBER
ences in negative cognitions (e.g., Hankin et al., 2001; Thompson et

al., 1998), but are not congruent with studies in which girls and boys
have been shown to differ in their cognitive style (e.g., Calvete &
Cardenoso, 2005; Gladstone et al., 1997; Hankin & Abramson, 2002)
or levels of stress (e.g., Rudolph, 2002; Rudolph & Hammen, 1999).
Results of studies testing whether sex moderates the cognitive-stress
interaction also have been inconsistent (e.g., Abela, 2001; Abela &
Payne, 2003; Prinstein & Aikins, 2004; Prinstein et al., 2005). The
present prospective study found that sex did not significantly moderate the relations between negative cognitions and stress in the prediction of depressive episodes or symptoms. Although sex differences
have been found in the rates and chronicity of depression beginning in
adolescence (Costello, Erkanli, & Angold, 2006; Essau, Lewinsohn,
Seeley, & Sasagawa, 2010; Hankin et al., 1998) and in mean levels of
negative cognitions and stress (e.g., Hankin & Abramson, 2002;
Rudolph, 2002), the relations among these constructs could be similar
for males and females (Nolen-Hoeksema & Girgus, 1994). In particular, certain interpersonal stressors such as romantic break-ups, which
are common during adolescence, are related to depression in both
males and females (Monroe, Rohde, Seeley, & Lewinsohn, 1999).
Moreover, youth at increased familial risk for depression, as in the
current sample, might be especially vulnerable to depression regardless of gender (e.g., Brennan, Le Brocque, & Hammen, 2003). That is,
girls and boys who are exposed to similar risk factors, such as
maternal depression and stress, may be equally likely to develop
depression. We need to be cautious, however, not to over interpret
null effects, particularly given the challenges of finding significant
statistical interactions (McClelland & Judd, 1993).
dropped out had higher levels of achievement stress at Wave one

and were more likely to be male. This might have biased the results
and reduced the power to find significant effects for achievement
stress on depression by reducing its variability in the remaining
sample. Additionally, the higher attrition of males may have affected our ability to detect sex differences.
Finally, in the DTHM analyses, the measure of the cognitive
vulnerability was assessed at Wave 1 (Grade 6) only and therefore
might have been a less adequate predictor of depression than had
cognitions been assessed monthly across the six years. Nevertheless, this cognitive vulnerability index interacted significantly with
interpersonal stressors to predict the first onset of a depressive
episode, and was a significant direct predictor of an MDE in the
model with achievement stressors. Although more frequent and
proximal assessments of the cognitive vulnerability would have
been preferred, the cognitive measures were found to be moderately stable across time; that is, cognitions assessed at Wave 1
correlated significantly with those measured at subsequent time
points (ranging from .58 at wave 2 to .32 at wave 5).
We also explored whether this Wave 1 index of cognitive vulnerability used in the DTHM analyses would produce similar results in
the MLM analyses predicting depressive symptoms. Consistent with
the DTHM results, in the achievement stressor model the Wave 1
cognitive vulnerability index significantly predicted depressive symptoms; the interaction of Wave 1 cognitions with interpersonal stressors
was not significant, however. Future tests of the cognitive-stress
interaction in youth should measure cognitions as frequently as stressors to provide a more sensitive and temporally linked measure of the
cognitive vulnerability over time.
Limitations
Limitations of the current study provide directions for future
research. First, both cognitive vulnerability and depressive symptoms were assessed with self-report measures. In contrast, depressive disorders were diagnosed by clinicians and contextual stress
levels were rated by independent judges based on separate interviews with the adolescent and mother. Thus, although the relation
between negative cognitions and depressive symptoms could have
been inflated due to the use of a common assessment procedure
(i.e., self-report), the link between stressors and depressive disorders is less likely to have been the result of such shared method
variance. Alternative ways of assessing cognitions include
laboratory-based tasks, such as the attentional dot-probe and incidental recall tasks (Garber & Kaminski, 2000), have shown promise for examining cognitive vulnerability in the context of stress
(Jacobs, Reinecke, Gollan, & Kane, 2008).
Second, both a strength and a limitation of the current study was
that participants were recruited so as to oversample offspring of
mothers with histories of depression (high risk). This strategy was
used in order to have greater variability on the measures of interest
(e.g., stress, cognitions, and depression). Interestingly, over and
above the potent predictor of maternal depression, the cognitivestress interaction was still significant indicating that more variance
needed to be explained beyond risk. The results of the current
study, however, may not generalize to a purely community sample
or to offspring of depressed fathers (Connell & Goodman, 2002;
Kane & Garber, 2004).
Third, as with any longitudinal study, some participants were
lost over the six years. Attrition analyses revealed that those who
Clinical Implications
An important clinical implication of the current study is that
interventions that teach youth strategies for coping with negative
life events, particularly interpersonal stressors, may be especially
useful in treating and preventing depression. Indeed, interventions
that emphasize interpersonal communication (Young, Mufson, &
Davies, 2006), social skills (Gillham, Hamilton, Freres, Patton, &
Gallop, 2006; Reinecke, Ryan, & DuBois, 1998), and coping
(Compas et al., 2009) have been found to be efficacious in the
treatment and prevention of depression in youth. Approaches that
integrate both cognitive and interpersonal strategies may be especially effective and therefore should be the focus of future intervention efforts (Garber, 2006; Jaycox, Reivich, Gillham, & Seligman, 1994). Finally, results of the current study highlight the need
for the continued development and dissemination of depression
prevention programs that target cognitive restructuring and coping
with stress, particularly in offspring of depressed parents (e.g.,
Compas et al., 2009; Garber et al., 2009).
References
Abela, J. R. Z. (2001). The hopelessness theory of depression: A test of the
diathesis-stress and causal mediation components in third and seventh
grade children. Journal of Abnormal Child Psychology, 29, 241254.
doi:10.1023/A:1010333815728
Abela, J. R. Z., Aydin, C., & Auerbach, R. P. (2006). Operationalizing the
vulnerability and stress components of the hopelessness theory of
depression: A multi-wave longitudinal study. Behavior Research and
Therapy, 44, 15651583. doi:10.1016/j.brat.2005.11.010

Abela, J. R. Z., & Hankin, B. L. (2008). Cognitive vulnerability to depression in
children and adolescence: A developmental psychopathology approach. In
J. R. Z. Abela & B. L. Hankin (Eds.), Handbook of child and adolescent
depression (pp. 3578). New York, NY: Guilford Press.
Abela, J. R. Z., & McGirr, A. (2007). Operationalizing cognitive vulnerability and stress from the perspective of the hopelessness theory: A
multi-wave longitudinal study of children of affectively ill parents.
British Journal of Clinical Psychology, 46, 377395. doi:10.1348/
014466507X192023
Abela, J. R. Z., & Payne, A. V. L. (2003). A test of the integration of the
hopelessness and self-esteem theories of depression in schoolchildren.
Cognitive Therapy and Research, 27, 519 535. doi:10.1023/A:
1026303020478
Abela, J. R. Z., & Skitch, S. A. (2007). Dysfunctional attitudes, selfesteem, and hassles: Cognitive vulnerability to depression in children of
affectively ill parents. Behaviour Research and Therapy, 45, 11271140.
doi:10.1016/j.brat.2006.09.011
Abela, J. R. Z., Skitch, S. A., Adams, P., & Hankin, B. L. (2006). The
timing of parent and child depression: A hopelessness theory perspective. Journal of Clinical Child and Adolescent Psychology, 35, 253263.
Abela, J. R. Z., & Veronneau-McArdle, M. (2002). The relationship
between self-complexity and depressive symptoms in third and seventh
grade children: A short-term longitudinal study. Journal of Abnormal
Child Psychology, 30, 155166. doi:10.1023/A:1014705316224
Abramson, L. Y., Alloy, L. B., & Metalsky, G. I. (1995). Hopelessness
depression. In G. M. Buchanan & M. E. P. Seligman (Eds.), Explanatory
style (pp. 113134). Hillsdale, NJ: Erlbaum.
Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness
depression: A theory-based subtype of depression. Psychological Review, 96, 358 372. doi:10.1037/0033-295X.96.2.358
Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D. (1978). Learned
helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology, 87, 49 74. doi:10.1037/0021-843X.87.1.49
Adams, P., Abela, J. R. Z., & Hankin, B. L. (2007). Factorial categorization
of depression-related constructs in early adolescents. Journal of Cognitive Psychotherapy, 21, 123139. doi:10.1891/088983907780851540
Aiken, L. S., West, S. G., & Reno, R. R. (1991). Multiple regression:
Testing and interpreting interactions: Thousand Oaks, CA: Sage.
Akaike, H. (1973). Information theory as an extension of the maximumlikelihood principle. In B. N. Petrov & F. Caski (Eds.), Second International Symposium on Information Theory (pp. 267328). Budapest,
Hungary: Akademiai Kiado.
Allgood-Merten, B., Lewinsohn, P. M., & Hops, H. (1990). Sex differences
in adolescent depression. Journal of Abnormal Psychology, 99, 55 63.
doi:10.1037/0021-843X.99.1.55
Alloy, L. B., Abramson, L. Y., Whitehouse, w. G., Hogan, M. E., Panzarella, C., & Rose, D. T. (2006). Prospective incidence of first onsets and
recurrences of depression in individuals at high and low cognitive risk
for depression. Journal of Abnormal, 115, 145156. doi:10.1037/0021843X.115.1.145
Alloy, L. B., Lipman, A. J., & Abramson, L. Y. (1992). Attributional style
as a vulnerability factor for depression: Validation by past history of
mood disorders. Cognitive Therapy and Research, 16, 391 407. doi:
10.1007/BF01183164
Ambrosini, P. J. (2000). Historical development and present status of the
schedule for affective disorders and schizophrenia for school-age children (K-SADS). Journal of the American Academy of Child & Adolescent Psychiatry, 39, 49 58. doi:10.1097/00004583-200001000-00016
Avenevoli, S., & Steinberg, L. (2001). The continuity of depression across
the adolescent transition. In H. Reese & R. Kail (Eds.), Advances in child
development and behavior (Vol. 28, pp. 139 173). New York, NY:
Academic Press.
Barnett, P. A., & Gotlib, I. H. (1988). Psychosocial functioning and
depression: Distinguishing among antecedents, consequences, concom-
793
itants, and consequences. Psychological Bulletin, 104, 97126. doi:

10.1037/0033-2909.104.1.97
Beardslee, W. R., Versage, E. M., & Gladstone, T. R. (1998). Children of
affectively ill parents: A review of the past 10 years. Journal of the
American Academy of Child and Adolescent Psychiatry, 37, 1134 1141.
doi:10.1097/00004583-199811000-00012
Beck, A. T. (1967). Depression: Clinical, experiential, and theoretical
aspects. New York, NY: Harper & Row.
Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New
York: Meridian.
Beck, A. T. (1991). Cognitive therapy: A 30-year retrospective. American
Psychologist, 46, 368 375.
Beck, A. T. (1983). Cognitive therapy of depression: New perspectives. In P. J.
Clayton & J. E. Barrett (Eds.), Treatment of depression: Old controversies and
new approaches (pp. 265290). New York, NY: Raven Press.
Beck, A. T., Ward, C. F., Mendelson, M., Mock, J., & Erlbaugh, J. (1961).
An inventory for measuring depression. Archives of General Psychiatry,
4, 53 63.
Brennan, P. A., Le Brocque, R., & Hammen, C. (2003). Maternal depression, parent child relationships and resilient outcomes in adolescence.
Journal of the American Academy of Child & Adolescent Psychiatry, 42,
1469 1477. doi:10.1097/00004583-200312000-00014
Brown, G. W., Bifulco, A., & Harris, T. O. (1987). Life events, vulnerability and onset of depression: Some refinements. The British Journal of
Psychiatry, 150, 30. doi:10.1192/bjp.150.1.30
Brown, G. W., & Harris, T. O. (1978). Social origins of depression: A
study of psychiatric disorder in women. New York, NY: Free Press.
Brown, G. W., & Harris, T. O. (1989). Depression. In G. W. Brown &
T. O. Harris (Eds.), Life Events and Illness (pp. 49 93). New York:
Guilford Press.
Calvete, E., & Cardenoso, O. (2005). Gender differences in cognitive vulnerability
to depression and behavior problems in adolescents. Journal of Abnormal Child
Psychology, 33, 179192. doi:10.1007/s10802-005-1826-y
Chambers, W. J., Puig-Antich, J., Hirsch, M., Paez, P., Ambrosini, P. J.,
Tabrizi, M. A., & Davies, M. (1985). The assessment of affective
disorders in children and adolescents by semistructured interview. Archives of General Psychiatry, 42, 696 702.
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20, 37 46. doi:10.1177/
001316446002000104
Cole, D. A., Hoffman, K., Tram, J. M., & Maxwell, S. E. (2000). Structural
differences in parent and child reports of childrens symptoms of depression and anxiety. Psychological Assessment, 12, 174 185. doi:
10.1037/1040-3590.12.2.174
Compas, B. E., Ey, S., & Grant, K. E. (1993). Taxonomy, assessment, and
diagnosis of depression during adolescence. Psychological Bulletin, 114,
323344. doi:10.1037/0033-2909.114.2.323
Compas, B. E., Forehand, R., Keller, G., Champion, J. E., Rakow, A.,
Reeslund, K. L., . . . Hardcastle, E. (2009). Randomized controlled trial
of a family cognitive-behavioral preventive intervention for children of
depressed parents. Journal of consulting and clinical psychology, 77,
1007. doi:10.1037/a0016930
Compas, B. E., Slavin, L. A., Wagner, B. M., & Vannatta, K. (1986).
Relationship to life events and social support with psychological dysfunction among adolescents. Journal of Youth and Adolescence, 15,
205221. doi:10.1007/BF02139123
Conley, C. S., Haines, B. A., Hilt, L. M., & Metalsky, G. I. (2001). The
Childrens Attributional Style Interview: Developmental tests of cognitive diathesis-stress theories of depression. Journal of Abnormal Child
Psychology, 29, 445 463. doi:10.1023/A:1010451604161
Connell, A. M., & Goodman, S. H. (2002). The association between
psychopathology in fathers versus mothers and childrens internalizing
and externalizing behavior problems: A meta-analysis. Psychological
Bulletin, 128, 746 773. doi:10.1037/0033-2909.128.5.746
794
CARTER AND GARBER
Costello, E. J., Erkanli, A., & Angold, A. (2006). Is there an epidemic of

child or adolescent depression? Journal of Child Psychology and Psychiatry, 47, 12631271.
Coyne, J. C., & Gotlib, I. H. (1983). The role of cognition in depression:
A critical appraisal. Psychological Bulletin, 94, 472505.
Coyne, J. C., & Whiffen, V. E. (1995). Issues in personality as diathesis for
depression: The case of sociotropydependency and autonomyself-criticism.
Psychological Bulletin, 118, 358378. doi:10.1037/0033-2909.118.3.358
Cummings, E. M., Keller, P. S., & Davies, P. T. (2005). Towards a family
process model of maternal and paternal depressive symptoms: Exploring
multiple relations with child and family functioning. Journal of Child
Psychology and Psychiatry, 46, 479 489. doi:10.1111/j.14697610.2004.00368.x
Duggal, S., Malkoff-Schwartz, S., Birmaher, B., Anderson, B. P., Matty,
M. K., Houck, P. R., . . . Frank, E. (2000). Assessment of life stress in
adolescents: Self-report versus interview methods. Journal of the American Academy of Child & Adolescent Psychiatry, 39, 445 452. doi:
10.1097/00004583-200004000-00013
Essau, C. A., Lewinsohn, P. M., Seeley, J. R., & Sasagawa, S. (2010).
Gender differences in the developmental course of depression. Journal
of Affective Disorders, 127, 185190. doi:10.1016/j.jad.2010.05.016
Flett, G. L., Vredenburg, K., & Krames, L. (1997). The continuity of
depression in clinical and nonclinical samples. Psychological Bulletin,
121, 395 416. doi:10.1037/0033-2909.121.3.395
Foster, C., Webster, M., Weissman, M., Pilowsky, D., Wickramaratne, P., Rush,
A., . . . King, C. A. (2008). Course and severity of maternal depression: Associations with family functioning and child adjustment. Journal of Youth and
Adolescence, 37, 906916. doi:10.1007/s10964-007-9216-0
Garber, J. (2006). Depression in children and adolescents: Linking risk
research and prevention. American Journal of Preventive Medicine, 31,
104 125. doi:10.1016/j.amepre.2006.07.007
Garber, J. (2007). Depression in youth: A developmental psychopathology
perspective. In A. Masten & A. Sroufe (Eds.), Multilevel dynamics in
developmental psychopathology: Pathways to the future (Vol. 34, pp.
181242). New York, NY: Erlbaum.
Garber, J., Clarke, G., Weersing, V. R., Beardslee, W. R., Brent, D.,
Gladstone, T., . . . Iyengar, S. (2009). Prevention of depression in at-risk
adolescents: A randomized controlled trial. Journal of the American
Medical Association, 301, 22152224. doi:10.1001/jama.2009.788
Garber, J., & Kaminski, K. M. (2000). Laboratory and performance-based
measures of depression in children and adolescents. Journal of Clinical
Child & Adolescent Psychology, 29, 509 525. doi:10.1207/
S15374424JCCP2904_5
Garber, J., Keiley, M. K., & Martin, N. C. (2002). Developmental trajectories of adolescents depressive symptoms: Predictors of change. Journal of Consulting and Clinical Psychology, 70, 79 95. doi:10.1037/
0022-006X.70.1.79
Garber, J., & Robinson, N. (1997). Development and validation of the Life
Events Interview for Adolescents. Paper presented at the biennial meeting of the Society for Research in Child Development, Washington, DC.
Gibb, B. E., & Alloy, L. B. (2006). A prospective test of the hopelessness
theory of depression in children. Journal of Clinical Child & Adolescent
Psychology, 35, 264 274. doi:10.1207/s15374424jccp3502_10
Gillham, J. E., Hamilton, J., Freres, D. R., Patton, K., & Gallop, R. (2006).
Preventing depression among early adolescents in the primary care
setting: A randomized controlled study of the Penn Resiliency Program.
Journal of Abnormal Child Psychology, 34, 203219. doi:10.1007/
s10802-005-9014-7
Ginsburg, G. S., Silva, S. G., Jacobs, R. H., Tonev, S., Hoyle, R. H.,
Kingery, J. N., . . . March, J. S. (2009). Cognitive measures of adolescent
depression: Unique or unitary constructs? Journal of Clinical Child &
Adolescent Psychology, 38, 790 802. doi:10.1080/15374410903259015
Gladstone, T. R., & Kaslow, N. J. (1995). Depression and attributions in
children and adolescents: A meta-analytic review. Journal of Abnormal

Child Psychology, 23, 597 606. doi:10.1007/BF01447664
Gladstone, T. R., Kaslow, N. J., Seeley, J. R., & Lewinsohn, P. M. (1997).
Sex differences, attributional styles, and depressive symptoms among
adolescents. Journal of Abnormal Child Psychology, 25, 297305. doi:
10.1023/A:1025712419436
Goodman, S. H., Adamson, L. B., Riniti, J., & Cole, S. (1994). Mothers
expressed attitudes: Associations with maternal depression and childrens self-esteem and psychopathology. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 12651274. doi:10.1097/
00004583-199411000-00007
Goodman, S. H., & Gotlib, I. H. (1999). Risk for psychopathology in the
children of depressed mothers: A developmental model for understanding mechanisms of transmission. Psychological review, 106, 458. doi:
10.1037/0033-295X.106.3.458
Gotlib, I. H., & Hammen, C. L. (1992). Psychological aspects of depression:
Toward a cognitive-interpersonal integration. New York, NY: Wiley.
Gotlib, I. H., Lewinsohn, P. M., Seeley, J. R., Rohde, P., & Redner, J. E.
(1993). Negative cognitions and attributional style in depressed adolescents: An examination of stability and specificity. Journal of Abnormal
Psychology, 102, 607 615. doi:10.1037/0021-843X.102.4.607
Grant, K. E., Compas, B. E., Thurm, A. E., McMahon, S. D., & Gipson, P. (2004).
Stressors and child and adolescent psychopathology: Measurement issues and
prospective effects. Journal of Clinical Child and Adolescent Psychology, 33,
412425. doi:10.1207/s15374424jccp3302_23
Grant, K. E., & McMahon, S. (2005). Conceptualizing the role of stressors
in the development of psychopathology. In B. L. Hankin & J. R. Z.
Abela (Eds.), Development of psychopathology: A vulnerability-stress
perspective. Thousand Oaks, CA: Sage.
Haeffel, G. J., Abramson, L. Y., Voelz, Z., Metalsky, G. I., Halberstadt, L.,
Dykman, B., . . . Alloy, L. B. (2003). Cognitive vulnerability to depression and lifetime history of axis I psychopathology: A comparison of
Negative Cognitive Styles (CSQ) and Dysfunctional Attitudes (DAS).
Journal of Cognitive Psychotherapy: An International Quarterly, 17,
322. doi:10.1891/jcop.17.1.3.58269
Halligan, S. L., Murray, L., Martins, C., & Cooper, P. J. (2007). Maternal
depression and psychiatric outcomes in adolescent offspring: A 13-year
longitudinal study. Journal of Affective Disorders, 97, 145154. doi:
10.1016/j.jad.2006.06.010
Hammen, C. (1988). Self-cognitions, stressful events, and the prediction of
depression in children of depressed mothers. Journal of Abnormal Child
Psychology, 16, 347360. doi:10.1007/BF00913805
Hammen, C. (1991). Generation of stress in the course of unipolar depression. Journal of Abnormal Psychology, 100, 555. doi:10.1037/0021843X.100.4.555
Hammen, C., Adrian, C., Gordon, D., Burge, D., Jaenicke, C., & Hiroto, D.
(1987). Children of depressed mothers: Maternal strain and symptom
predictors. Journal of Abnormal Psychology, 96, 190198. doi:10.1037/
0021-843X.96.3.190
Hammen, C., Adrian, C., & Hiroto, D. (1988). A longitudinal test of the
attributional vulnerability model in children at risk for depression. British Journal of Clinical Psychology, 27, 37 46. doi:10.1111/j.20448260.1988.tb00751.x
Hammen, C., Shih, J., Altman, T., & Brennan, P. A. (2003). Interpersonal
impairment and the prediction of depressive symptoms in adolescent
children of depressed and nondepressed mothers. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 571577. doi:
10.1097/01.CHI.0000046829.95464.E5
Hammen, C., Shih, J. H., & Brennan, P. A. (2004). Intergenerational
transmission of depression: Test of an interpersonal stress model in a
community sample. Journal of Consulting and Clinical Psychology, 72,
511522. doi:10.1037/0022-006X.72.3.511
Hankin, B. L., & Abramson, L. Y. (2001). Development of gender differences in depression: An elaborated cognitive vulnerability-transactional

stress theory. Psychological Bulletin, 127, 773796. doi:10.1037/00332909.127.6.773
Hankin, B. L., & Abramson, L. Y. (2002). Measuring cognitive vulnerability to depression in adolescence: Reliability, validity, and gender
differences. Journal of Clinical Child and Adolescent Psychology, 31,
491504.
Hankin, B. L., Abramson, L. Y., Miller, N., & Haeffel, G. J. (2004).
Cognitive vulnerability-stress theories of depression: Examining affective specificity in the prediction of depression versus anxiety in three
prospective studies. Cognitive Therapy and Research, 28, 309 345.
doi:10.1023/B:COTR.0000031805.60529.0d
Hankin, B. L., Abramson, L. Y., Moffitt, T. E., Silva, P. A., McGee, R., &
Angell, K. A. (1998). Development of depression from preadolescence
to young adulthood: Emerging gender differences in a 10 year longitudinal study. Journal of Abnormal Psychology, 107, 128 141. doi:
10.1037/0021-843X.107.1.128
Hankin, B. L., Abramson, L. Y., & Siler, M. (2001). A prospective test of
the hopelessness theory of depression in adolescence. Cognitive Therapy
and Research, 25, 607 632. doi:10.1023/A:1005561616506
Hankin, B. L., Fraley, R. C., Lahey, B. B., & Waldman, I. D. (2005). Is
youth depression best viewed as a continuum or discrete category? A
taxometric analysis of childhood and adolescent depression in a
population-based sample. Journal of Abnormal Psychology, 114, 96
110. doi:10.1037/0021-843X.114.1.96
Hankin, B. L., Mermelstein, R., & Roesch, L. (2007). Sex differences in
adolescent depression: Stress exposure and reactivity models. Child
Development, 78, 279 295. doi:10.1111/j.1467-8624.2007.00997.x
Harter, S. (1985). Self-Perception profile for children: Revision of the
Perceived Competence Scale for Children (manual). Denver, CO: University of Denver.
Hilsman, R., & Garber, J. (1995). A test of the cognitive diathesis-stress
model in children: Academic stressors, attributional style, perceived
competence and control. Journal of Personality and Social Psychology,
69, 370 380. doi:10.1037/0022-3514.69.2.370
Hollingshead, A. A. (1975). Four-factor index of social status. Unpublished manuscript, Yale University. New Haven, CT.
Holmbeck, G. N. (2002). Post-hoc probing of significant moderational and
mediational effects in studies of pediatric populations. Journal of Pediatric Psychology, 27, 8796. doi:10.1093/jpepsy/27.1.87
Hyde, J. S., Mezulis, A. H., & Abramson, L. Y. (2008). The ABCs of
depression: Integrating affective, biological, and cognitive models to
explain the emergence of the gender difference in depression. Psychological Review, 115, 291313. doi:10.1037/0033-295X.115.2.291
Jacobs, R. H., Reinecke, M. A., Gollan, J. K., & Kane, P. (2008). Empirical
evidence of cognitive vulnerability for depression among children and
adolescents: A cognitive science and developmental perspective. Clinical Psychology Review, 28, 759 782. doi:10.1016/j.cpr.2007.10.006
Jaenicke, C., Hammen, C., Zupan, B., Hiroto, D., Gordon, D., & Adrian, C.
(1987). Cognitive vulnerability in children at risk for depression. Journal of
Abnormal Child Psychology, 15, 559572. doi:10.1007/BF00917241
Jaser, S. S., Langrock, A. M., Keller, G., Merchant, M. J., Benson, M. A.,
Reeslund, K., . . . Compas, B. E. (2005). Coping with the stress of
parental depression II: Adolescent and parent reports of coping and
adjustment. Journal of Clinical Child and Adolescent Psychology, 34,
193205. doi:10.1207/s15374424jccp3401_18
Jaycox, L. H., Reivich, K. J., Gillham, J., & Seligman, M. E. P. (1994).
Prevention of depressive symptoms in school children. Behaviour Research and Therapy, 32, 801 816. doi:10.1016/0005-7967(94)90160-0
Joiner, T. E., & Coyne, J. C. (1999). The interactional nature of depression: Advances in interpersonal approaches. Washington, DC: American Psychological Association. doi:10.1037/10311-000
Kane, P. P., & Garber, J. (2004). The relation between fathers depression
and childrens externalizing and internalizing symptoms and conflict: A
795
meta-analysis. Clinical Psychology Review, 24, 339 360. doi:10.1016/

j.cpr.2004.03.004
Kaufman, J., Birmaher, B., Brent, D., Rao, U., Hammen, C., Daley, S. E.,
Flynn, C., Moreci, P., . . . Ryan, N. (1997). Schedule for Affective
Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL): Initial reliability and validity data. Journal
of the American Academy of Child and Adolescent Psychiatry, 36,
980 988. doi:10.1097/00004583-199707000-00021
Keller, M. B., Lavori, P. W., Friedman, B., Nielsen, E., Endicott, J.,
McDonald-Scott, P., . . . Andreasen N. C. (1987). The Longitudinal
Interval Follow-up Evaluation: A comprehensive method for assessing
outcome in prospective longitudinal studies. Archives of General Psychiatry, 44, 540 548.
Kendler, K. S., Karkowski, L. M., & Prescott, C. A. (1999). Causal
relationship between stressful life events and the onset of major depression. American Journal of Psychiatry, 156, 837.
Kovacs, M. (1981). Rating scales to assess depression in school-aged
children. Acta Paedopsychiatrica: International Journal of Child and
Adolescent Psychiatry, 46, 305315.
Kovacs, M. (1985). The Childrens Depression Inventory. Psychopharmacology Bulletin, 21, 995998.
Kraaij, V., Garnefski, N., van de Wilde, E. J., Dijkstra, A., Gebhardt, W.,
Maes, S., & ter Doest, L. (2003). Negative life events and depressive
symptoms in late adolescence: Bonding and cognitive coping as vulnerability factors? Journal of Youth and Adolescence, 32, 185193. doi:
10.1023/A:1022543419747
Lakdawalla, Z., Hankin, B. L., & Mermelstein, R. (2007). Cognitive
theories of depression in children and adolescents: A conceptual and
quantitative review. Clinical Child and Family Psychology Review, 10,
124. doi:10.1007/s10567-006-0013-1
Langrock, A. M., Compas, B. E., Keller, G., Merchant, M. J., & Copeland,
M. E. (2002). Coping with the stress of parental depression: Parents
reports of childrens coping, emotional, and behavioral problems. Journal of Clinical Child and Adolescent Psychology, 31, 312324.
Lewinsohn, P. M., Joiner, T. E., Jr., & Rohde, P. (2001). Evaluation of
cognitive diathesisstress models in predicting Major Depressive Disorder in adolescents. Journal of Abnormal Psychology, 110, 203215.
doi:10.1037/0021-843X.110.2.203
Lewinsohn, P. M., Rohde, P., & Gau, J. M. (2003). Comparability of
self-report checklist and interview data in the assessment of stressful life
events in young adults. Psychological Reports, 93, 459 471. doi:
10.2466/PR0.93.6.459-471
Little, S. A., & Garber, J. (2004). Interpersonal and achievement
orientations and specific stressors predict depressive and aggressive
symptoms. Journal of Adolescent Research, 19, 63 84. doi:10.1177/
0743558403258121
McClelland, G. H., & Judd, C. M. (1993). Statistical difficulties of detecting interactions and moderator effects. Psychological Bulletin, 114,
376 376. doi:10.1037/0033-2909.114.2.376
McMahon, S. D., Grant, K. E., Compas, B. E., Thurm, A. E., & Ey, S.
(2003). Stress and psychopathology in children and adolescents: Is there
evidence of specificity? The Journal of Child Psychology and Psychiatry
and Allied Disciplines: Annual Research Review, 44, 107133.
McQuaid, J. R., Monroe, S. M., Roberts, J. E., Kupfer, D. J., & Frank, E. (2000).
A comparison of two life stress assessment approaches: Prospective prediction
of treatment outcome in recurrent depression. Journal of Abnormal Psychology,
109, 787791. doi:10.1037/0021-843X.109.4.787
Monroe, S. M., Rohde, P., Seeley, J. R., & Lewinsohn, P. M. (1999). Life events
and depression in adolescence: Relationship loss as a prospective risk factor for
first onset of major depressive disorder. Journal of Abnormal Psychology, 108,
606614. doi:10.1037/0021-843X.108.4.606
Monroe, S. M., & Simons, A. D. (1991). Diathesis-stress theories in the context of
life stress research: Implications for the depressive disorders. Psychological
Bulletin, 110, 406425. doi:10.1037/0033-2909.110.3.406
796
CARTER AND GARBER
Morris, M. C., Ciesla, J. A., & Garber, J. (2010). A prospective study of

stress autonomy versus stress sensitization in adolescents at varied risk
for depression, Journal of Abnormal Psychology, 119, 341354. doi:
10.1037/a0019036
Natsuaki, M. N., Klimes-Dougan, B., Ge, X., Shirtcliff, E. A., Hastings,
P. D., & Zahn-Waxler, C. (2009). Early pubertal maturation and internalizing problems in adolescence: Sex differences in the role of cortisol
reactivity to interpersonal stress. Journal of Clinical Child & Adolescent
Psychology, 38, 513524. doi:10.1080/15374410902976320
Nolen-Hoeksema, S., & Girgus, J. S. (1994). The emergence of gender
differences in depression during adolescence. Psychological Bulletin,
115, 424 443. doi:10.1037/0033-2909.115.3.424
Nolen-Hoeksema, S., Girgus, J. S., & Seligman, M. E. (1992). Predictors
and consequences of childhood depressive symptoms: A 5-year longitudinal study. Journal of Abnormal Psychology, 101, 405 422. doi:
10.1037/0021-843X.101.3.405
Nunnally, J. C., & Bernstein, I. H.(1994). Psychometric Theory. New
York, NY: McGraw-Hill.
Panak, W. F., & Garber, J. (1992). Role of aggression, rejection, and
attributions in the prediction of depression in children. Development and
Psychopathology, 4, 145165. doi:10.1017/S0954579400005617
Pawlby, S., Hay, D. F., Sharp, D., Waters, C. S., & OKeane, V. (2009).
Antenatal depression predicts depression in adolescent offspring: Prospective longitudinal community-based study. Journal of Affective Disorders, 113, 236 243. doi:10.1016/j.jad.2008.05.018
Prinstein, M. J., & Aikins, J. W. (2004). Cognitive moderators of the
longitudinal association between peer rejection and adolescent depressive symptoms. Journal of Abnormal Child Psychology, 32, 147158.
doi:10.1023/B:JACP.0000019767.55592.63
Prinstein, M. J., Cheah, C. S. L., & Guyer, A. E. (2005). Peer victimization, cue interpretation, and internalizing symptoms: Concurrent
and longitudinal findings for children and adolescents. Journal of
Clinical Child and Adolescent Psychology, 34, 1124. doi:10.1207/
s15374424jccp3401_2
Reinecke, M. A., Ryan, N. E., & DuBois, D. L. (1998). Cognitivebehavioral therapy of depression and depressive symptoms during adolescence: A review and meta-analysis. Journal of the American Academy
of Child and Adolescent Psychiatry, 37, 26 34. doi:10.1097/00004583199801000-00013
Roberts, J. E., & Monroe, S. M. (1994). A multidimensional model of
self-esteem in depression. Clinical Psychology Review, 14, 161181.
doi:10.1016/0272-7358(94)90006-X
Roberts, J. E., & Monroe, S. M. (1999). Vulnerable self-esteem and social
processes in depression: Toward an interpersonal model of self-esteem
regulation.
Robins, C. J., & Hinkley, K. (1989). Social-cognitive processing and
depressive symptoms in children: A comparison of measures. Journal of
Abnormal Child Psychology, 17, 29 36. doi:10.1007/BF00910768
Rudolph, K. D. (2002). Gender differences in emotional responses to
interpersonal stress during adolescence. Journal of Adolescent Health,
30, 313. doi:10.1016/S1054-139X(01)00383-4
Rudolph, K. D. (2008). Developmental influences on interpersonal stress
generation in depressed youth. Journal of Abnormal Psychology, 117,
673. doi:10.1037/0021-843X.117.3.673
Rudolph, K. D., & Hammen, C. (1999). Age and gender as determinants of stress
exposure, generation, and reactions in youngsters: A transactional perspective.
Child Development, 70, 660677. doi:10.1111/1467-8624.00048
Rudolph, K. D., Hammen, C., & Burge, D. (1997). A cognitiveinterpersonal approach to depressive symptoms in preadolescent children. Journal of Abnormal Child Psychology, 25, 33 45. doi:10.1023/
A:1025755307508
Rudolph, K. D., Hammen, C., Burge, D., Lindberg, N., Herzberg, D., &
Daley, S. E. (2000). Toward an interpersonal life-stress model of depression: The developmental context of stress generation. Development
and Psychopathology, 12, 215234. doi:10.1017/S0954579400002066
Ruscio, J., & Ruscio, A. M. (2000). Informing the continuity controversy:
A taxometric analysis of depression. Journal of Abnormal Psychology,
109, 473 487. doi:10.1037/0021-843X.109.3.473
Sandberg, S., Rutter, M., Pickles, A., McGuinness, D., & Angold, A.
(2001). Do high-threat life events really provoke the onset of psychiatric
disorder in children? Journal of Child Psychology and Psychiatry, 42,
523532. doi:10.1111/1469-7610.00746
Seiffge-Krenke, I., & Stemmler, M. (2002). Factors contributing to gender
differences in depressive symptoms: A test of three developmental
models. Journal of Youth and Adolescence, 31, 405 417. doi:10.1023/
A:1020269918957
Seligman, M. E. P., Peterson, C., Kaslow, N. J., Tanenbaum, R. L., Alloy,
L. B., & Abramson, L. Y. (1984). Attributional style and depressive
symptoms among children. Journal of Abnormal Psychology, 93, 235
238. doi:10.1037/0021-843X.93.2.235
Shih, J. H. (2006). Sex differences in stress generation: An examination of
sociotropy/autonomy, stress, and depressive symptoms. Personality and
Social Psychology Bulletin, 32, 434. doi:10.1177/0146167205282739
Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis:
Modeling event change and event occurrence. New York, NY: Oxford
University Press.
Smucker, M. R., Craighead, W. E., Craighead, L. W., & Green, B. J.
(1986). Normative and reliability data for the Childrens Depression
Inventory. Journal of Abnormal Child Psychology, 14, 2539. doi:
10.1007/BF00917219
Southall, D., & Roberts, J. E. (2002). Attributional style and self-esteem in
vulnerability to adolescent depressive symptoms following life stress: A
14-week prospective study. Cognitive Therapy and Research, 26, 563
579. doi:10.1023/A:1020457508320
Spitzer, R. L., Williams, J. B. W., Gibbon, M., & First, M. B. (1990).
Users Guide for the Structured Clinical Interview for DSMIIIR.
Washington, DC: American Psychiatric Press.
Thompson, M., Kaslow, N. J., Weiss, B., & Nolen-Hoeksema, S. (1998).
Childrens Attributional Style Questionnaire Revised: Psychometric
examination. Psychological Assessment, 10, 166 170. doi:10.1037/
1040-3590.10.2.166
Tompson, M., Pierre, C., Boger, K., McKowen, J., Chan, P., & Freed, R.
(2009). Maternal depression, maternal expressed emotion, and youth
psychopathology. Journal of Abnormal Child Psychology, 38, 105117.
doi:10.1007/s10802-009-9349-6
Turner, J. E., & Cole, D. A. (1994). Development differences in cognitive
diatheses for child depression. Journal of Abnormal Child Psychology,
22, 1532. doi:10.1007/BF02169254
Williamson, D. E., Birmaher, B., Frank, E., Anderson, B. P., Matty, M. K.,
& Kupfer, D. J. (1998). Nature of life events and difficulties in depressed
adolescents. Journal of the American Academy of Child & Adolescent
Psychiatry, 37, 1049 1057. doi:10.1097/00004583-199810000-00015
Young, J. F., Mufson, L., & Davies, M. (2006). Efficacy of interpersonal
psychotherapy-adolescent skills training: An indicated preventive intervention for depression. Journal of Child Psychology and Psychiatry, 47,
1254 1262.
Zuroff, D. C., Mongrain, M., & Santor, D. A. (2004). Conceptualizing and
measuring personality vulnerability to depression: Comment on Coyne
and Whiffen (1995). Psychological Bulletin, 130, 489 511.
Received April 20, 2010

Revision received June 27, 2011
Accepted July 22, 2011
All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.

Predictors of The First Onset of A Major Depressive Episode and Changes in Depressive Symptoms Adolescence: Stress and Negative Cognitions

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Predictors of The First Onset of A Major Depressive Episode and Changes in Depressive Symptoms Adolescence: Stress and Negative Cognitions

Uploaded by

Copyright:

Available Formats

Journal of Abnormal Psychology

2011, Vol. 120, No. 4, 779 796

2011 American Psychological Association

Predictors of the First Onset of a Major Depressive Episode

self-schema (e.g., I am worthless, I cant do anything right, I

Cognitive theories of depression propose that individuals with

This article was published Online First September 19, 2011.

CARTER AND GARBER

Predicting Onset of a Major Depressive Episode and

PREDICTING FIRST ONSET OF MAJOR DEPRESSION

Risk: Maternal Depression

depression in youth is having a parent with a mood disorder

CARTER AND GARBER

Stemmler, 2002). Given the increasing rate of depression during

The Current Study

further moderate the relation between stress and cognitions to

PREDICTING FIRST ONSET OF MAJOR DEPRESSION

ioral, and somatic symptoms. For each of the 21 items, respondents

head, Craighead, & Green, 1986). Internal consistency of the CDI

An abbreviated data collection was conducted when participants were

CARTER AND GARBER

suring the degree to which children are satisfied with themselves

Predicting Onset of a Major Depressive Episode:

Results available from authors upon request.

PREDICTING FIRST ONSET OF MAJOR DEPRESSION

Negative cognitions Cognitive composite.

Depressive symptoms measured with the Childrens Depression Inventory.

episode at study entry. Given the large number of waves in the

Note. W1 Wave 1; W2 Wave 2; W3 Wave 3; W4 Wave 4; W5 Wave 5.

p .05. p .01. p .001.

.56 .02 .02 .42

PREDICTING FIRST ONSET OF MAJOR DEPRESSION

Predicting Change in Depressive Symptoms Over

CARTER AND GARBER

Main Effects: Achievement Stress

Note. Haz Hazard Function; Neg Cog Negative Cognitions.

p .10. p .05. p .01. p .001.

from the persons mean level of that variable; positive scores

High Negative Cognitions

Low Negative Cognitions

Figure 1. The interaction between interpersonal stressors and negative

determine whether the slopes were significantly different from

PREDICTING FIRST ONSET OF MAJOR DEPRESSION

Note. CDI Childrens Depression Inventory.

High Neg Cognitions

Low Neg Cognitions

Finally, we conducted a supplemental analysis to explore

Figure 2. The interaction between interpersonal stressors and negative

Analyses of the individual cognitive measures (i.e., self-worth or

CARTER AND GARBER

however, the odds of an onset of a depressive episode were not

The findings that negative cognitions significantly interacted

PREDICTING FIRST ONSET OF MAJOR DEPRESSION

Finally, in this study we categorized life events as interpersonal

attributional style, higher levels of interpersonal stressors were

Risk: Maternal Depression

Results of this study were consistent with the extant literature on

CARTER AND GARBER

ences in negative cognitions (e.g., Hankin et al., 2001; Thompson et

dropped out had higher levels of achievement stress at Wave one

PREDICTING FIRST ONSET OF MAJOR DEPRESSION

itants, and consequences. Psychological Bulletin, 104, 97126. doi:

CARTER AND GARBER