A Longitudinal Study of Children's Depressive Symptoms, Self Perceptions, and Cognitive Distortions about the Self

Depressive Symptoms 1

Journal of Abnormal Psychology, in press
Copyright 2001, American Psychological Association








A Longitudinal Study of Children's Depressive Symptoms,


Self-Perceptions, and Cognitive Distortions about the Self




Emily P. McGrath and Rena L. Repetti



University of California, Los Angeles






Author Note. The present study is part of the doctoral dissertation of the first author. We are
very grateful to the children, teachers, and parents who participated in the study. The study was
supported by a grant from the National Institute of Mental Health (R29-48593) awarded to Rena
Repetti. We are thankful for comments from Joan Asarnow, Connie Hammen, Marian Sigman
and John Weisz on an earlier version of this manuscript. Correspondence concerning this article
should be addressed to Emily McGrath, Ph.D., 300 Medical Plaza, Suite 1524, Los Angeles, CA
90095-1759. Telephone: (310) 794-1835. Electronic mail: emcgrath@ucla.edu.

---

Depressive Symptoms 2


Abstract

The purpose of this longitudinal study was to examine how depressive symptoms relate to
children's self-perceptions and estimates of children's cognitive distortions about the self in a
non-clinical sample of children who were followed from fourth- grade (n=248) through sixth-
grade (n=227). Report card grades in reading and math were obtained to measure children's
academic competence, and teachers' ratings of children's level of peer acceptance at school
served as the indicator of social acceptance. The longitudinal data suggested that depressive
symptomatology may have a negative impact on a child's ability to develop a healthy self-
concept. Self-reported depressive symptoms predicted a change in children's negative views of
the self. Moreover, the self- perceptions of children who exhibited more symptoms of
depression appeared to reflect an underestimation of their actual competence, as represented by
the more objective indicators of performance. Children's negative self-perceptions and
underestimations about the self were not associated with a subsequent change in depressive
symptoms. The implications of the findings for cognitive theories of depression and future
research with this population are discussed.

---

Depressive Symptoms 3
A Longitudinal Study of Children's Depressive Symptoms,
Self-Perceptions, and Cognitive Distortions about the Self

Cognitive theory suggests that cognitions, particularly negative beliefs about the self, are
related to the etiology of depression (Beck, 1967, 1976). Two of the most basic assumptions of
the theory, that (a) cognition has causal priority over emotions, and that (b) depressed children's
negative beliefs about the self reflect distortions of reality, have rarely been tested in the
childhood depression literature (for reviews, see Garber, Quiggle, & Shanley, 1990; Hammen,
1990). The current study used a prospective longitudinal design to test associations between
depressive symptoms and two dimensions of children's self-perceptions: the overall evaluation of
the self and the accuracy of that judgment. First, the study tested whether negative self-
perceptions are associated with subsequent increases in signs of depressive symptoms or whether
the reverse is true, i.e. depressed mood predicts to a more negative view of the self. Second,
estimating negative biases in children's judgments about the self, the study examined whether
underestimations of competence are a risk factor for future depressive symptomatology. The
reverse causal model was also tested.
Negative Self-Perceptions

Both children with non-clinical levels of depressive symptoms, as well as clinically
depressed children, view themselves negatively. Evidence suggests that depressed children are
unhappy with themselves and hold negative expectations about the future (e.g. Asarnow & Bates,
1988; McCauley, Mitchell, Burke & Moss, 1988), do not believe they can solve their problems
(Weisz, Sweeney, Proffitt & Carr, 1993), and are critical of their academic and social
competence (see Hammen, 1990; Hammen & Rudolph, 1996 for reviews). However, there is
disagreement in the childhood depression literature as to whether negative self-perceptions may
lead to depression or whether depression causes a negative view of the self (Garber, Quiggle, &
Shanley, 1990). Three longitudinal studies of middle childhood and early adolescence

---

Depressive Symptoms 4
(Robinson, Garber, & Hilsman, 1995; DuBois, Felner, Bartels, & Silverman, 1995; Hammen,
1988) found that lower self-esteem predicted a change in depressive symptoms over 6-12
months. However, none of these studies tested the reverse causal model to determine whether
depressive symptoms predicted a change in self-esteem over time. Findings such as these (based
on both clinical and non-clinical samples of depressed children) indicate that reports of low self-
esteem may temporally precede changes in depressive symptoms and are consistent with the
cognitive theory of depression. In an attempt to replicate these research findings, this study
tested the hypothesis that children's negative self-perceptions predict a change in depressive
symptomatology over time (Hypothesis 1). Three different components of children's self-
perceptions were examined: global self- worth, self-perceived academic competence, and self-
perceived social acceptance. Prior research with this age group has demonstrated the importance
of assessing domain-specific judgments of competence as well as overall perceptions of one's
value as a person (Harter, 1985).

While negative self-perceptions are believed to make one vulnerable to depression, it is
also possible that depressive symptoms influence one's negative self-perceptions (Teasdale,
1983). Proponents of this viewpoint have argued that negative self-perceptions are symptoms of,
rather than contributors to, depression. At least four studies have found that clinically depressed
children's and adolescent's self-perceptions improve as their depressive symptoms remit
(Asarnow & Bates, 1988; Gotlib, Lewinsohn, Seeley, Rohde & Redner, 1993; McCauley et al.,
1988; Tems, Stewart, Skinner, Hughes & Emslie, 1993). These findings suggest that a negative
self-perception may be a state-dependent symptom of depression rather a stable characteristic of
depressed children.

While negative self-perceptions may remit with depression, it remains unclear whether or
not they play any role in the onset, maintenance or exacerbation of depression. In addition,
whether or not the level of depressive symptoms typically observed in a non-clinical sample of

---

Depressive Symptoms 5
children would influence the children's self-perceptions over time has rarely been examined in
the literature. This study tested the hypothesis that depressive symptoms predict a change in
children's negative self-perceptions over time (Hypothesis 2). It is important to note that
Hypothesis 1 and Hypothesis 2 are not mutually exclusive. That is, negative self-perceptions
may make children more vulnerable to other symptoms of depression, and depressive symptoms,
in turn, may diminish feelings of self- worth.
Cognitive Distortions

Beck's (1967; 1976) cognitive theory of depression suggests that depressed children's
negative self-perceptions reflect cognitive distortions about the self. Beck's theory has prompted
research testing whether children with depressive symptomatology engage in distorted patterns
of thinking. Some of these studies have employed measures that ask children about hypothetical
situations to determine if their thinking reflects distorted processing of information (e.g.,
Children's Negative Cognitive Error Questionnaire (CNCEQ), Leitenberg, Yost, & Carroll-
Wilson, 1986; Cognitive Bias Questionnaire for Children (CBQC), Haley, Fine, Marriage,
Moretti, & Freeman, 1985). Others have used more objective indicators of children's
circumstances, in addition to subjective reports, in order to demonstrate that depressed children
distort information about themselves (e.g., Asarnow, Carlson, & Guthrie, 1987; McGee,
Anderson, Williams & Silva, 1986; Meyer, Dyck, & Petrinack, 1989; Kendall, Stark, & Adam
1990). For example, Asarnow et al. (1987) found that although depressed children saw
themselves as less academically competent, they did not differ from nondepressed children in
terms of IQ or actual achievement. These studies support the cognitive model in showing that
depressed children's negative self-appraisals may, at least in part, reflect cognitive distortions. In
particular, the evidence suggests that children with depressive symptomatology negatively distort
information about their academic competence.

Longitudinal studies are needed in order to determine whether cognitive distortions

---

Depressive Symptoms 6
precede or follow the appearance of depressive symptoms. Cognitive theories maintain that
cognitive distortions place one at risk for depression and not the reverse (Beck, 1967). However,
three recent analyses of data from a non-clinical sample of children addressed this question and
reported conflicting results. Cole and his colleagues (Cole, Martin, Peeke, Seroczynski, &
Hoffman, 1998; Cole, Martin, Peeke, Seroczynski, & Fier, 1999) examined third through eighth
graders' perceptions of their competence in several areas in relation to ratings provided by others,
such as teachers, peers, and parents. There was very little support for the hypothesis that
cognitive errors of underestimation are associated with a change in depressive symptoms (for the
single exception, see Hoffman, Cole, Martin, Tram, & Seroczynski, in press). However, in all
but one of the analyses reported by Cole and his colleagues, depression scores at the beginning
of the school year predicted children’s underestimation of their own academic and social
competence at the end of the school year.

The present study also tested the theoretical position that cognitive distortions play an
etiologic role in depression among children. The accuracy of children's self-judgments were
assessed in two life domains that are critical for this age group: academic performance and social
acceptance at school. Hypothesis 3 stated that children's underestimation of their actual social
and/or academic competence predicts a change in depressive symptomatology over time.
Cognitive distortions were measured by creating variables that reflect the accuracy of children's
self-perceptions, relative to independent indicators of children's competence. Report card grades
and teacher-reported peer problems served as more objective indicators of children's academic
and social performance. Of course, teacher evaluations are not purely objective nor completely
accurate indicators of children's academic and social performance. Other factors, such as
additional information children have about their academic and social performance (e.g., feedback
from parents; friendships outside of the classroom), may also contribute to children’s self-
evaluations (McGrath & Repetti, 2000; Repetti, McGrath, & Ishikawa, 1999). Hence, we

---

Depressive Symptoms 7
attempt only to estimate children's cognitive distortions about the self, with the understanding
that discrepancies between self- and teacher-ratings do not always necessarily reflect distortions.

Whereas cognitive theories suggest that negative cognitive errors make one vulnerable to
depression, some of the findings reported by Cole and his colleagues (Cole et al., 1998; Cole et
al., 1999; Hoffman et al., in press) suggest the reverse. That is, in a non-clinical sample of
children, depressive symptoms predicted more negative perceptual biases over time. This study
also tested the hypothesis that symptoms of depression predict a change in children’s
underestimation of their academic and/or social competence over time (Hypothesis 4). A
replication of this pattern in a different sample, utilizing different measures of children's
cognitive distortions about the self, would question a basic assumption of cognitive theories of
depression as applied to non-clinical samples of depressed youth.

The current study differs from most previous work in several important ways. First,
official report-card grades are used as the more objective indicator of academic performance, as
opposed to questionnaire data collected from teachers within the context of a research study. As
noted by Cole and his colleagues (Cole et al., 1999), in order for a negative self-perception to be
considered erroneous, children must receive and misinterpret positive feedback related to their
competence. By using report card grades, the current approach allowed for a more direct
assessment of the extent to which a child distorts information contained in feedback that was
actually received about his or her academic performance. Second, the present study uses
multiple reporters of children’s symptoms of depression in order to gain a more comprehensive
picture of children’s emotional functioning. Depressive symptoms are measured by (a)
children’s self- reports and (b) the average of mothers’ and teachers’ reports of depressive
symptomatology (Hoffman et al., in press, also obtained parents’ reports). Third, this study
extends the current literature by controlling for children's externalizing symptoms in tests of all
the hypotheses in order to assess the unique association between depression and self-evaluations.

---

Depressive Symptoms 8
Externalizing symptoms have been linked to children's depressive symptoms (Cole &
Carpentieri, 1990), negative self-perceptions (Compas, Phares, Banez, & Howell, 1991),
academic underachievement (Hinshaw, 1992) and problematic peer relationships (Parker, Rubin,
Price, & DeRosier, 1995). Findings such as these raise important questions about the specificity
of risk factors associated with depressive symptoms (Compas & Hammen, 1994; Hammen
1990). It is therefore important to rule out the possibility that externalizing behavior problems
account for some or all of the observed links between depressive symptoms, children's self-
perceptions, and cognitive distortions about the self.
Method
Procedures

The data for this study were collected as part of a larger longitudinal investigation of
stress and family development, involving annual collections of interview data from elementary-
school age children and questionnaire data from their parents and teachers over a period of three
years (fourth- through sixth- grade). Parents of fourth- grade children from three schools, one
parochial and two public schools in a large metropolitan area, were sent letters describing the
study. Children who agreed to participate, and whose parents returned consent forms, completed
annual interviews and their parents and teachers were asked to complete questionnaires. In
exchange for their participation each year, children received $5.00 to $10.00 and parents
received $10.00 to $20.00, with the honorarium increasing over the 3 years of the family’s
participation. Teachers received $5.00 for each completed child questionnaire.
Participants

Cohorts of fourth-grade children were recruited into the study in each of 3 consecutive
years. A total of 677 families with fourth graders were invited to participate, and parental
consent was obtained from 248 (37%) of the families. Time 1 interviews were completed with
248 children (116 girls and 132 boys; children's average age was 9.5 years) and 218 mothers

---

Depressive Symptoms 9
returned questionnaires, reflecting an overall maternal response rate of 88%. The study
maintained high retention rates over 3 years (230 child interviews and 195 parent questionnaires
when the children were in fifth- grade, at Time 2; 227 child interviews and 186 parent
questionnaires when the children were in sixth-grade, at Time 3). Teacher questionnaires were
completed for approximately 86% of the children at each time point. In order to analyze the
effects of attrition, children who were retained from Time 1 to Time 3 were compared to children
who withdrew from the study after Time 1 participation. T-tests indicated there were no
significant differences between the two groups of children on Time 1 scores on depressive
symptoms, externalizing symptoms, self-perceptions, or cognitive distortions. The sample
consisted primarily of high income, highly educated Caucasian parents and their children.
Approximately 81% of the parents who participated identified themselves as Caucasian (8% as
Asian/Pacific Islander, 4% as Latino, 1% as African American, 1% as Native American, and 5%
as Other). More than half of the families (54%) reported earning more than $80,000 per year,
and over 80% of the parents were college graduates.
Measures
Emotional/Behavioral Functioning

Depressive Symptoms: Children's Self- Reports. The first indicator of children's
depressive symptoms was children's self- reports on the Children's Depression Inventory (CDI)
(Kovacs, 1992). The CDI is a 27- item questionnaire that assesses affective, behavioral, somatic,
and cognitive aspects of depressive symptomatology. Evidence suggests that this widely used
scale has good internal consistency (Cronbach's alpha = .85 in the current study), test-retest
reliability, concurrent validity, and validity as a screening measure for depressive
symptomatology in non-clinic samples (Kovacs, 1992; Carey, Faulstich, Gresham, Ruggiero, &
Enyart, 1987).

For each item on the CDI, children were asked to indicate which of three sentences best

---

Depressive Symptoms 10
described how they had been feeling during the last two weeks. An example is "I am sad once in
awhile; I am sad many times; I am sad all the time." Each sentence was assigned a value from 0-
2, with higher scores reflective greater symptom severity. Responses to the 27 items were
summed. Children's scores (see Table 1) were lower but fell within one standard deviation of the
national norms (Kovacs, 1992). Approximately 80% of children scored below 9 at each time
point, a cutoff score used to signify mild depressive symptomatology (e.g., Rudolph, Hammen,
and Burge, 1997). Children's scores ranged from 0-38 at Time 1, 0-29 at Time 2, and 0-35 at
Time 3. T-tests indicated there were no differences between girls' and boys' self-reports at any of
the three time points. There were high rates of stability in the depression variable over three
years (r =.55 to r =.67, p< .001).

Depressive Symptoms: Parents' Reports. Mothers' reports on the 14- item
anxious/depressed syndrome scale of the Child Behavior Checklist (CBCL) served as the second
indicator of children's depressive symptoms. The CBCL consists of 118 items that assess
children's internalizing and externalizing behavior problems. This widely- used measure has
been shown to be reliable (Cronbach's alpha for the anxious/depressed syndrome scale = .81 in
the current study), stable, and valid (Achenbach, 1978; Achenbach & Edelbrock, 1979).
Utilizing principal components analyses, Achenbach found that depressive and anxious
symptoms on the CBCL are closely intertwined and difficult to tease apart (Achenbach, 1991a),
which is consistent with research indicating high rates of comorbidity between depression and
anxiety in children (Compas & Hammen, 1994). A sample item from the anxious/depressed
syndrome scale is "Unhappy, sad, or depressed." Mothers were asked to circle (0) if the item was
not true of the child, (1) if the item was sometimes true of the child, and (2) if the item was often
true of the child during the previous 6 months. On average, mothers reported low levels of
depressive symptoms in their children (see Table 1). These ratings were comparable to norms on
the anxious/depressed subscale from a national sample of non-referred 4-11 year-old children. T-

---