Evaluation of subclinical depression in children using self-, peer-, and teacher-report measures

This study reports on several classification issues utilizing the Children's Depression Inventory (CDI), the Child Behavior Checklist-Teacher Form (CBCL-T), and the Peer Nomination Inventory of Depression (PNID). The first sample contained 752 public school children in grades 5, 6, and 9. A second sample of 142 fifth-graders was included to have an additional PNID comparison group. Results revealed moderate correlations between the CDI and the CBCL-T and PNID. Use of multiple criteria for selecting children as depressed was discussed, and selection rates using multiple measures with either strict or lenient cutoff scores were given. Normative data on the PNID were evaluated, and it was suggested that norms for the PNID may not be generalizable to other samples. Finally, race and gender analyses showed that CDI scores for females were significantly higher than for males, whereas CDI scores for Hispanics were significantly higher than for non-Hispanic whites. However, Omega-squared analyses for these two variables showed this to be of little practical significance.     

Self-Report,Peer-Report,and Teacher-Report Measures of Childhood Depression: An Analysis by Item

We explored the low correlation among different types of childhood depression measures at the item level. The items from the Children's Depression Inventory (CDI), Peer Nomination Inventory of Depression (PNID), and the Child Behavior Checklist-Teacher Report Form (CBCL-T) were combined, and both first- and second-order factor analyses were conducted. Results indicate that self-report, peer-report, and teacher-report assessments of depression measure generally uncorrelated constructs. Second-order analysis suggests that depression as a global construct is being measured to some degree by items from all three instruments. Canonical analysis was employed to identify items that best predicted CDI, PNID, and CBL-T summary scores simultaneously. Also, the relationship between specific items with similar content was investigated. Results from these analyses generally supported a conclusion that the three types of measures yield scores that are primarily independent and that the use of summary scores is not masking stronger relationships within measures. These findings have implications for clinical practice and construct elaboration.     

Self-report, peer-report, and teacher-report measures of childhood depression: an analysis by item.

We explored the low correlation among different types of childhood depression measures at the item level. The items from the Children's Depression Inventory (CDI), Peer Nomination Inventory of Depression (PNID), and the Child Behavior Checklist-Teacher Report Form (CBCL-T) were combined, and both first- and second-order factor analyses were conducted. Results indicate that self-report, peer-report, and teacher-report assessments of depression measure generally uncorrelated constructs. Second-order analysis suggests that depression as a global construct is being measured to some degree by items from all three instruments. Canonical analysis was employed to identify items that best predicted CDI, PNID, and CBL-T summary scores simultaneously. Also, the relationship between specific items with similar content was investigated. Results from these analyses generally supported a conclusion that the three types of measures yield scores that are primarily independent and that the use of summary scores is not masking stronger relationships within measures. These findings have implications for clinical practice and construct elaboration.     

Teachers'' ratings of social competence of children with high versus low levels of depressive symptoms

This study examined the relationship between children's self-ratings of depressive symptoms on the Children's Depression Inventory (CDI) and teachers' ratings of the situational social competence of these children based on the Taxonomy of Problematic Social Situations (TOPS). Children in the high CDI group showed significantly higher teacher ratings (indicating lower social competence) than children in the low CDI group on four TOPS scales: peer group entry, response to provocation, response to success, and teacher expectations. Of the children in the high CDI group, 85% could be correctly classified with a discriminant analysis on the basis of TOPS ratings; of the children in the low CDI group, 77% were correctly classified. Implications of these findings, as well as limitations of the study, are discussed.     

A psychometric assessment of a two-factor solution for the preschool behavior questionnaire in mid-childhood

In this study, we examined Fowler and Park's (1979) two-factor solution (aggressive-hyperactive-distractible and anxious-fearful) for the Preschool Behavor Questionnaire (PBQ; Behar & Stringfield, 1974) in a sample of 179 (93 males, 86 females) grade 1 and 155 (73 males, 82 females) grade 2 children. The validity of this two-component solution was assessed by relating the PBQ factor to indices of peer popularity and observations and peer nominations of aggressive, sociable, and isolative/fearful behavior. Popularity was measured using a rating scale sociometric. Behavioral data were obtained by observing children at free play and recording the frequency of aggression, withdrawal, and sociability. Peer assessments of children's social behaviors were obtained using The Revised Class Play (Masten, Morison, & Pellegrini, 1985). A two-factor solution of the PBQ resulted in the identification of an externalizing (EXT) and an internalizing (INT) factor. These data replicated earlier findings employing younger children (e.g., Fowler & Park, 1979). Each child's EXT and INT score was then correlated (with age partialled out) with his/her scores on each of the observational and peer assessment variables. Teacher ratings of EXT were significantly related to indices of aggression and unpopularity. The teacher ratings of INT were related to measures of anxiety, withdrawal, and unpopularity. These results suggest that the two-factor solution employed herein is an economical and viable indicator of social maladjustment in samples of “normal” elementary school-aged children.     

The differential validity of teacher ratings of Inattention/Overactivity and Aggression

The ability of school-based measures of child behavior to predict differentially the IOWA Conners teacher rating scale Inattention/Overactivity (IO) and Aggression (A) factors was evaluated in a sample of 71 school-aged boys. Multiple measures and multiple methods of assessment were utilized to provide a comprehensive assessment of social and academic behavior. These included direct observations of children in the classroom and on the playground, examination of the organization of children's desks, measures of academic performance, peer nominations of popularity and rejection, and sociometric ratings using the Pupil Evaluation Inventory. Despite moderately high correlations between the teacher rating factors (r = .60), considerable evidence was provided for differential validity on measures of academic performance, peer rating measures, and measures of disruptive or inappropriate classroom and playground behavior. These differences between IO and A factors indicated that the factors were logically independent in important ways, supporting prior work validating separate hyperactivity and aggression dimensions.     

The relationship of young adolescent Children's Depression Inventory (CDI) scores to their social and cognitive functioning

The relationship of young adolescents' scores on the Children's Depression Inventory (CDI) to subjective and objective indices of their social and cognitive functioning was examined. Subjects were 89 young adolescents (11–15 years old), their parents, and their social studies teachers. The correlations of adolescents' CDI scores with recent school grades, objective ratings of overt social behavior, and measures of adolescent social and cognitive competence as perceived by the adolescent, both parents, and the teacher were examined. CDI scores were significantly and negatively correlated with school grades, behavioral ratings of positive social communication, and adolescent, mother, father, and teacher perceptions of the adolescents' social and cognitive competence. The relationship of these findings to those of other CDI validity studies is discussed.This research was supported, in part, by the William T. Grant Foundation and the University of Georgia's Institute for Behavioral Research.     

An Examination of the Association between Social relationships and depression in early adolescence

The purpose of this study was to examine the association between social relationships and depression in early adolescence. Subjects were 216 sixth and seventh graders. Depression was measured by the Children's Depression Inventory; social relationships were measured by three sociometric techniques (positive nominations, negative nominations, peer ratings) and estimates by the subject of the number of nominations and the rating (s)he received. The results indicated that depression was associated with pear and self-estimate negative nominations and ratings more than with peer and self-estimate positive nominations. A stepwise multiple regression predicting depression scores included two of the six predictors and accounted for 14% of the variance. The implications of the results for the assessment and treatment of depression are discussed.     

Peer acceptance: The correspondence between children's sociometric scores and teachers' ratings of peer interactions

In view of the current interest in children's peer relationships and social skills, and the need for valid assessment procedures for children's peer problems, the intent of the present study was to examine the correspondence between peer ratings of acceptance and teacher ratings of a child's social behavior and likability. The 92 children were males and females from the third, fourth, and fifth grades. Classroom teachers rated each of the children on withdrawn, aggressive, and likable behavior using the Pupil Evaluation Inventory. Peer ratings of the child's acceptance in play and work situations were obtained from same-sex classmates. The teacher rating of likability was the best predictor of peer acceptance scores for males; withdrawn behavior was the best predictor of peer acceptance scores for females. Ratings of withdrawn and aggressive behavior contributed to the prediction of peer acceptance scores for males; for females, only withdrawn behavior contributed to the prediction of peer acceptance. In terms of the clinical utility of teacher ratings for assessing children with peer problems, the use of the withdrawn and/or aggressive ratings scales appeared to offer promise as a means of identifying such children. Implications for the area of children's social skills and directions for future research were discussed.     

Peer relationships in ADDH and normal children: A developmental analysis of peer and teacher ratings

Peer and teacher perceptions of younger and older ADDH and normal children were compared. Peer nominations on the Pupil Evaluation Inventory revealed that ADDH children received more nominations on the Aggression factor and fewer on the Likability factor. ADDH boys received more nominations on the Withdrawal factor. Generally, older children received fewer nominations, but age did not interact with diagnosis; that is, younger and older ADDH children were perceived by peers as equally deviant. Discriminant analyses showed that peer ratings were useful in discriminating between ADDH and normal boys, but not between ADDH and normal girls. Teacher ratings of peer behavior suggested that older ADDH girls were perceived as less disturbed than younger girls. However, teacher ratings of boys revealed no age by diagnosis interactions. Low to moderate correlations were found between peer and teacher ratings of ADDH boys, and the pattern of relationship varied with age.During the writing of this report, Charlotte Johnston was supported by a Sir James Lougheed Fellowship, Alberta Heritage Scholarship Fund.     

Parent-completed and adolescent-completed CDIs: Relationship to adolescent social and cognitive functioning

The purpose of this study was to examine the relationship of both young adolescents' scores on the Children's Depression Inventory (CDI) and their mothers' ratings of the adolescents' depression (parent-completed CDI) to indices of their social and cognitive functioning obtained from a source outside the home. Subjects consisted of a nonclinic sample of 85 young adolescents (11–15 years of age), their mothers, and their social studies teachers. Recent school grades and teacher-completed measures served as the indices of adolescent social and cognitive competence. The results indicated that both adolescents' and mothers' CDI scores were significantly and negatively correlated with measures of adolescent cognitive and social functioning. Multiple regression analyses, utilizing adolescent- and parent-completed CDI scores as predictors, indicated that both predictor variables entered into the equations for cognitive functioning while only the mother-completed CDI entered into those for social functioning. When maternal depression was also entered into the multiple regressions, the findings regarding mother- and adolescent-completed CDI scores were not altered. The relationship of these findings to other somewhat disparate findings concerning the utility of mother and child reports of child depression is discussed.This research was supported, in part, by the William T. Grant Foundation and the University of Georgia's Institute for Behavioral Research.     

Moderators of the Relation Between Popularity and Depressive Symptoms in Children: Processing Strength and Friendship Value

Children with low (n = 25) and high (n = 38) peer-rated popularity completed an emotional Stroop task, using negative social words, a self-report measure of friendship value relative to other domains of competence, and the Child Depression Inventory (CDI). Six months later, they completed the CDI again. In regression analyses, after controlling for prior CDI scores, social status interacted significantly with both Stroop and value measures (separate regressions). For unpopular children, both greater friendship valuing and greater negative social word Stroop interference predicted increases in depressive symptoms. In contrast, neither predictor was significant for popular children. In a third regression that included friendship value and Stroop interference as joint predictors of depression change, their effects remained significant and independent. We discuss these findings' implications for 3 models of depression; Champion and Power's social-cognitive theory of depression (L. A. Champion and M. J. Power, 1995), Pyszczynski and Greenberg's self-regulatory perseveration theory of depression (T. Pyszczynski & J. Greenberg, 1992), and Harter's model of global self-worth (S. Harter, 1985).     

Peer ratings of depression,anxiety, and aggression in inpatient and elementary school children: Rating biases and influence of rater's self-reported depression,anxiety, and aggression

Using parallel self-, peer, and teacher rating scales, several rating biases in children's peer ratings of depression, anxiety, and aggression were examined. Participants were 66 inpatient and 133 elementary school children (N = 199, 109 boys, 90 girls; 61% white, 39% black) aged 8 to 12, and their teachers. Results showed significant halo bias in both the children's peer ratings and the teachers' ratings. Children's self-reports on each of the three traits were significantly related to their peer ratings of the same trait, while adjusting for socioeconomic status and the peers' teachers' ratings of the same trait. Children who rated themselves as high on each trait rated their peers significantly higher on the same trait than children who rated themselves as medium or low; and for depression and anxiety, those who rated themselves as medium rated their peers significantly higher on those traits than those who rated themselves as low. For both depression and aggression, children's self-reports on the trait were significantly related to their peer ratings of the same trait, but not significantly related to their peer ratings of different traits. Disagreements between children's and teachers' ratings of the peers on all three traits were significantly related to child self-reports on each trait, indicating a possible distortion in children's peer ratings due to self-report. The implications of the results for both peer and others' assessments are discussed, and further investigation of rating biases in other informants' assessments is encouraged.These data were collected as part of the author's doctoral dissertation submitted to Memphis State University. Appreciation is expressed to Stacey Donegan for assistance with the literature review for an earlier version of this paper presented at the meeting of the Society for Research in Child Development, New Orleans, March 1993.     

The relations of thematic and nonthematic childhood depression measures to defensiveness and gender

The susceptibility of self-report and thematic measures of depression to the effects of a defensive test-taking approach was assessed among 44 child and adolescent psychiatric inpatients. It was hypothesized that scores on a measure of defensiveness would be associated with self-reported but not thematic measures of depression, and furthermore, that this finding would be more pronounced for girls than boys. Patients completed the Lie Scale of the Revised Children's Manifest Anxiety Scale (RCMAS), the Children's Depression Inventory (CDI), and the Roberts Apperception Test for Children. As predicted, RCMAS-Lie scores were significantly associated with CDI but not Roberts Depression scores, and this effect was more pronounced among females, primarily due to the high CDI scores of nondefensive girls. Implications of results for psychological assessment of children and adolescents are discussed.     

Norms and screening utility of the Dutch version of the Children's Depression Inventory in clinical and nonclinical youths

This study aimed to (a) assess relationships between the Children's Depression Inventory (CDI) and DSM-oriented depression and anxiety scales of the Youth Self Report, (b) develop reliable norms for the CDI, and (c) determine CDI cutoff scores for selecting youngsters at risk for depression and anxiety. A total of 3,073 nonclinical and 511 clinically referred children and adolescents from The Netherlands and Belgium were included. Results showed that CDI scores were significantly related to DSM-oriented symptoms of both depression and anxiety. CDI scores correlated highly with depression symptoms and moderately with anxiety symptoms. Norms for the CDI were determined by means of multiple regression analysis and depended on sex, age, and country. CDI cutoff scores for selecting individuals at risk for depression and anxiety as measured by the DSM-oriented depression and anxiety scales of the Youth Self Report were determined by means of multiple regression analysis and receiver operating characteristic analysis. A CDI score of 16 was found to have the most optimal balance between sensitivity and specificity for depression, whereas a score of 21 provided the best sensitivity and specificity for anxiety in a subsample of children. We conclude that the CDI is an effective instrument for screening depression and to a lesser extent anxiety in primary and secondary care centers, before applying further assessment of high-risk individuals.     

Early emotional instability,prosocial behaviour,and aggression: some methodological aspects

The internal consistency of Emotional Instability, Prosocial Behaviour, and Aggression scales was examined in elementary school children aged between 7 and 10 years. Multiple informants (self-report, teacher rating, mother rating, and peer nomination) were obtained on the above three dimensions. Relationships between the scales were first examined within the single informant frame of reference and then between informants. The concurrent validity was assessed using sociometric measures (popularity, rejection, social impact, and social preference) and the Achenbach and Edelbrock Child Behavior Checklist (teacher and parent form). The results confirm the internal validity of the three scales measuring emotional instability, prosocial behaviour, and aggression in child self-report, teacher rating, and mother rating. The use of self-report measures seems promising within the multiple informant strategy of research on child behaviour. Correlational results show satisfactory concurrent validity for the three scales, especially for teacher rating and peer nomination. The strict relation between emotional instability and aggression poses problems of discriminant validity which need to be further investi gated.     

The relationship of anxiety to childhood depression

In order to investigate the relationship between anxiety and depression in emotionally disturbed children, 30 hospitalized inpatient children were individually administered the Children's Depression Inventory (CDI), the Children's Manifest Anxiety Scale-Revised (CMAS-R), and the State-Trait Anxiety Inventory for Children (STAIC). Results indicated a significant relationship between CDI scores, the CMAS-R and its factors, and the STAIC. Correlations between the various factors of anxiety and depression suggest a complex relationship between the two constructs. Stepwise regression analyses indicated further the complexity of this relationship. Results were discussed in terms of the possible differential role which the different anxiety factors play in depression.     

The relationship of three measures of childhood depression to academic underachievement

Clinical features of depression (short attention span, lethargy, poor memory and shortened task persistence) appear to be incompatible with effective learning. This has led several investigators to suggest that depression is a significant cause of academic underachievement. The most adequate test of this hypothesis (Tesiny, Lefkowitz, & Gordon, 1980), however, found only small correlations between depression and academic achievement when neither IQ nor socioeconomic status were controlled. In the present study, three measures of childhood depression (Peer Nomination Inventory for Depression, Children's Depression Inventory, and an ad hoc teacher rating) were correlated with scores from the Peabody Individual Achievement Test and the Stanford Achievement Test. IQ was statistically controlled to provide a direct measure of underachievement and analyses were conducted separately for males, females, and the combined sample in a sample that was homogeneous for socioeconomic status. Although intercorrelations among the variables indicated that the sample was appropriate to test the hypothesis, only a few significant correlations were found. These findings suggest that if depression causes academic underachievement, it is a weak or uncommon effect.     

Heightened subjective experience of depression in borderline personality disorder

Despite the frequent comorbidity of major depression and borderline personality disorder (BPD), limited research has examined what effect this comorbidity has on the severity, course, and presentation of depression. The purpose of this study was to examine whether the severity of major depressive disorder (MDD) in the context of comorbid borderline personality disorder (BPD) differs from MDD when comorbid BPD is not present and to determine whether different measures of depression yield convergent findings. Sixty patients diagnosed with DSM-IV MDD participated in this study. Twenty-nine were diagnosed with DSM-IV BPD, while the remaining 31 had no Axis II diagnosis. Depression was evaluated with both clinician (Hamilton Rating Scale for Depression) and self-report (Beck Depression Inventory) ratings. While the two groups were rated as similarly depressed by clinicians on the overall rating and the factor scores, the MDD/BPD group reported more severe depressive symptoms on the self-report measure. This difference was significant even after controlling for clinician-rated severity. Gender interacted with diagnosis, males in the BPD group showed the largest discrepancies between clinician ratings and self-reports. Posthoc analyses of HDRS factors with the BDI showed that the clinicianrated cognitive disturbance and retardation factors were correlated with self-rated severity overall. Within subgroups, only the retardation factor was correlated with the BDI. Our results suggest that while depressed individuals with and without BPD may be rated as similarly depressed when assessed with objective rating methods, the subjective experience of the depression may be rated as more intense or severe by patients with comorbid BPD. The mechanism underlying this effect remains unknown, and requires further research.     

The assessment of depression in children: The internal structure of the child depression inventory (CDI)

In the present report three separate studies of childhood depression were conducted. First, the internal structure of the Child Depression Inventory (CDI), with 216 children representing various ethnic groups with equal numbers of boys and girls, was evaluated through a factor analysis and by various internal-reliability measures (e.g. split-half reliabilities, Pearson correlations of each item to the total score). Four factors were established and internal reliability of the scale proved to be high. The relationship of the factor structure of the CDI to Kendell's Type A and B categorization of depression are discussed. In Study 2 the relationship of demographic variables to the CDI using the same group of children described for Study 1 was employed. Evaluating the characteristics of depression across age, sex and so on in children has not been frequently studied, and was deemed appropriate for the present investigation. Age proved to be a significant factor in depression scores although race and gender did not. With respect to age, older children tended to display more symptomatology. Comparisons of depressed children to nondepressed children also showed that age was a factor in the obtained scores, and range of severity in both groups. Depressed children differed from nondepressed children on all 27 items indicating that all the items on the CDI seem to be measuring a unitary concept. Study 3 compared CDI scores to a measure of social behavior, the Matson Evaluation of Social Skills with Youngster (MESSY). Seventy-six children (36 girls and 40 boys), ages 4–10 yrs $\text{(}\text{X}\text{?}\text{= 7)}$ were evaluated. Appropriate Social Skills was negatively correlated with childhood depression, and Inappropriate Impulsive/Assertiveness was positively correlated with depressive features described under the factor Guilt/Irritability. The implications of these data for further research on assessment, differential diagnosis and evaluation of treatment research are discussed.