Off-campus Pacific University users: To download campus access theses and dissertations, please log into our proxy server with your PUNet ID and password.

Non-Pacific University users: Please talk to your librarian about requesting this thesis or dissertation through interlibrary loan.

Theses or dissertations that have a specific embargo period indicated below will not be available to anyone until the date indicated.

Date of Award


Degree Type

Dissertation (On-Campus Access Only)

Degree Name

Doctor of Psychology (PsyD)


The Roberts Apperception Test for Children (RATC) was investigated for construct validity in terms of its ability to discriminate depressed from nondepressed school-aged children. The performances of relatively depressed and nondepressed children on the RATC and the Children's Depression Inventory (CDI), a well validated measure of childhood depression, were compared. A literature review of a number of replicated studies indicated that depressed children have negative attributional styles and interpersonal problem solving deficits that are characteristic of depressed adults. An examination of the RATC indicated that four basic scales: Problem Identification, Resolutlon-1, Resolution-2, and Unresolved have the promise of sampling specific components of problem solving deficits and attribution bias. This study tested the efficacy of these four scales relative to the RATC Depression scale, which is a thematic scale for depression. It was hypothesized that the four problem solving scales would perform as well or better than the thematic Depression scale 1n 1dentifying relatively depressed school children. Fourth and fifth graders from two schools, aged nine through eleven, were screened with the CDI in a group format. Subjects were assigned to one of three pools based on CDI scores; Nondepressed (scores of 0-5), Mixed (6-12), and Depressed (13-34). Three groups of 20 subjects each were then formed by randomly pulling students from these pools. The subjects In each group were administered a standard RATe by trained examiners who rema1ned bl1nd to group membership. Analyses of variance tested differences between the three subject groups for each of the five scales in question. The results indicated that none of these five scales were able to d1scriminate between groups to a sign1ficant degree, although a scale measur1ng a higher quality of problem resolution, Resolution-2, approached significance. An intercorrelation matrix was also calculated and correlations between the f1ve scales and the collapsed CD! distribution from all groups were tested for Significance. The RATC Depression scale outperformed the other four scales by correlating significantly with the CDI in the expected direction, as well as with several other RATC scales be11eved to be associated with depression in children, such as the Aggess10n and Anxiety scales. The four problem solving RATC scales generally showed very little promise as indices of ch1ldhood depression for this nonclinical group of school children. The results were discussed as evidence for the construct validity of the RATC Depression scale, and it was noted that the problem solving scales still requ1re val1dation against a validated scale of problem solving to fully Interpret their performance in a study such as this one. The results did not appear to either validate or invalidate the cognitive hypothesis regarding depression in children. The results of the study were also discussed in light of the psychometric issues inherent in developing objective scoring systems for thematic projective instruments. Specifically, the results seemed to underscore the difficulties involved in constructing objective scales for thematic tests that are both reliably scored and valid. The utility of configural scoring systems were addressed in light of the psychometric properties indicated by the descriptive statistics and other results found here. Implications for further studies of this type were discussed. It was suggested that clinical samples of depressed children, identified by multiple measures of depression., would add to internal validity. In addition, a larger subject population and regression analyses were discussed as being important design issues in future research regarding construct validation of the RATC scales.