Date of Award

5-1-1995

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Counseling Psychology & Community Services

Abstract

This study examined parental responses on the Child Behavior Checklist (CBCL), and considered variables that may affect rating reliability. Subjects for the study were 82 matched parent-pairs who completed CBCL protocols on their clinic-referred children. Research variables included parent gender, child gender, and level of item objectivity. The data was analyzed at the CBCL scale (narrowband) and broadband (Externalizing and Internalizing) levels, utilizing analysis of mean differences, repeated-measures analysis of variance, and clinical classification agreement.

Results showed little or no mean difference between mothers' and fathers' scores, but a large degree of discrepancy for parent-pairs was evident. None of the associated ANOVA F-tests were statistically significant. On analyses that investigated parent-gender by child-gender interaction effects, no significant gender differences were found between mother vs. father discrepancies.

Diagnostic classification analyses revealed a tendency for mothers to classify children in the clinical range more often than fathers, although differences were not statistically significant. Kappas and Occurrence Agreement Indices were low to moderate (k ranged from .30 to .55, and OAIs ranged from .28 to .53, respectively). Regarding item objectivity. Externalizing Scales (composed of items rated as more objective or observable), demonstrated greater parental agreement (higher kappas) and higher reliabilities than Internalizing Scales.

These results revealed no systematic differences between mother and father ratings, and therefore continued use of common norms is supported. However, the low to moderate reliabilities and indices of agreement suggest that the CBCL be used with caution when making diagnostic or treatment decisions.

Download

Share

COinS