Abstract
The construct validity of four test batteries (i.e., the Wechsler Intelligence Scale for Children-Revised; Cognitive Levels Test; Peabody Individual Achievement Test; and Academic Levels Test) was investigated for a group of adjudicated delinquent youth. The degree of convergent and discriminant validity indicated by the intercorrelations among the 18 scores obtained was evaluated. Results revealed that the constructs of cognitive ability and academic achievement were supported insofar as the correlations were generally higher than those typically found in the literature. Also, the ranks of groups of correlations were ordered as hypothesized. Briefly, (a) reliabilities of individual tests were highest in magnitude; (b) correlations between scores matched on the basis of content (e.g., Verbal Reasoning and Verbal IQ) were second; (c) correlations between measures of the same broad construct (e.g., cognitive ability: Quantitative Reasoning and Verbal IQ) ranked third; and (c) correlations between measures of unmatched constructs (e.g., cognitive ability [Performance IQ] and achievement [Spelling]) were lowest. Multiple comparisons to examine differences between mean scores revealed few significant differences. On the whole, the results were interpreted as providing strong support for the validity of the batteries. Departures from the generally positive conclusions were discussed.