WISC-III predictors of academic achievement for children with learning disabilities: Are global and factor scores comparable?

The Wechsler Intelligence Scale for Children-Third Edition (WISC-III) provides a Full Scale Intelligence Quotient (FSIQ), four factor standard scores, and subtest scale scores. Although the incremental validity of the WISC-III factor scores over FSIQ in predicting academic achievement has been questioned, this finding was based on the assumption that FSIQ should be entered first into the regression equation and failed to examine the shared variance among predictors. In contrast with previous findings, this study of 174 children meeting criteria for learning disabilities revealed that the WISC-III factors accounted for a large portion of achievement variance during hierarchical regression analyses, yet FSIQ added little predictive power. A commonality analysis of FSIQ indicated that it is largely comprised of unique, not shared, factor variance. Analyzing the WISC-III subtests from a Gf-Gc theoretical framework, academic achievement commonality analyses revealed complex relationships among the predictors, with crystallized, quantitative, and short-term memory factors accounting for the most achievement variance, regardless of academic domain. Results suggest that simple rejection of factor or subtest scores based on hierarchical regression techniques is unwarranted and that systematic exploration of nomothetic and idiographic patterns of performance is recommended for practitioners.