Establishing Covariance Continuity between the WISC--R and the WISC-III.

Much of the research based on Wechsler Intelligence Scale for Children—Revised (WISC—R; D. Wechsler, 1974) subtest covariances may not validly be applied to the Wechsler Intelligence Scale for Children-Third Edition (WISC-III; D. Wechsler, 1991) without ascertaining whether the covariance structures of the normative samples for the two versions are equivalent. The covariance matrices for the WISC—R and the WISC-III were tested for equality by age group using the EQS Structural Equations Program (P. M. Bentler, 1989). Chi-square tests revealed significant differences between the covariance matrices for 10- and 16-year-olds; however, the more stable normed fit index revealed essentially no difference across all age groups. It can be concluded that, for all practical purposes, covariation among WISC-III subtests does not differ from covariation among WISC—R subtests.