Factor Structure of the WISC-R.

Abstract

A simultaneous maximum likelihood confirmatory factor analysis was performed on the 12 subtests the WISC-R in the 11 age groups of normative sample. Several models were fit to the data and compared, including: a single-factor model; orthogonal and oblique two-factor models, comprised of the Virbal and Performance subtests, respectively; and orthogonal and two oblique three-factor models, the third factor involving the Digit Span, Arthmetic, and Coding subtests. These analyses were subsequentlyb replicated in 11 samples in reported in 9 separate published articles. Results of the analyses in the normative sample suggest: 1) the single-factor model fits the data fairly well, although there is clearly some misspecification in this model; 2) the oblique multi-factor models do fit the data better than the single-factor model, though these models are also misspecified; and, 3) the single-factor, oblique two-factor and oblique three-factor structures are fairly consistent across the age groups. Results of the re-analyses of the published studies are quite similar to the results found in the normative sample, regardless of the population studied. It is concluded: 1) that the WISC-R largly consists of a general intellectual factor; and 2) all the critical models proposed in the literature to explain WISC-R performance are only partially correct.