No abstract

Scatter analysis of IQ profiles has a long and controversial history. We conducted this study to determine whether the validity of scatter information is any greater for 2 new IQ batteries, the Kaufman Assessment Battery for Children (K-ABC) and the Fourth Edition Stanford-Binet (SB4), than for the Wechsler Intelligence Scale for Children-Revised (WISC-R). Within a sample of referred children, we computed numerical indexes of profile elevation, shape, and variability for all IQ tests. Using hierarchical multiple regression with achievement scores as dependent variables, we found shape information had marginal incremental validity over elevation as a predictor for the WISC-R, even less for the SB4, and virtually none for the K-ABC. We discuss implications of these results.

We evaluated whether abilities measured by the Kaufman Assessment Battery for Children (K-ABC)are as complex as those measured by the Wechsler Intelligence Scale for Children-Revised (WISC-R)within a sample of referred children. Results of factor and multidimensional scaling analyses indicated that abilities tapped by the K-ABC Mental Processing scale are not as complex as those assessed by the WISC-R. K-ABC Achievement scale subtests, however, seemed to assess skills as complex as those measured by many WISC-R subtests. The relative cognitive complexity of the K-ABC and the WISC-R seem comparable, but only when all the K-ABC subtests are administered.

R. W. Kamphaus and C. R. Reynolds developed a scoring system for the Kaufman Assessment Battery for Children (K-ABC) in which test users derive alternative summary scores, including General Intelligence, Verbal Intelligence, and Reading composites. In this study, we evaluated the construct validity of this supplementary scoring model with data from a sample of referred school-age children and from the K-ABC normative sample. Within the referred sample, we conducted a factor analysis to determine the joint factor structure of subtests from the K-ABC and other IQ and achievement measures. We evaluated within the normative sample the relative fits of the factor models implied by the original and Kamphaus-Reynolds scoring models to subtest inter-correlations. Our results supported the construct validity of only part of the Kamphaus-Reynolds scoring model.

No abstract