Ethnic ability patterns on the Wechsler Intelligence Scale for Children-Revised: Relationship to theories of intelligence and cognition

This study examined the profiles of abilities for different ethnic groups on the WISC-R. The derived profiles were related to theories of intelligence and cognition. A total sample of 6,869 nonreferred children were included in this study. Data sets were obtained from sources throughout the United States and Japan. The ethnic groups were White, Hispanic, Black, American Indian, and Japanese. Ethnic groups were equated for Full Scale IQ utilizing analysis of covariance. Within ethnic group differences were examined and each group was divided to enable cross-validation of procedures. Discriminant analyses (i.e., classification, canonical and stepwise) yielded results indicating that the different ethnic groups could be discriminated (with the exception of the Hispanic group) with hit rates better than chance. Correct classifications ranged from 9-17% (Hispanic) to 68% (Japanese). Approximately 37-43% of the variance among the ethnic groups could be accounted for by the discriminant procedures. The subtests that consistently accounted for the largest amount of variance were the Vocabulary, Block Design, and Coding subtests. Particular ethnic profiles showed a Vocabulary $\u3c$ Block Design/Coding pattern. Cluster analyses revealed that the natural groupings of cases were not homogeneous with respect to ethnicity. The theories which accounted for the differences in the WISC-R ethnic profiles were those which separated spatial abilities from verbal abilities. These included: Wechsler\u27s verbal-performance distinction, Cattell\u27s crystallized and fluid intelligences, Kaufman\u27s hemispheric processing, and Vernon\u27s verbal-educative and spatial-mechanical model. The ethnic profiles which fit these theoretical perspectives were the Hispanic, Japanese, and one of the American Indian groups. Limitations of the study were addressed and caution in interpretation of the profiles was emphasized given that some ethnic groups were sampled regionally and the amount of variance accounted for by ethnicity was moderate. This study suggests future research be conducted incorporating socioeconomic status and gender variables. In addition, examination of a possible American profile in comparison to international profiles is noted