Ability patterns of black and white adults on the Wechsler Adult Intelligence Scale - Revised independent of general intelligence and as a function of socioeconomic status

Abstract

The national standardization sample from the Wechsler Adult Intelligence Scale - Revised was examined to determine the relative impact of the general factor of intelligence, group factors, subtest specificity, and socioeconomic status upon the profile of subtest scores that emerged from the scores of Blacks and of Whites in the standardization sample. Black-White ability profiles were developed for the full standardization sample (192 Blacks and 1664 Whites) as well as for those adults 34 years of age and younger (100 Blacks and 784 Whites) and 35 years of age and older (92 Blacks and 880 Whites). For the full sample, Block Design correlated positively and significantly with White classification while Digit Span and Picture Arrangement correlated positively and significantly with Black classification after the groups were statistically equated on general intelligence. The only significant partial correlations found for both the younger and older samples were with Block Design, and the direction of the relationship again favored Whites. The Black and the White subtest intercorrelation matrices were submitted to hierarchical factor analyses. Coefficients of congruence were calculated between corresponding factors and revealed substantial similarity between factor structures of the WAIS-R for Blacks and for Whites. Estimated factor scores were then determined for all Black and White individuals. Further analyses indicated that the general factor contained within all the WAIS-R subtests accounted for the majority of variance between groups when compared to group factors and subtest specificity. The profile of Black-White differences on the WAIS-R subtests did not correlate with SES profiles developed separately for each group, suggesting that the pattern of subtest differences is not explainable by differences in SES levels. Stepwise multiple regressions with estimated factor scores as dependent variables and Black-White classification and level of SES as independent variables revealed both demographic variables to be linked primarily to the general factor. However, partial correlations between factor scores and each demographic variable (with the other held constant) resulted in only one significant relationship. Only the partial correlation between group classification and general intelligence factor scores maintained statistical significance after the demographic variables were partialled from each other.