Investigating black-white differences in psychometric IQ: Multi-group confirmatory factor analyses of the WISC-R and K- ABC and a critique of the method of correlated vectors

analyses of the WISC-R and K-ABC, and a critique of the method of Abstract According to Spearman's hypothesis, observed black-white (b-w) differences in psychometric measures of cognitive abilities are attributable to a fundamental difference in general intelligence. Efforts to investigate this hypothesis have centered on the method of correlated vectors. This method involves calculating the correlation between the factor loadings of cognitive tests on the common factor identified as g and the (standardized) observed differences in means on the tests. This correlation is called a Spearman correlation. A large and positive Spearman correlation is supposed to be indicative of the importance of general intelligence as the source of b-w differences. Several commentators have argued that the issue of b-w differences in psychometric IQ test should be addressed by means of multi-group confirmatory factor analysis (MGCFA). The software to carry out such analyses is in place, as are the exact requirements within this model for a meaningful comparison of groups. The aims of the present chapter are to provide a detailed account of the problems associated with the method of correlated vectors in the light of MGCFA, and to present suitable multi-group factor models to investigate Spearman's hypothesis and other competing hypotheses. These models are fitted to a published data set. The upshot of our critique of the method of correlated vectors is that the the Spearman correlation, as a test of the importance of g in b-w differences, is lacking in specificity. The results of our MGCFA's suggest that it is very difficult to distinguish between competing hypotheses concerning the latent sources of b-w differences.