ABSTRACT
The Cattell-Horn-Carroll (CHC) taxonomy has been used to classify and describe human cognitive abilities. The ability factors derived from the CHC taxonomy are often assumed to be invariant across multiple populations and intelligence batteries, which is an important assumption for research and assessment. In this study, data from five different test batteries that were collected during separate Kaufman Assessment Battery for Children-Second Edition (KABC-II; Kaufman & Kaufman, 2004) concurrent validity studies were factor-analyzed jointly. Because the KABC-II was administered to everyone in the validity studies, it was used as a reference battery to link the separate test batteries in a "cross-battery" confirmatory factor analysis. Some findings from this analysis were that CHC-based test classifications based on theory and prior research were straightforward and accurate, a first-order Fluid/Novel Reasoning (Gf) factor was equivalent to a second-order g factor, and sample heterogeneity related to SES and sex influenced factor loadings. It was also shown that a reference variable approach, used in studies that incorporate planned missingness into data collection, may be used successfully to analyze data from several test batteries and studies. One implication from these findings is that CHC theory should continue to serve as a useful guide that can be used for intelligence research, assessment, and test development.
