Woodcock-Johnson-III, Kaufman Adolescent and Adult Intelligence Test (KAIT), Kaufman Assessment Battery for Children (KABC), and Differential Ability Scales (DAS) Support Carroll but Not Cattell-Horn
Recently emerging evidence suggests that the dominant structural model of mental abilities—the Cattell-Horn-Carroll (CHC) model—may not adequately account for observed scores for mental abilities batteries, leading scholars to call into question the model’s validity. Establishing the robustness of these findings is important since CHC is the foundation for several contemporary mental abilities test batteries, such as the Woodcock-Johnson III (WJ-III). Using confirmatory factor analysis, we investigated CHC’s robustness across 4 archival samples of mental abilities test battery data, including the WJ-III, the Kaufman Adolescent & Adult Intelligence Test (KAIT), the Kaufman Assessment Battery for Children (KABC), and the Differential Ability Scales (DAS). We computed omega hierarchical (ωH) and omega subscale (ωS) coefficients for g and the broad factors, which estimated the relationship of composite scores to g and the broad factors, respectively. Across all 4 samples, we found strong evidence for a general ability, g. We additionally found evidence for 3 to 9 residualized, orthogonal broad abilities existing independently of g, many of which also explained reliable variance in test battery scores that cannot be accounted for by g alone. The reliabilities of these broad factors, however, were less than desirable (i.e., <.80) and achieving desirable reliabilities would be practically infeasible (e.g., requiring excessively large numbers of subtests). Our results, and those of CHC critics, are wholly consistent with Carroll’s model. Essentially, both g and orthogonal broad abilities are required to explain variance in mental abilities test battery scores, which is consistent with Carroll but not Cattell-Horn.