The WAIS-III as a Nested Factors Model

Abstract

Past confirmatory factor analytic (CFA) research that has examined the factor structure of the WAIS-III has only investigated the more popular models, such as oblique factors and/or higher-order models. In contrast, CFA modeling based on nested factors modeling has been neglected. Consequently, this study investigated the model fit of various WAIS-III nested factors models, in comparison to the more traditional higher-order models. Based on the WAIS-III standardization sample, the results associated with the nested factors modeling indicated that Digit Span, Arithmetic, and Letter-Number Sequencing did not share any variance with VIQ, independently of “g.” Further, across all age groups, there was only very weak evidence in favor of the plausibility of a Perceptual Organization (PO) factor, independently of a general factor. The results are discussed in light of the distinction between modeling “g” as a higher-order factor and a first-order factor. Researchers are encouraged to model the subtests of intelligence batteries as nested factor models, because of their tendency to be associated with greater model fit, as well the possibility to obtain less ambiguous factor loading estimates and their associated statistical significance.