A confirmatory maximum likelihood factor analysis of the WPPSI

Abstract

A simultaneous confirmatory maximum likelihood factor analysis was conducted on the 11 subtests of the WPPSI in the 6 age groups in the normative sample. Four substantive models were fitted to the data and compared: (1) a single-factor model; (2) orthogonal and oblique 2-factor models, in which the 2 factors were composed of the Verbal and Performance subtests, respectively; (3) orthogonal and oblique 3-factor models, based on factor solutions for the WISC-R; and, (4) orthogonal and oblique 4-factor models, based on a Bannatyne recategorization of the WPPSI subtests. Results of these analyses indicate that: (1) the single-factor model fits the data quite well, though there is some mis-specification in this model; (2) the oblique 2-factor model does fit the data slightly better than the single-factor model; and, (3) the 3- and 4-factor models either produce improper parameter estimates or do not converge to a proper solution. It is concluded that the WPPSI largely reflects a general intellectual factor, and that any additional common ability factors play only a small role in determining WPPSI scores.