Structure of the Woodcock–Johnson III cognitive tests in a referral sample of elementary school students.

Abstract

The structure of the Woodcock-Johnson Cognitive Battery-Third Edition (WJ Cog) has been extensively explored via confirmatory factor analysis (CFA) with its normative sample, but there has been little research to verify that the same structure holds for students referred for special education services. Likewise, research on the structure of the WJ Cog with exploratory factor analysis (EFA) methods has been rare. Consequently, this study applied both EFA and CFA methods to the scores of 529 elementary school students referred for special education services (95.5% eligible) on the 14 tests of the WJ Cog extended battery. EFA results suggested only 2 or 3 factors, whereas CFA results favored the theoretical 7 factors posited by McGrew and Woodcock. In this theoretical model, a strong general factor accounted for 27% of the total variance and 57% of the common variance, whereas the 7 group factors combined accounted for 21% of the total variance and 43% of the common variance. Reliability, as quantified by ωH, was good for the general factor, marginal for the Gs factor, and poor for the other group factors. Nine of the 14 WJ Cog tests displayed uniqueness values that exceeded their communality. On the basis of this evidence from a referral sample, interpretation of the WJ Cog should be restricted to the Gs and g factors: the Gs factor because it exhibited considerable independence and precision of measurement and the g factor because it has emerged in all investigations of the WJ Cog.

Exploring the latent structure of the Luria model for the KABC‐II at school age: Further insights from confirmatory factor analysis

“…Whereas previous investigations of related ability measures have supported a bifactor structure in which the general factor and the group‐specific factors both have direct effects on observed indicators (e.g., Canivez, ; Canivez, Watkins, & Dombrowski, ; Gignac & Watkins, ; McGill & Dombrowski, ; Strickland, Watkins, & Caterino, ; Watkins & Beaujean, ), the present study indicated that a higher‐order solution is preferred for the KABC‐II Luria model configuration. However, due to model equivalence and the estimation problem encountered (i.e., Heywood case) in specifying the publisher suggested higher‐order model (Figure ) at ages 7–12, it may be argued that the bifactor model should have been given preference because of its advantages (e.g., Reise, ; Rodriguez, Reise, & Haviland, ).…”

Construct validity of the Wechsler Intelligence Scale For Children – FifthUKEdition: Exploratory and confirmatory factor analyses of the 16 primary and secondary subtests

“…Similar results have been observed in studies of the WISC‐IV (Bodin, Pardini, Burns, & Stevens, ; Canivez, ; Gomez, Vance, & Watson, ; Keith, ; Styck & Watkins, ; Watkins, , ) and with other Wechsler scales (Canivez & Watkins, ; Canivez, Watkins, Good, et al ., ; Gignac, , ; Golay & Lecerf, ; McGill & Canivez, , ; Watkins & Beaujean, ; Watkins et al ., ). Nor are these results unique to Wechsler scales as similar findings have been reported with other cognitive scales (Canivez, , ; Canivez, Konold, Collins, & Wilson, ; Canivez & McGill, ; Cucina & Howardson, ; DiStefano & Dombrowski, ; Dombrowski, ; Dombrowski, McGill, & Canivez, ; Dombrowski & Watkins, ; Dombrowski et al ., ; Dombrowski, McGill, et al ., ; Nelson & Canivez, ; Strickland, Watkins, & Caterino, ).…”

Construct validity of the Wechsler Intelligence Scale for Children – Fourth UK Edition with a referred Irish sample: Wechsler and Cattell–Horn–Carroll model comparisons with 15 subtests

“…This has been documented in both EFA and CFA studies of the WISC–IV (Bodin et al ., ; Canivez, ; Keith, ; Nakano & Watkins, ; Styck & Watkins, ; Watkins, , ; Watkins et al ., ) and with other versions of Wechsler scales (Canivez & Watkins, ,b; Canivez et al ., , ; Dombrowski et al ., ; Golay & Lecerf, ; Golay et al ., ; Gignac, , ; Lecerf et al ., ; McGill & Canivez, ; Watkins & Beaujean, ; Watkins et al ., ). Further, these results are not unique to Wechsler scales as similar results were also observed with the DAS–II (Canivez & McGill, ), SB5 (Canivez, ), WASI and WRIT (Canivez et al ., ), RIAS (Dombrowski et al ., ; Nelson & Canivez, ; Nelson et al ., ), CAS (Canivez, ), WJ III (Dombrowski, , ,b; Dombrowski & Watkins, ; Strickland, Watkins, & Caterino, ), and the WJ IV Cognitive (Dombrowski et al ., ). The implication of these consistent findings is that primary interpretive weight should be placed on the omnibus FSIQ rather than the first‐order group factor‐based index scores.…”