A Revisitation of the WISC-R Factor Structure with Adequate Sample Size, Systematic Control of Both IQ and Age Ranges, and the Use of Procedures to Limit Overfactorization

Problem statement. This study addressed the problem of inconsistency in the results of factor-analytic studies of the WISC-R by performing analyses where age and size of Full Scale IQ ranges were held constant, and by using adequate sample sizes. Methodology. Factor analyses were performed on the WISC-R scale scores. Principal Axes Factor method with squared multiple Rs in the diagonals was used. Oblimin was the rotation of choice. Sample size varied for each analysis, the largest being 3,493; and the smallest, 225. Results. At the broad IQ ranges a general factor was found. At narrower IQ ranges, evidence of the general factor disappeared. Two factors emerged at the 40 IQ points range; one represented some aspect of verbal ability; the other, some aspect of nonverbal ability. Object Assembly and Block Design formed one factor, whereas Vocabulary, Information, and in some cases, Similarities and Comprehension formed the other factor. At the narrowest IQ ranges, fewer common factors were found at the higher IQ levels. Picture Arrangement, Picture Completion, Coding, Digit Span, and Arithmetic did not load significantly with any of the factors in most of the reduced ranges. Arithmetic and Digit Span had acceptable loadings only in ranges below IQ 100. Whenever Coding remained in the analysis, it acted more as a Verbal subtest than as a Performance subtest. Whenever Picture Arrangement remained in the analysis, it did not load exclusively on thePerformance Scale. Conclusions. This study raised a number of questions with regard to the utility of the Full Scale, Verbal, and Performance IQs. Its results also add weight to those studies suggesting that the WISC-R should not be interpreted in terms of a 3-factor solution. Caution should be exercised when interpreting the traditional WISC-R factors beyond Verbal Comprehension. Caution should also be exercised when using the WISC-III factor interpretations since they were made in light of the WISC-R research