Journal of Psychoeducational Assessment

Intuition suggests that when the score on one subtest is compared with the average score on a group of subtests, the outcome ought to depend on whether the score in question is included in or excluded from the average. In fact, it does not. The equivalence of the two procedures, inclusion and exclusion, is demonstrated for three varieties of pattern analysis. In addition, some historical observations are offered that testify to the almost irresistible intuitive appeal of the notion that inclusion or exclusion does make a difference.

A confirmatory factor analysis was used to determine the respective viability of factor structure models for the WAIS-R standardization data. The traditional Verbal-Performance split was compared to an alternative model proposed by Gutkin, Reynolds, and Galvin (1984). The confirmatory analysis indicated that the Gutkin model provides a superior fit to the data at all age levels and for the total sample. In addition, the results from Gutkin model analyses at each age level were iterated, yielding a refinement in the respective age-specific fits of the model. Implications for test interpretation are discussed.

The validity of Bannatyne's recategorization of WISC-R subtests was tested by applying multiple group factor analysis to the intercorrelation matrices of the 7.5-, 10.5-, and 13.5-year samples from the WISC-R standardization. Although the Acquired Knowledge category had a high reliability coefficient, it was almost completely linearly dependent on the other categories, particularly Conceptualization. Consequently, differences between these two categories will be unreliable. It was concluded that scores on Acquired Knowledge provide no unique information, and their interpretation as a measure of educational attainment, apart from verbal ability, was discouraged.

This article provides a discussion of a number of statistical and procedural aspects of psychoeducational studies that can influence the interpretations and conclusions drawn from such research. The topics covered relate to situations that are commonly encountered by researchers and practitioners in the conduct and interpretation of psychoeducational studies. In many cases, guidelines are provided for the practitioner regarding how each of the aspects discussed is most likely to influence the results of a research study. Most of the topics included are not covered in depth but are accompanied by sufficient references for the interested reader to pursue each topic in more depth.

The Woodcock-Johnson Battery contains four Scholastic Aptitude clusters designed to predict achievement. Based on stepwise multiple regression development procedures, the subtests within these clusters are weighted differently. Concerns regarding the inclusion of the Quantitative Concepts subtest in two of these aptitude clusters were investigated in a referral sample (N = 386). The results suggest that the achievement content of Quantitative Concepts subtest, particularly because of its relatively high weighting, may inappropriately affect the scores on the Written Language and Knowledge Aptitude clusters. Implications for interpretation are discussed.

Procedures for validating the complexity level of test items include empirical examinations of item characteristics and subjective judgments of expert raters. These methods have yielded mixed and inconsistent results. This study employs a technique derived from cognitive psychology research in the examination of levels of item complexity. Using a microcomputer to administer test items, response time to a secondary task was used as an index of the level of item complexity. A series of three studies are reported that (1) employ the technique on items of marked discrepancy in complexity level and (2) contrast the sensitivity of the technique to more traditional analyses of level of item complexity. The results suggest that the technique has promise as a mechanism for validating or establishing level of item complexity.

The school performance of low-income Mexican-American children was predicted from scores on three intelligence tests. The Stanford-Binet (S-B) was given at 3 years, S-B and McCarthy Scales of Children's Abilities (MSCA) at 4 years, and the Wechsler Intelligence Scale for Children-Revised (WISC-R) at 7-9 years. School grades, Iowa Test of Basic Skills (ITBS) scores, and the Classroom Behavior Inventory (CBI) comprised the school performance measures. Grades were predicted significantly by S-B, but there were no significant correlations for the other tests. All the intelligence tests had some significant correlations with some of the ITBS and CBI scores. The S-B and MSCA were about equal as preschool predictors of ITBS and CBI scores. The controversy over the value of the WISC-R in assessing minority-group children was discussed in view of the finding that WISC-R scores were significantly correlated with the ITBS but not with grades.