Diagnostic use of the WISC-R and McCarthy Scales: A regression analysis approach to learning disabilities

Thirty learning disabled and 30 nonlearning disabled students were individually administered the WISC-R, McCarthy Scales of Children's Abilities, and the Wide Range Achievement Test. Regression analyses were conducted to determine the combination of scores from the WISC-R and McCarthy Scales that best predicted the achievement level of the subjects. In addition, the scores that best predicted group status (LD or nonLD) were determined. In general, the WISC-R Comprehension, Arithmetic, and Object Assembly, and the McCarthy Quantitative and Memory Indices were most sensitive to the LD students' achievement. Conversely, the WISC-R Similarities and Arithmetic and the McCarthy Verbal Index were most sensitive to the achievement of nonLD students. Finally, the McCarthy Perceptual-Performance Index and the WISC-R Vocabulary subtest best discriminated group status. The diagnostic implications of these results were discussed.     

An investigation of WISC-R verbal-performance differences as a function of ethnic status

The Wechsler Intelligence Scale for Children-Revised was administered to 187 black, 184 Hispanic, and 184 white children aged 6–11. The Verbal-Performance (V-P) IQ discrepancies were determined for the total group and for each ethnic group. The overall mean of 10.74 was similar to the V-P discrepancy found in the WISC-R standardization sample. The Hispanic group had a significantly greater V-P split than did the black group or the white group. Implications of these results were discussed, focusing on the various cautions of using V-P discrepancies for diagnostic decision making.     

Factor structure of the McCarthy screening test

The factor structure of the McCarthy Screening Test was investigated with a sample of 555 children. The sample consisted of 300 boys and 255 girls, whose mean age was 61.4 months; approximately 80% were white and 20% were black. The data from the present study and from the MST standardization samples were analyzed by a principal components analysis (with ones in the diagonal) followed by a varimax rotation. The results from the data analysis seem to indicate that the subtests all measure parts of the same attribute (cognitive and sensorimotor functions) in varying degrees.     

Subtest scatter on the WISC-R with children of superior intelligence

Subtest scatter on the WISC-R was examined by analyzing Verbal-Performance IQ discrepancies, subtest scaled-score ranges and subtest scaled-score standard deviations of 290 psychoeducationally normal 9-year-old children whose IQs ranged from 100 to 140+. Although there were no differences between average and superior IQ groups on the Verbal-Performance discrepancy measure, substantial differences were found on the measures of subtest scatter where higher IQ groups exhibited substantially more scatter. The results suggest the need for caution in attempting to employ WISC-R subtest scatter as an indicator of learning disabilities with gifted students.     

Subtest scatter and Kaufman regroupings on the WISC-R in non-learning-disabled and learning-disabled children

The present study was undertaken to examine scatter analysis and Kaufman regroupings of WISC-R scores. Thirty-one non-learning-disabled children (NLD) and 29 learning-disabled children (LD) determined by individual IQ and achievement scores served as subjects. Scatter indices were computed within the Verbal and Performance subtests as well as between the Verbal and Performance IQ scales. Also, the regrouping categories proposed by Kaufman (1979) were examined. A significant difference in Verbal and Performance IQs was found between the LD and NLD group. Significantly more students in the LD group exhibited the Kaufman regrouping pattern than in the NLD group (p < .01). These results suggest that Kaufman regroupings and Verbal-Performance scatter analysis may be used to assist in diagnosing LD children.     

Do normal children have “flat”; ability profiles?

Scatter on the five specific scales in the McCarthy was examined for the standardization sample to explore the types of ability profiles exhibited by normal children. Contrary to the common stereotype of “flat” profiles, normal children aged 21/2 to 81/2 years evidenced much variability in their performance on the various cognitive and motor dimensions.     

Use of the WISC-R with a learning disabled population: Some diagnostic implications

A group of 41 learning disabled children in Hawaii, aged 6 to 151/2, were tested on the WISC-R. Two main questions were explored: (a) Do the separate WISC-R tests assess “g” for learning disabled children to the same degree that they do for normal children? (b) Is there significantly more scatter among the tests for learning disabled than for normal youngsters? The answers to these questions were interpreted in terms of their diagnostic significance.     

Relationship between PPVT and WISC-R in children with reading disabilities

The Peabody Picture Vocabulary Test (PPVT), Form A was compared to the Wechsler Intelligence Scale for Children-Revised (WISC-R) in two samples of children with reading disabilities. One group of 14 children, referred to a university clinic, were administered the WISC-R, followed by the PPVT. The second group of 38 children from a private learning disability center were administered the PPVT first, followed by the WISC-R. In the combined sample, the PPVT IQ (X̄ = 109.2) was significantly higher than the WISC-R Verbal IQ (X̄ = 98.9), Performance IQ (X̄ = 97.0), and Full Scale IQ (X̄ = 97.5). Similarly, the PPVT IQ was significantly higher than the WISC-R Full Scale in both samples separately, regardless of which test was administered first. In one case, the PPVT IQ was 50 points higher than the WISC-R IQ. Correlations between the PPVT and WISC-R Verbal, Performance, and Full Scale IQs were significant (rs = .56, .29, and .50, respectively). The results suggest that the two tests do not provide interchangeable IQs for a population of reading disabled children.     

The long-term stability of WISC-R scores in mildly retarded and learning-disabled children

This study addresses the need for systematic longitudinal research documenting the stability of WISC-R scores in special education populations. WISC-R scores of 100 learning-disabled and 60 mildly retarded children retested on three separate occasions at three-year intervals were examined. The stability of WISC-R scores was evaluated according to three different criteria: (a) the consistency of group means over time, (b) the frequency of significant changes in individual scores, and (c) correlations between administrations as an index of stability of subjects' relative positions in the group. Different results were obtained depending on the criterion considered. Examination of group means and correlation coefficients indicated that Full Scale IQ was fairly stable over a period of six years for both learning-disabled and mildly retarded samples. However, greater variability was noted when examining the frequency of changes in individual subject's scores. Verbal IQ and Performance IQ demonstrated somewhat more variability by all criteria examined. The implications of these results with regard to the importance assigned to IQ in special education classification decisions, the usefulness of retesting IQ in three-year reevaluations, and the efficacy of special education are discussed.     

The reliability of the WISC-R performance scale with deaf children

Though the Wechsler tests are widely used to assess the intellectual ability of hearing impaired children, data on the reliability of the WISC-R with this population have been lacking. The Performance Scale of the WISC-R was administered to 59 prelingually deaf children attending a state-supported day school program. The reliability coefficients and standard errors of measurement were computed for each subtest and for the Performance Scale IQ. The results compare favorably with those found by Wechsler with the standardization sample.     

Comparison of the slosson intelligence test,revised norms,and WISC-R for children with learning problems and for gifted children

The Slosson Intelligence Test (revised norms) (SIT) and the WISC-R were compared in two samples of children. In the first sample, there were 34 black and 27 white rural Southeastern Alabama children being considered for special education classes. In the second sample, there were 4 black and 81 white suburban Alabama children being considered for classes for the talented and gifted. In both samples, correlations between the SIT IQ and WISC-R Full Scale IQ were significant (rs of .70 and .48, respectively). However, in the special education sample, SIT IQs were significantly higher than WISC-R Full Scale IQs by about 7 points. The results from both samples provide a moderate degree of support for the concurrent validity of the revised SIT norms, using the WISC-R as the criterion. However, the IQs on the two tests may not be interchangeable.     

A comparison of the factor structure of the WISC-R for blacks and whites

Scaled scores of the 10 mandatory WISC-R subtests were factor analyzed for separate groups of black (N = 430) and white (N = 274) public school students. Two-, three- and four-factor solutions were examined using the iterated principal factor method. The factorial solutions were compared using coefficients of congruence, and each solution was compared to the standardization sample. Coefficients of congruence between the black and white first factor were .96, second factor .99, and third factor .74. Congruency between Kaufman's results and each racial group was also quite high, suggesting that the WISC-R measures the same theoretical abilities.     

A longitudinal study of the Stanford-Binet and WISC-R with special education students

A group of 95 black and 52 white low socioeconimic children assigned to special education calsses were initially given the Stanford-Binet (1972 norms) and three years later the WISC-R. Correlations between the Stanford-Binet and WISC-R Full Scale IQs were significant in both ethnic groups, with r=.60 for the total group. Correlations between the Stanford-Binet and WISC-R subtests also were generally signiificant (rs between .29 and .52 for the total group). Stanford-Binet and WISC-R Full Scale IQs did not differ significantly. For this sample of special education children, the Stanford-Binet was found to have satisfactory predictive validity.     

The WISC-R verbal comprehension,perceptual organization,and freedom from distractibility deviation quotients: Data for practitioners

Utilizing data from the WISC-R standardization sample (Wechsler, 1974), Gutkin (1978) presented formulae to calculate factorially pure verbal comprehension, perceptual organization, and freedom from distractibility deviation quotients. Data are presented concerning standard errors of measurement and the score discrepancies that are required for statistical significance between these WISC-R scores.     

Factor structure of the McCarthy Scales for Children experiencing problems in school

The McCarthy Scales of Children's Abilities (MSCA) were factor analyzed for a group of 6- to 8 1/2-year-old children referred for school-related problems. The sample was composed of 300 rural eastern North Carolina public school children with a mean GCI of 88.1. Evidence was found for factors similar to three of the six MSCA Scales: General Cognitive, Verbal, and Motor. Coefficients of congruence were computed and compared with earlier analyses of the standardization data. The results were interpreted as generally supporting the MSCA's validity with children experiencing problems in school. Implications for interpretation of the MSCA Scales were presented.     

Scatter on the WISC as an indicator of intellectual potential

The study was undertaken to investigate the relationship between subtest scatter on the WISC and higher intellectual potential. All subjects at one time obtained Full Scale IQ scores within the retarded range. Group I consisted of those children who upon reexamination with the WISC obtained Full Scale IQ scores of 75 or below, while Group 2 consisted of those children whose Full Scale IQ scores were now found to be nonretarded upon reexamination. The results indicated that the scatter on the initial examinations failed to predict the future group membership. Irrespective of group membership, scatter was found not to be related to change in Full Scale IQ scores. It was concluded that subtest scatter is a poor indicator of the presence of higher intellectual potential. It was also suggested that these findings are relevant to the WISC-R.     

A comparison of the Stanford-Binet Intelligence Scale and the McCarthy Scales of Children's Abilities with preschool children

The purpose of the present study was to determine the degree to which performance on the McCarthy Scaies of Children's Abilities correlated with performance on the Stanford-Binet for a group of preschoolers. The sample consisted of 44 children ranging in age from 3–11 to 5–4. It was found that the General Cognitive Index scores of the McCarthy Scales correlated well with the Stanford-Binet IQ scores, r^xy = .90. However, 40 of the 44 subjects scored higher on the Stanford-Binet than on the McCarthy Scales.     

A short-form version of the WISC-R for the learning disabled

This study investigated the best short form version of the WISC-R for learning disabled (LD) students. The WISC-R was administered to 234 LD and 290 children who were referred for psychological evaluation but were found not eligible (NE) for special education services from 45 school districts in Michigan. Stepwise multiple regressions were calculated to ascertain the best tetrads of Full Scale IQ (FS IQ) for the two groups. For the NE sample, Vocabulary, Picture Arrangement, Picture Completion, and Information were the best four indicators of FS IQ, accounting for about 80% of the variance in Full Scale scores. For the LD sample, the best four predictors of FS IQ were Similarities, Block Design, Picture Completion, and Vocabulary, accounting for about 80% of criterion variance. Using the WISC-R standardization data, Sattler identified the Information, Vocabulary, Comprehension, and Block Design subtests as one of the best short form versions for predicting FS IQ. For the LD sample of the present study, Sattler's composite yielded a multiple R of .873 with FS IQ, whereas our best four predictors yielded a multiple R of .896. It appears that Sattler's tetrad is about as good a predictor of FS IQ for LD students as that found in the present data.     

A four-test short form of the WISC-R

A four-test short form of the WISC-R was developed on the basis of both rational and statistical criteria, using data from the representative standardization sample. The short-form comprised Arithmetic and Vocabulary from the Verbal Scale along with Picture Arrangement and Block Design from the Performance Scale. Various psychometric properties of the abbreviated battery were defined, and indicators of the short form's accuracy in predicting Full-Scale IQ (the standard error of estimate, effectiveness in classifying a child's intelligence level) were provided. A major topic of discussion was the use and abuse of short forms.     

The use of Wechsler Scales in the assessment of Native Americans of the Columbia River Basin

Native American secondary students from the Columbia Basin were found to have significant Verbal-Performance discrepancies on the WISC-R and WAIS. Mean Verbal scores were significantly below the normative mean, while Performance scores were at, or above, the normative means. These findings substantiated research with other Native American groups. Also, the Verbal and Performance scales correlated so low as to preclude the Full Scale from being an accurate representation of the “g” factor of intelligence. Further, the predictive validity of the WISC-R and the WAIS for reading and math achievement was found to be at variance with the standardization group.