Using Two WISC-IV Indices to Identify the Gifted

Whenever a new version of an IQ test is released, it is necessary to determine its effectiveness in identifying gifted students. The Wechsler Intelligence Scale for Children, Fourth Edition (WISC-IV), released in August of 2003, is structured in a significantly different manner from its predecessor, the WISC-III. The Verbal and Performance sections have been replaced by four indices: Verbal Comprehension, Perceptual Reasoning, Working Memory and Processing Speed. Object Assembly, Picture Arrangement and Mazes have been removed. Word Reasoning, Matrix Reasoning, Picture Concepts, Letter-Number Sequencing and Cancellation have been added. The Arithmetic subtest, which is now a supplementary subtest, has been altered so that subjects are given only 30 seconds to answer each item and none of the items for older children is presented visually. It is clustered with the tests of Working Memory, although it loads heavily on fluid reasoning (Keith, Fine, Taub, Reynolds & Kranzler, 2004).

The WISC-IV allows substantially more time for responses than the WISC-III, which benefits gifted students but adds administration time. It also has added more difficult items for better differentiation in the higher IQ ranges. There are now 15 subtests: 10 required and 5 optional. Substitutions of optional tests are allowed, which gives the test greater flexibility; however, the long administration time will discourage school psychologists from giving the optional tests.

The norms for 63 gifted students reported in the Technical Manual are significantly lower on the WISC-IV than on previous IQ tests. The mean Full Scale score of the gifted group was 123.5; Verbal Comprehension Index, 124.7; Perceptual Reasoning Index, 120.4, Working Memory Index, 112.5 and Processing Speed Index, 110.6. Similar discrepancies between the first two and last two indices have been revealed in our sample at the Gifted Development Center in Denver, Colorado. When the 15 subtests are ranked from highest to lowest for the gifted sample, it is apparent that five of the six subtests that produced the lowest scores for the gifted group are required to obtain a Full Scale IQ score, whereas four of the five optional subtests, unlikely to be administered due to time constraints, yield scores among the highest of the gifted group (Flanagan & Kaufman, 2004). The subtests on which the gifted sample scored the lowest make up the Working Memory and Processing Speed Indices.

Should the cutoff scores for gifted programs be reduced when using the WISC-IV Full Scale IQ scores for selection? Alternatively, are Verbal Comprehension and the Perceptual Reasoning Indices more appropriate for assessing giftedness than Working Memory and Processing Speed Indices? The purpose of this paper was to address these questions with results of assessments of 103 gifted children.

In a presentation given last year to the Research and Evaluation Division of the National Association for Gifted Children (Falk, Silverman & Moran, 2003), we compared IQ scores on the WISC-III and the Stanford-Binet Intelligence Scale (Form L-M) (SBL-M) for a stratified random sample of 120 cases in four IQ ranges from the clinical files of the Gifted Development Center. The SBL-M, which has a higher ceiling, generated significantly higher scores than the WISC-III, particularly in the highest IQ ranges. In that study, any inflation caused by the Flynn Effect (Flynn, 1999) was corrected at the rate of 0.3 per year for an estimated 3 decades; a total of 9 points was subtracted from each SBL-M score. Based on this analysis, we recommended the continued use of the SBL-M as a supplemental test for children who attain scores near the ceiling of the Wechsler scales, in order to be able to identify children in the highly, exceptionally, and profoundly gifted ranges.

In the past year, we administered the SBL-M to children who obtained 2 or more scores of 17, 18 or 19 on the WISC-IV. At the time of writing the proposal for this presentation, we anticipated that at least 100 children would take the WISC-IV and that 40% of our sample would be given the SBL-M as a supplemental test. The data on 103 cases were analyzed to determine the rank order of indices and subtests of the WISC-IV.

Research Questions

Sample

The Gifted Development Center is a nonprofit, tax exempt agency, which originated in June, 1979, and is now in its 25th year. It is completely supported by fees for services. Most of the families represented are from Caucasian middle and upper-middle class urban backgrounds. Approximately 55% of the clients are from Colorado, 40% are from other states and 5% are from other countries. Parents seek assessment at the Center to determine if their children are eligible for acceptance in special programs for the gifted and to assist them in understanding their children so that they can meet their needs at home and at school. The age range of the children served is 4 through 16; services are also available for adults. The number of children assessed annually ranges between 220 and 450. Several different instruments are used for assessing intelligence.

This study concentrated on the assessment results of 103 children, aged 6-14, who were administered the WISC-IV. In addition, a substantial number were administered the SBL-M as a supplemental test. When children attained 2 subtest scores at or above the 99 th percentile, it was generally recommended that the SBL-M be administered as a supplemental test. While we anticipated that at least 40% of the sample would be administered the SBL-M, in fact we only have 36 subjects reflecting 35% of the sample.

Data Analysis

The first three hypotheses were answered with descriptive statistics, using rank ordering and tests of mean differences. We had hoped to examine hypotheses 4 through 7 with first order correlations and linear regression analyses, but the results were unstable. Instead, descriptive graphs are used to depict trends.

Research Hypotheses

1. The Verbal Comprehension Index will provide the highest score among the indices.

2. The Perceptual Reasoning Index will provide the second highest score among the indices.

Friedman’s test of rank differences for multiple dependent samples and Wilcoxon’s Sign Rank Matched Pairs test support these two hypotheses. The results show the mean and rank order differences for all four index scores.

The Friedman test shows that there is an overall difference between the four indices. To test for the individual differences the Wilcoxon test is reported below. Here we see that the Verbal Comprehension Index Score is significantly higher than the Perceptual Reasoning, Working Memory, and Processing Speed Indices. The Perceptual Reasoning Index Score is significantly higher than the Working Memory and Processing Speed Indices, and the Working Memory Index Score is significantly higher than the Processing Speed Index.

Hypothesis 3 states that subjects with Verbal Comprehension Indices in the gifted range (130 or above) will have Working Memory and Processing Speed scores below the gifted range. The following two graphs vividly portray the empirical support for this hypothesis.

Chi-square analyses for the 63 subjects with Verbal Comprehension indices in the gifted range clearly support the third hypothesis. In the following tables, the value 1 reflects index scores below 130 and value 2 reflects index scores at 130 and above. As can be clearly seen, subjects who scored in the gifted range on the Verbal Comprehension index often did not score in the gifted range on the Working Memory or Processing Speed Indices.

For the following research hypotheses we have too few cases to establish stable predictions and rather than present data which could be misleading we have chosen not to report regression results. In all cases, the standard errors for the unstandardized regression coefficients were too large to establish any significant results.

4. Vocabulary, Similarities and Comprehension will predict SBL-M Formula IQ scores.

5. Block Design, Matrix Reasoning and Picture Concepts will be the second best set of predictors of SBL-M Formula IQ scores.

6. Digit Span and Arithmetic will be the third best set of predictors of SBL-M Formula IQ scores.

Research Questions

We can provide some visual graphic information that is suggestive of some answers, but again these are based on only 36 cases. The graphs show the relationship between Stanford-Binet Intelligence Scale Formula IQ score (Form L-M) and the Index Scores for the WISC-IV. The answer to the first research question appears to be the Verbal Comprehension Index.

The answer to the first research question appears to be that the Verbal Comprehension Index is the best of the four indices for predicting giftedness.

With regard to the second research question preliminary analyses suggest that Arithmetic, Comprehension, Digit Span and Matrix Reasoning appear to predict higher performance on the SBL-M. These initial findings, based on only 36 subjects, are necessarily quite tentative. As they are not consistent with the rank order of subtest performance of the full sample, they are likely to be modified considerably when more cases are available. At this time, we have decided to not report actual results, which could be illusory. The last two research questions are still to be resolved with more data:

However, the following graphs suggest that using the Full Scale IQ Score of 123 on the WISC-IV may be a reasonable alternative to using the Verbal Comprehension Index or the Perceptual Reasoning Index. These three graphs demonstrate that when you select subjects with a Full Scale IQ Score of 123.5 and above, the distribution of VCI and PRI scores goes well below the score of 130. Additionally, the Full Scale IQ Score of 123 and above picks up 75% of the subjects who had SBL-M Formula Scores of 130 and above.

Conclusion

The WISC-IV is a useful test for the gifted population due to its strengths in measuring verbal and visual reasoning; however, it does have drawbacks that adversely affect its ability to adequately locate gifted children for special programs. If the Verbal Comprehension Index and the Perceptual Reasoning Index are combined into a General Ability Index, as suggested by Flanagan and Kaufman (2004) and by recent information from the publisher, Psychological Corporation (Harcourt Assessments), the test increases in power to select gifted students. Alternatively, if the Full Scale IQ requirement for gifted programs is reduced to 123, the test can also find most of the students who would have been selected by either a WISC-III or an SBL-M. Other measures with higher ceilings, such as the Stanford-Binet Intelligence Scales, will need to be administered to locate children for highly gifted programs. These measures, too, will probably require modifications in scoring practices to be fully applicable.

References

Falk, R. F., Silverman, L. K., & Moran, D. M. (2003, November). WISC-III and Stanford-Binet L-M Scores for gifted children. Paper presented at the 50 th Annual Convention of the National Association for Gifted Children, Indianapolis, IN.

Flanagan, D. P., & Kaufman, A. S. (2004). Essentials of WISC-IV assessment. Hoboken, NJ: John Wiley & Sons.

Flynn, J. R. (1999). Searching for justice: The discovery of IQ gains over time. American Psychologist, 54, 5-20.

Keith, T. Z., Fine, J. G., Taub, G. E., Reynolds, M. R., & Kranzler, J. H. (2004). Hierarchical multi-sample, confirmatory factor analysis of the Wechsler Intelligence Scale for Children-Fourth Edition: What does it measure? (Manuscript submitted for publication)

To read more on assessment, and our work with the WISC-IV, please see the following articles: