In the WAIS-R Manual, Wechsler (1981, Table 17) provided correlations between the WAIS and WAIS-R for 72 adults aged 35-44, tested on both instruments in a counterbalanced order with an interval of 3 to 6 weeks between administrations. He reported WAIS versus WAIS-R coefficients of .91 for the Verbal Scale, .79 for the Performance Scale, and .88 for the Full Scale.

Ryan, Nowak, and Geisser (1987, Table 2) summarized the results of eight groups, including the one cited by Wechsler (1981), that were tested on both the WAIS and WAIS-R. These samples ranged in size from 29 mildly and moderately retarded adults (Simon & Clopton, 1984) and 29 medical and psychiatric patients (Warner, 1983) to 88 college students (Mishra & Brown, 1983). Table 3.9 summarizes the correlations between IQ scales and subtests of the same name on the WAIS and WAIS-R based on the results of these eight samples, which totaled 420 individuals. Median coefficients are reported, rather than means, because of several uncharacteristically low values reported by R. S. Smith (1983) for 70 college students. Mean test-retest reliability coefficients for the WAIS-R, obtained from data on two samples provided by Wechsler (1981, Table 11), are also shown in Table 3.9 to serve as a reference point for evaluating the WAIS/WAIS-R coefficients.

Overall, the corresponding WAIS and WAIS-R IQs correlated quite substantially (.84 to .92) and not much lower than the respective WAIS-R IQs correlated with each other in a test-retest situation (.90 to .96). The correlation of .87 between V-IQ on the W-B I and WAIS-R, reported by Stewart (1981) for 44 normal Caucasian, middle-class adults (median age = 31.5), is wholly consistent with the coefficients observed for the W-B I and WAIS and for the WAIS and WAIS-R.

Five of the 11 subtests on the WAIS and WAIS-R correlated about as high with each other as one could reasonably expect in view of WAIS-R

TABLE 3.9 Summary of correlations between WAIS and WAIS-R IQs and scaled scores for eight samples tested on both instruments

Mean WAIS-R Range of Median Stability

Subtest or Scale Correlations Correlation Coefficient

Verbal

Information .47 to

Digit Span .66 to

Vocabulary .26 to

Arithmetic .27 to

Comprehension .51 to

Similarities .33 to

Performance

Picture Completion .30 to

Picture Arrangement .15 to

Block Design .74 to

Object Assembly .14 to

Digit Symbol .29 to

IQ Scales

Verbal .73 to

Performance .76 to

Full Scale .85 to

Note: Data for the WAIS/WAIS-R comparisons are based on Table 2 in a summary article by Ryan et al. (1987). Mean WAIS-R stability coefficients, provided for comparison purposes, are computed from data provided by Wechsler (1981, Table 11) for 71 people aged 25—34 and 48 individuals aged 45-54.

stability coefficients: Information, Vocabulary, Arithmetic, Similarities, and Block Design each correlated within .05 of the test-retest coefficient. If .10 is selected as an arbitrary "allowable" discrepancy between WAIS/WAIS-R correlation and WAIS-R stability coefficient, then only two subtests, Picture Completion and Picture Arrangement, show an inadequate relationship. The WAIS and WAIS-R Picture Completion subtests correlated .68, well below its WAIS-R stability coefficient. The two Picture Arrangement tasks correlated a moderate .58, not nearly as high as its test-retest coefficient of .72 on the WAIS-R.

The three IQs yielded by the WAIS and WAIS-R correlated so well that their comparability is axiomatic. Nine of the 11 subtests also correlated sufficiently well to support their comparability on the WAIS and WAIS-R. However, the two Picture Completion and the two Picture Arrangement subtests do not relate well enough to support their continuity of measurement from the WAIS to the WAIS-R. Not coincidentally, these two subtests were modified substantially when the WAIS was revised (Wechsler, 1981). They retained only 60% to 70% of the WAIS items; among other subtests, only Information rivaled that percentage (see Table 3.1).

The WAIS-III/WMS-III Technical Manual (Psychological Corporation, 1997) provided data on a sample of 192 adults who were administered both the WAIS-R and WAIS-III. The sample was administered the two tests in counterbalanced order with a 2- to 12-week interval between the tests (median = 4.7 weeks). The correlation coefficients were calculated in a two-step process to account for differential practice effects; the results of the analysis are presented in Table 3.10. As shown in the table, which also includes WAIS-III stability coefficients for comparison purposes, the correlation between tests was .94 for Verbal IQ, .86 for Performance IQ, and .93 for Full Scale IQ.

Overall, the corresponding WAIS-R and WAIS-III Verbal subtests correlated better than the Performance subtests of the two versions of the test. However, this finding is to be expected, given that the test-retest coefficients for the WAIS-III are generally more stable for the Verbal subtests (range =.75-.94; median = .83) than the Performance subtests (range = .69-.86; median = .78). Six of the 11 WAIS-III subtests that are the same for the two versions of the test correlated strongly with one another (given what one could reasonably expect based on their WAIS-III stability coefficients). Digit Span, Vocabulary, Comprehension, Similarities, Picture Completion, and Block Design each correlated within .05 of the test-retest coefficient. Only one subtest, Information, had a discrepancy between the correlation and stability coefficient that was larger than 0.10 (correlation = .83, stability coefficient = .94). Of the Verbal subtests, Information retained the lowest amount of items from its previous version (65.5%). However, given the very high stability coefficient, the .83 correlation between the two Information subtests is still relatively strong. The two subtests that had the lowest correlations between the WAIS-R and WAIS-III were Picture Completion (.50) and Picture Arrangement (.63). However, this lack of continuity is not surprising given that only 50% of the items from the WAIS-R were retained on the WAIS-III for each of these two subtests (see Table 3.1). Like the relationships between subtests on the WAIS and WAIS-R, most coefficients obtained for the WAIS-R and WAIS-III were impressive.

Was this article helpful?

Whenever a doctor informs the parents that their child is suffering with Autism, the first & foremost question that is thrown over him is - How did it happen? How did my child get this disease? Well, there is no definite answer to what are the exact causes of Autism.

## Post a comment