Via the grapevine, I have become aware of some misinterpretation of select material and slides I presented recently at the Ohio and GA school psychology conferences. This note is to clarify the point I made (and did not make)
I presented the published intercorrelation matrix for the WISC-IV subtests from Table 5.1 of the WISC-IV Integrated manual. I used those correlations to demonstrate that certain field-based rules-of-thumb (subtests need to be 1 or 1.5 SD discrepant; which typically then refers to subtest scales scores of 3 and 5 on the WISV-IV) can cause problems and is not always accurate. I presented similar correlations and findings for select WJ III tests.
Their is NOTHING incorrect with the WISC-IV Table 5.1 Also, at the GASP workshop I made it clear, in response to a question, that the statistical significance and base rate information provided for the WISC-IV manuals and scoring software are technically correct and psychometrically sound. To be clear, the information in the various Wechsler manuals is correct and should be used to evaluate subtest discrepancies.
The point made was that certain field based "rules-of-thumb" cannot assume that 1SD or 1.5SD equals 3 and 5 in the case of scaled scores or 15 and 23 in the case of measures with a scale of M=100 and SD=15.
The problem is in the generic "rules-of-thumb" that fail to recognize that the proper statistic in terms of SD of discrepancies varies as a function of the purpose of the comparison (ability-achievement discrepancy; evaluating difference between two tests in same CHC domain where there is an implied correlation---unitary vs non-unitary indicators of domain; and simple score differences). There is NOTHING wrong in the Wechsler manuals or scoring software. The problem lies in potential misuse and understanding of "SD" based rules-of-thumb that have rattled around our field as clinical lore for decades.
I am working on a post or small report to explain this better.
- iPost using BlogPress from my Kevin McGrew's iPad