Subtest specificities were determined for each of the 11 WAIS-R subtes
ts for a sample of 200 psychiatric inpatients. The subtest's shared va
riance (defined as the multiple squared correlation between a subtest
and all remaining subtests) was subtracted from reliable variance (def
ined as the subtest's reliability) to yield a measure of subtest speci
ficity. The Information, Arithmetic, Vocabulary, Comprehension, Simila
rities, Picture Arrangement, and Digit Symbol subtests were judged to
have adequate to ample specificity for ipsative interpretation. The Ob
ject Assembly, Digit Span, Picture Completion, and Block Design subtes
ts were judged to have inadequate specificity for ipsative interpretat
ion. Implications for the interpretation of WAIS-R results with psychi
atric inpatients are discussed.