Internal consistency reliability (Cronbach's alpha) estimates and standard errors of measurement were determined for the Personality Assessment Inventory (PAI; Morey, 1991) with a group of psychiatric inpatients (N = 111). Full-scale reliabilities were large and acceptable, averaging .82. Subscale reliabilities were lower, averaging .66. Reliability estimates for full scales and subscales were comparable to those reported for the clinical portion of the PAI standardization group. Of the 20 full scales examined, only 2 (Somatization and Paranoia) had reliabilities (.85 and .80, respectively) that were significantly lower than comparable values reported for the standardization group. Despite statistically significant differences, these values were still considered large and acceptable.

The relative performance of Kaufman's (Kaufman, 1990; Kaufman & Ishikuma, 1989) dyad, triad, and tetrad subtest‐reduction short forms of the WAIS‐R was evaluated for a group of 100 psychiatric inpatients. Kaufman's triad, which consists of Information, Picture Completion, and Digit Span, was judged superior, based on its brief administration time, ease of scoring, reliability, and predictive accuracy. The Kaufman triad also was compared with the Silverstein (1982) and the Reynolds, Willson, and Clark (Reynolds, Willson, & Clark, 1983; Silverstein, 1990b) tetrads and performed just as well as, if not better than, both of these short forms on all of the evaluative criteria.

The relative performance of six Wechsler Adult Intelligence Scale‐Revised short forms (Brooker & Cyr, 1986; Reynolds, Wilson, & Clark, 1983; Silverstein, 1982) was examined for a group of psychiatric inpatients (N = 50). Short forms were evaluated according to their ability to predict the Full Scale IQ using regression analysis. The standard error of estimate was significantly smaller for Silverstein's (1982) four‐subtest short form, which consists of Vocabulary, Arithmetic, Picture Arrangement, and Block Design. Using the Silverstein (1982) tables, 90% of the short form IQs fell within ±6 points of the actual Full Scale IQ.

Internal consistency reliability (Cronbach's alpha) estimates and standard errors of measurement were determined for the Personality Assessment Inventory (PAI; Morey, 1991) with a group of psychiatric inpatients (N ϭ 111). Full-scale reliabilities were large and acceptable, averaging .82. Subscale reliabilities were lower, averaging .66. Reliability estimates for full scales and subscales were comparable to those reported for the clinical portion of the PAI standardization group. Of the 20 full scales examined, only 2 (Somatization and Paranoia) had reliabilities (.85 and .80, respectively) that were significantly lower than comparable values reported for the standardization group. Despite statistically significant differences, these values were still considered large and acceptable.

Intrasubtest scatter values for eight WAIS-R subtests were estimated for a sample of 150 psychiatric inpatients and compared with those reported for a comparable portion of the WAIS-R standardization group by Kaplan, Fein, Morris, and Delis in 1991, using the normal deviate (z) test. The standardization group's intrasubtest scatter was significantly greater than those reported for the psychiatric inpatients on all subtests except Arithmetic and Information. These results suggested that, contrary to expectation, limited intrasubtest scatter may characterize WAIS-R protocols of psychiatric inpatients, while the presence of scatter may characterize normal WAIS-R performance.