To examine the validity of short forms of the Wechsler Intelligence Scale for Children-Revised (WISC-R), the WISC-R was first administered to 106 hospitalized psychiatric patients, aged 8-16. No subjects had a primary diagnosis of mental retardation or learning disability, and one-third were receiving psychotropic medication. WISC-R IQ scores ranged from 70 to 133 (mean of 93.76). The WISC-R was administered during the first month of hospitalization. To obtain short form WISC-R scores, Hobby's Split Half Short Form procedure was used. Scoring was modified for all subtests except Digit Span, Coding, and Mazes, usually resulting in a 40-45% shorter test. Both subtest and IQ scores were obtained. Correlations between long form and short form scores on subtests ranged from .76 to .92. Correlations between long and short form results on derived verbal, performance, and full scale IQs were .96 to .98. However, in all instances the mean value yielded by the short form was higher than the corresponding long form result. Differences were statistically significant for four of the nine subtests. Routine use of the short form was not recommended, because of the relatively high rates of classification error, extreme score deviation, and profile dissimilarity. (GDC)