This study investigated the validity of two- and four-subtest combinations as estimates of WAIS-R Full Scale IQ among clients of low IQ (N = 100). Differences between the short-form means were small and nonsignificant. The correlation between Full Scale IQ and two-subset estimates was .86 and was .92 for the four-subtest estimate. These account for 74% and 85% of the variance, respectively. The two-subtest form correctly classified 69% of the subjects by intelligence category, and the four-subtest form correctly classified 85%. It was concluded the four-subtest form is superior as a screening device when complete administration of the WAIS-R is not feasible.