Short forms of the Wechsler Memory Scale Revised: Cross-validation and derivation of a two-subtest-form

Revised General Memory and Delayed Recall Indexes using the initial and delayed portions of only 3 subtests. The original work found nearly perfect estimates. The present study indicated excellent cross-validation for the original formulas with correlations of .99. Over 85% of the estimated scores were within 6 points of the actual score. Using estimates from only 2 subtests rather than 3, correlations were only. slightly lower with 85% of the estimated scores within 10 points of the actual scores. Age, education, diagnosis, and IQ had no effects on the accuracy of these equations. Although the initial 3 test formulas are robust, the results suggest that using only 2 subtests results in higher estimation errors, which can be used when less accurate estimates are acceptable (such as in group research).