WISC-R subtest reliability over time: implications for practice and research

This study investigated the test-retest reliabilities of the WISC-R IQs and scaled scores on subtests over 3- and 6-yr. intervals for a learning disabled sample. Statistical procedures and individual analyses were conducted on a sample of 107 students who had at least two evaluations. For the total sample, 3-yr. test-retest reliabilities of the Verbal IQ, Performance IQ, and Full Scale IQ were .69, .72, and .77, respectively. Subtest reliabilities ranged from .33 to .71, with a median of .49, but nine subtests had reliabilities of .62 or lower. Individual analyses indicated considerable variability, with some students decreasing as much as 33 IQ points on retest. Similar results were noted for a subsample of 41 students who had three evaluations over a 6-yr. period. Problems with subtest interpretation due to the inconsistency of the WISC-R scores and extensions of this research to the WISC-III are discussed.