IQ SUBTEST ANALYSIS Clinical Acumen or Clinical Illusion ? Marley

Subtest analysis is pervasive in psychological training and practice. That is, the speculation that the variability or profile of an individual's scaled scores across the subtests of an intelligence test have meaning beyond that provided by global IQ measures. A review of subtest analysis research revealed that neither subtest scatter nor subtest profiles demonstrate acceptable accuracy in discriminating among diagnostic groups. The evidence that exists regarding relations between subtest profiles and socially important academic and psychosocial outcomes is, at best, weak: subtest profile information contributes 2% to 8% variance beyond general ability to the prediction of achievement and 2% to 3% to the prediction of learning behaviors and test-session behaviors. Hypothesized relationships between subtest profiles and other psychosocial behaviors persistently fail to achieve statistical or clinical significance. Methodological problems in research and practice that cause subtest analysis results to be more illusory than real and to represent more of a shared professional myth than clinically astute detective work are explicated.