EQUATING INVERSE PROBABILITIES IN IMPLICIT PERSONALITY JUDGMENTS

Meehl and Rosen (1955) observed that their colleagues tended to equate the sensitivity of a sign (its probability given a diagnostic categor y) with its predictive accuracy (the probability of the diagnostic cat egory given the sign). Such equating of inverse probabilities implies ignoring base rates. Later, many researchers demonstrated such neglect in a number of experimental contexts involving hypothetical judgments ; others have found some use of base rates-but only when base rates we re supplied by the experimenter, in which case even irrelevant base ra tes influenced judgment. The present study tested for equating conditi onal probabilities in a context similar to the contexts involved in th e initial observations-that is, subjects strove for accuracy rather th an attempted to make good judgments about hypothetical problems, and t he base rates were generated by the subjects themselves. Neglect of ba se rates was revealed by within-individual comparisons showing subject s equated inverse probabilities without equating the corresponding unc onditional ones.