| Sumario: | This article presents results of the first large-scale test of Schmidt and Hunter's Bayesian validity generalization procedure. This procedure was applied to 56 distributions of validity coefficients drawn from 698 published and unpublished studies representing five clerical job families, 10 test types, and two classes of criteria-job proficiency and training success. Results showed that most of the between-study variation in empirical validity results was accounted for by four statistical artifacts, thus casting serious doubt on the traditional belief that employment test validities are situationally specific. it was also found that in most cases generalization of validity to similar clerical jobs or new settings was justified, even where the hypothesis of situational specificity could not be rejected with certainty. Further, validity generalization could be supported based on corrections for sampling error alone. The correlation between mean test type validities for proficiency and training criteria was found to be high, indicating that contrary to previous belief, similar ability measures are predictive of both criterion types. Implications of these findings are discussed in terms of both practical applications and theory development in industrial-organizational psychology.
|
|---|