Validity of Selection and Classification Procedures for Predicting Job Performance

摘要

This report reviews major validation studies and meta-analytical summaries to assess the effectiveness of selection and classification procedures for predicting job performance in military and civilian settings. Initially, the average job validity coefficient across all jobs was computed to be in the low .20s. Currently, when a more uniform collection of studies is considered, statistical artifacts have been corrected, and carefully developed job criteria are used, the average job validity increased to the low 60s. Findings support major validity generalization concepts; however, job complexity and also criterion dimensions within a job both moderate validity. Expanding the predictor space of ASVAB and criterion dimensions appears t offer promise of differential validity for assignment.