SUBGROUP VALIDATION BY SEX: CLERICAL APTITUDE AND PERFORMANCE.

Theoretical and methodological studies relating to differential validation suggest that the phenomenon seldom exists. In an experimental study, the Minnesota Clerical Test was predictively validated for 41 Ss (27 female, and 14 male). While the correlation in the total sample equaled .56, moderated regression and subgroup validation both increased predictive accuracy when sex was used as a grouping factor. The results are discussed in light of desired outcomes, EEOC guidelines and court tests, and statistical properties of prediction equation. It is suggested that subgroup validation should be considered wherever mean differences between major groups are found for the predictor variable. [ABSTRACT FROM AUTHOR]