A LOGISTIC-REGRESSION MODEL FOR PERSONNEL-SELECTION

A two-parameter logistic regression model for personnel selection is proposed. In addition to presenting a theoretical basis for the model, a uni fied approach is provided for studying selection, validity generalization, employee classification, selection bias, and utility-based fair selection. The new model was tested with a large database (N = 84,808). Results show the logistic regression model to be valid and also quite robust with respect to direct and indirect range restriction on the predictor. © 1991, Sage Publications. All rights reserved.