Predicting job choices with models that contain subjective probability judgments: An empirical comparison of five models

Abstract

Five mathematical models whose components were subjective probability judgments were used to predict the job choices of persons seeking professional employment in the public schools. The models differed in their algebraic forms and in the types of subjective probability judgments that they included. Some models were heuristic and some were Bayesian. The subjects differed in that some were experienced in public school work and some were not. Several hypotheses were tested concerning the effects on model validity of (1) algebraic form, (2) type of probability judgments, and (3) degree of previous experience in public school work. A Bayesian model using likelihood ratios derived from posterior-odds appeared to be the best predictor of job choices, but all models were good predictors.