The problem of this study is to refine the "policy-capturing" model used in the selection process of interns for the Administrative Leadership Training Program in a large metropolitan school district. The subjects for this study consisted of 416 candidates over a three-year period. The statistical procedure of multiple linear regression analysis was used to test the hypothesis that it would be possible to model the decision-making process so that the predictive value would be 90 per cent or higher. During the refinement process, the unique contribution of variance accounted for by each predictor variable was examined,and interactions between certain variables were tested. Two refined models were formulatedand the predictive value of each was calculated. The predictive values of all the models were less than 90 per cent; therefore, the hypothesis was rejected.