A new computer-based approach for teaching counselors about predictive validity and how to use tests in clinical decision making was developed and tested. The approach is based on concepts similar to those used in inferential statistics, particularly statistical hypothesis testing. To implement this model, the counselor must: (1) determine the criterion variable and the rule for inclusion in the diagnostic category; (2) consider the consequences of misclassification; (3) set the maximum probability of misclassification allowable for any client; (4) compute a cutoff value on the predictor variable; (5) compare the client's score on the predictor variable to the cutoff value and make a decision; and, optionally, (6) determine long-run probabilities of true positives, false positives, true negatives, and false negatives that would result if the cutoff score was used over time with all clients. A computer program in PASCAL was written to teach this method. Students' opinions about the program were obtained from 26 counseling students. Students liked the program and found it helpful. Those who perceived themselves as relatively more able in statistics received the program more favorably than did those who saw themselves as less able. Five figures illustrate the study, and an appendix gives examples of student reactions. (SLD)