The use of certain variables in predicting the probability of persistence to graduation for students majoring in business or nonbusiness was examined at Western State College of Colorado. The study sample consisted of 131 students from the 1977-78 graduating class, of whom 52 were nonbusiness majors and 79 were business majors. Predictions of major field of study also were run separately for the 46 men and the 33 women in business. The 16 predictor variables, taken from the American College Testing Program Assessment student profile reports, included aptitude scores, choice of major, and interest inventory scores. It was found that differences between groups were not entirely accounted for by discriminant function equations and the study variables. It appears that there is an extraneous variable that causes or encourages men to major in business while preventing (borderline) women from majoring in business. Of the six nonbusiness students who were predicted as most closely resembling the business groups, but who completed nonbusiness majors, all were women. Of the 14 business students who were predicted as most closely resembling the nonbusiness group, but who completed business majors, 12 were men. The findings generally indicate that intellective and nonintellective sets of measures are useful in predicting majors at the time of graduation. Prediction is better for men and women in nonbusiness than it is for those in business. There is a need to know more about characteristics of persons who attend college and about the impacts different kinds of college environments have on different categories of students. (SW)