The research evidence for selecting students in higher education with and without regard to race and sex is examined to assist admissions officers in light of the 1978 Bakke decision, which seems to give schools the option to use race in admissions decisions. Three clusters of studies supporting the consideration of race-sex subgroups in admissions decisions are considered: (1) those showing no relationship, or perhaps a negative relationship between the traditional predictors and college grades for minority students; (2) those indicating the need for separate equations or cutoffs for each subgroup if traditional predictors are employed; and (3) those involving the utility of noncognitive or nontraditional variables in predicting minority student success. Proponents of using a single prediction equation or cutoff score suggest that this approach is most fair to all concerned. Proponents of using separate equations or cutoffs for each subgroup when traditional predictors are employed note that black males tend to be the least predictable race-sex subgroup and any general equation would discriminate most against them. One reason for considering noncognitive or nontraditional variables in predicting minority student success is to ensure having comparable information for minority and majority group college applicants. Some research has indicated that the typical minority applicant is not as sure what is being solicited and is less likely to know how to supply the information the college wants. It is concluded that strong consideration should be given to race-sex subgroups in admissions procedures. (SW)