The predictive validity of the Graduate Record Examinations (GRE) aptitude test has been investigated for foreign students at the University of Kansas. A total of 148 students whose first language was other than English were selected for this study. Verbal and quantitative scores on the GRE, field of study, sex, and year of initial enrollment were used as predictors. The grade point average (GPA) in the graduate school was considered as the criterion. The results revealed that foreign students scored significantly lower than American students on the GRE scores and the GPA. Poor correlation between the GRE scores and the criterion suggested that the GRE aptitude test is not the most appropriate way to predict academic performance of foreign students. During multiple regression, field of study and verbal scores on the GRE were found to be significant predictors of the GPA. High within-group variability among foreign students suggested the study of subgroups formed on the basis of language, culture, or the geographic location. The use of separate norms for foreign students was also found desirable. (Author/PN)