The continued controversy concerning the predictive validity of the Graduate Record Examination (GRE) aptitude test and its influence on selection decisions, including admission and financial aid, has necessitated the establishment of local norms. The sample involved 407 University of Kansas education and computer science students. Information on GRE verbal (GRE-V), GRE quantitative (GRE-Q), GRE verbal and quantitative, undergraduate grade point average (UGPA), graduate grade point average (GGPA), major field of study, sex, and year of enrollment were recorded. GGPA was selected as a criterion variable. The remaining variables were treated as predictors. Stepwise multiple regression was applied as a statistical tool to analyze the data. Data analysis revealed that the verbal score on GRE was the best single predictor of GGPA for education students. Based on zero-order correlations, computer science students' UGPA was a better predictor than GRE scores. The findings agree with previous studies that the GRE-V is the best single predictor of GGPA for majors that are descriptive in nature and GRE-Q is the best predictor for symbol-oriented disciplines. Tests for the equality of regression equations developed for the two student groups found significant differences, suggesting separate selection procedures in the two departments. (Author/PN)