This study compared the ability of the ACT Assessment Battery (ACT), the Scholastic Aptitude Test (SAT), and six other variables to predict first semester Grade Point Average (GPA) at the University of Illinois at Urbana-Champaign, a "highly selective midwestern university". Additional purposes were to determine: (1) the effect of using the short form SAT instead of the long form SAT in multiple regression equations, and (2) whether a difference exists in the motivation of the same students, taking the SAT in the spring as part of the regularly scheduled national SAT testing program and in the fall as part of a special experimental testing session designed for entering freshmen. This was examined by comparing the correlation of spring and fall SAT scores with a third spring score, School and College Ability Tests (SCAT). The 12 variables used in this study were SAT short form (verbal and quantitative), ACT (English, mathematics, and composite), SCAT (verbal, quantitative, and total), SAT long form (verbal, quantitative, and total) and high school percentile rank (HSPR). Results suggest that, in general, the ACT and SAT are equally able to predict first semester GPA either alone or in multiple prediction with HSPR. HSPR is, by far, the best single predictor of success. Adding more variables to the prediction equation increases the multiple regression only slightly. Finally, the short form SAT appeared to be as good a predictor of first semester GPA as the long form SAT. (Author/MV)