ABSTRACT
Research on predicting college performance focuses on the relationship of a criterion measure with either (a) tests, such as the SAT, or (b) any other predictor. There are many more studies of the former than the latter. Correlations between the SAT Verbal and Math composite with firstyear college grade point average (FGPA) range from .44 to .62 (Hezlett et al., 2001). Studies show that the SAT accounts for over 30% of the variance in FGPA and an increment of about .07 to .1 over high school grade point average (HSGPA; Bridgeman, McCamley-Jenkins, & Ervin, 2000; Ramist, Lewis, & McCamley-Jenkins, 1993). Similarly, median multiple correlations between FGPA and the four ACT scores were .43 across 129 recent institutional validity studies (Maxey, 2001). Predictive validity coefficients of tests for graduate and professional are generally .3 to .4, and often exceed the validity of undergraduate GPA (Zwick, 2002). A recent meta-analysis of the predictive validity of the SAT I included approximately 3,000 studies, involving over 1 million students. The study found that this test is a valid predictor of performance early in college, with multiple correlations of SAT I Verbal and Math composite with FGPA ranging from .44 to .62 (Hezlett et al., 2001). This same study found that the SAT I is also a valid predictor of academic performance later in college
(e.g., graduation, cumulative GPA) with multiple correlations ranging from the mid-30s to the mid-40s.
