ABSTRACT
It is difficult to imagine a topic that has been the focus of more empirical research than the prediction of student performance in college. Literally thousands of analyses involving millions of students have been conducted by ACT, Inc. and the College Board, using tests and high school grades as predictors of success at individual colleges and universities. The most common measure of success in these studies has been freshman grade point average, but a number of other criterion measures have also been used, including, but not limited to, grades in individual courses, four-year cumulative grade point average, graduation, and ratings by faculty. For the past three decades many of the studies of the predictive value of tests have compared the validities and prediction equations obtained for subgroups of students defined by gender, race/ethnicity and disability status.
