The population of 674 first-year student-athletes culled from 5 successive freshman classes (2004-2008) at a mid-size midwestern university was examined to determine what combination of demographic, academic, and athletic variables best predicted retention into the 2nd academic year. The dependent variable of retention was chosen because it is a primary component of the Academic Progress Rate, a semester by semester evaluation of team academic performance used by the NCAA. Pearson correlations revealed student-athletes were less likely to drop out if they were Caucasian, attended college close to their hometown, scored well on standardized tests, had a respectable high school GPA, were ranked high in their graduating high school class, were not a member of a revenue sport, and earned a considerable amount of playing time. Binary logistic regression analysis demonstrated that predicting retention is possible with information about race, distance from hometown, type of sport, and amount of playing time.