The study explores the ways in which school size influences two important student outcomes commonly used in school effects research: growth in mathematics achievement and dropout rate. Past research suggests that smaller high schools can lead to increased benefits for students. In this study, multilevel analytic models of the first two waves of the Educational Longitudinal Study (ELS) of 2002 are used to better understand mediating factors in the small school size effect. Results support the benefits of smaller schools: students in smaller schools were less likely to drop out than their counterparts in large high schools. The relationship between high school size and math gain, however, is curvilinear; students attending very small (less than 674) or very large (greater than 2592) schools have the largest math gains, while students in schools with intermediate enrollment sizes demonstrated smaller math gains. Implications for school size decision making are discussed in light of these results. (Contains 3 tables and 2 figures.)