Using data from two rounds of surveys on students in the Washington State Achievers (WSA) program, this study examined the relationship between student engagement in college activities and student persistence in college. Different approaches using student engagement measures in the persistence models were compared. The results indicated that the relationship between student engagement and the probability of persisting was not linear. Even though a higher level of social engagement was related to an increased probability of persisting, a higher level of academic engagement was negatively related to such probability. The findings have strong implications for educational research, policy, and practice.