Using data from the 2006 cohort of the Wabash National Study of Liberal Arts Education, we examined the relationships between three approaches to measuring student learning outcomes (direct-assessment learning gains, self-reported gains, and college grades) and student persistence from the first to second year. Results from a series of logistic regressions indicated that students' grade-point averages had the largest explanatory power in student persistence, followed by self-reported gains. Direct-assessment learning gains had the least power in explaining persistence. The findings have implications for the national conversation on student success in college.