Association for Public Policy Analysis & Management

This study presents results from a five-year randomized trial in which a single cohort of more than two thousand high school ninth graders received a $12,000 performance-based college scholarship offer. The program increased some measures of college readiness during high school, but had no effect on overall college entry and shifted student enrollment to colleges with lower academic quality. This conflicts with the large quasi-experimental literature that has found consistently positive effects of similar aid programs on college outcomes. Additional analysis points to three explanations: First, with experiments and other small scale pilot programs, the frequency of communication and spillover effects are smaller than with scaled up programs. Consistent with this theory, plausibly exogenous increases in the number of communications students received about the experimental program led to larger treatment effects. Second, the performance requirements made it almost impossible for most students to receive scholarship funds. Third, the scholarship required students to attend college in-state and program administrators framed the scholarship as free community college, which apparently led students to shift from four- to two-year colleges, and to less competitive four-year colleges. While making college more affordable can increase efficiency and equity, some forms of “free” college and performance incentives may unintentionally undermine these objectives.