“Diamonds in the Rough”: Using Community Information to Target Grants to High Value-Added Students

Merit-based educational grants and scholarships are typically awarded on the basis of measures of academic performance (i.e. grades, test scores, etc.), even though these measures are not designed to predict the impact of aid. Grants administered like this overlook talented students who (perhaps rationally) underperform on these measures because they do not expect to reap the rewards of the investment or because they are unaware of the returns to scholastic achievement. Though scholarship-granting institutions generally do not have the local knowledge required to identify these “diamonds in the rough”, community members (such as peers, teachers, and counselors) may be in a better position to assess who will benefit most from an educational grant. This project will test the value of using community information to target educational grants. The project will have three parts. First, we will use an incentive compatible mecha-nism to elicit peer, teacher, and counselor rankings of “at risk” students on the basis of predicted value-added from an educational grant. Second, we will assign a ranking of these students purely on the basis of school performance as measured by grades and test scores. Finally, we will randomly distribute grants to some portion of these students and evaluate which measure better predicts gains in educational performance, graduation, college attendance, and other outcomes.