Regression Discontinuity in Prospective Evaluations

Abstract

While regression discontinuity has usually been applied retrospectively to secondary data, it is even more attractive when applied prospectively. In a prospective design, data collection can be focused on cases near the discontinuity, thereby improving internal validity and substantially increasing precision. Furthermore, such prospective application of regression discontinuity will often allow better understanding of the selection process, further increasing precision. As a result, regression discontinuity’s scope for application is considerably broadened in the prospective context. We demonstrate these ideas with an impact evaluation of the USDA Food and Nutrition Service’s Fresh Fruit and Vegetable Program (FFVP).