We bring together 40 randomized and non-randomized evaluations of education programs to compare cost-effectiveness, seeking to facilitate prioritization of different candidate interventions by policymakers. We examine cost-effectiveness across three outcomes (enrollment, attendance, and test scores) and find distinct "best interventions" for each outcome. For increasing enrollment, urban fellowships, school consolidation, and extra teachers have proven most cost effective. For school attendance, school-based deworming stands out as most cost effective. And for improving test scores, several interventions seem similarly cost effective, including providing blackboards, workbooks, training teachers, and others. We discuss some of the challenges inherent to comparing interventions.
_______________________________________________________________________
Acknowledgments:We are grateful for financial support from the RAND Corporation. We especially thank Lynn Karoly for helpful comments on the paper and Rachel Glennerster for helpful conversations, as well as the authors of many of the cited papers for additional cost information not available in the published versions of their papers. All errors are our own.