|
Abstract:
|
Randomized impact evaluations are increasingly conducted in international development to assess more rigorously which interventions best improve conditions in poor countries. Set against backdrops of diverse cultures, political instability, and difficult conditions, though, randomized studies face challenges that conspire to reduce carefully-calibrated selection methods and power calculations. This paper will examine how the Millennium Challenge Corporation, a United States foreign aid program focused on poverty reduction and economic growth, has overcome those challenges, examining randomized evaluations in Georgia, Benin, and Nicaragua. It will compare how their randomization procedures and power estimates were affected by shifting project conditions, and the mix of quantitative and methodological approaches used preserve statistical rigor to provide reliable evidence for analysis.
|
