In the early 90s, when a young economist named Michael Kremer finished his PhD, there had been a few economic studies based on randomized trials. But they were rare. In part because randomized trials ...