Reposted by Emilia Tjernstrom
The central issue is fundamental. The treatment was described as randomized at the village level. It was not. Participants instead retained their treatment assignments from two earlier studies run by the same NGO. Those earlier treatments themselves had large/persistent effects.
