Limits to Experimental Evaluation and Student Attendance in Rural Indian Schools: A Field Experiment

Diane Coffey, Princeton University

This paper reports the results of a field experiment in schools for the children of migrant workers in rural north India. The experiment tested whether the “megaphone program," an intervention using social psychology and focused on children's agency, would improve student attendance. While standard impact evaluation econometrics initially suggests the program may have been weakly effective, operational data reveals that the program probably had no impact on student attendance. Further results illustrate challenges for development policy and caveats about experimental evaluation. Teacher attendance, confirming previous findings, is a first-order problem for rural education. Impacts measured among capable implementers may have little validity when applied to less capable organizations. More attention should be given to the question of how to attract organizations that are dedicated to improving their effectiveness to the service of deprived populations.