Can Public Servant Performance Be Increased? Experimental Evidence on Efforts to Improve Teaching in India

In a cluster-randomized trial, two treatment arms promoted activity-based instruction by providing teaching materials and teacher training. One of these arms also promoted community engagement through community-led student contests. A third arm remained untreated. After 13 months, the version without contests improved teaching quality and learning among girls. Both versions improved student attitudes towards math. Yet, the addition of contests—which are intended to put pressure on teachers to increase their students’ performance—worsened instructional quality (especially classroom culture), and we can rule out that the contests added even small improvements in learning.