The Rajasthan Police implemented four reforms in a randomly selected group of 162 police stations: (1) weekly duty rosters with a guaranteed rotating day off per week; (2) a freeze on transfers of police staff between stations; (3) inâ€service training to update skills; and (4) placing community observers in police stations.
These reforms were subjected to a scientific evaluation, using data collected through two rounds of surveying including police interviews, decoy visits to police stations, and a large scale crime survey—the first of its scale in India. The results show that two of the interventions, the freeze on transfers and the training, do show potential to improve the public image of the police.
The other reforms showed no robust effects, an outcome that may be due to their incomplete implementation.
That from a paper Esther Duflo presented at Yale this week, written with coauthors Banerjee, Keniston, and Singh.
Next step: randomize vigilantes?