Bureaucratic Interest in Academic Evidence: A Field Experiment in Peru, India and Tanzania
Abstract
We investigate how different types of messaging by academics impact public servants’ behavioral demand for access to academic evidence by inviting nearly 9,000 national-level government bureaucrats in Tanzania, India and Peru to receive training on an original website that summarizes and graphically depicts findings from hundreds of randomized controlled trials. Implementing well-known interventions from the psychology of choice literature, we sought to test whether “nudges” might work across different national and bureaucratic contexts. We tested randomly assigned messaging characteristics related to messenger characteristics and social proof. In particular, we randomized whether invitations came from U.S. researchers on location or similar local researchers. We find that bureaucrats were significantly more responsive to U.S. researchers in Peru but not in India or Tanzania. We also randomly assigned information that other bureaucrats in the same ministry had previously expressed interest. Contrary to findings from other populations with similar “social proof” treatments, bureaucrats were less interested when informed that others had already accepted the demonstration. This novel finding is likely due to the institutional context of bureaucracies, where knowledge ist more likely to be either a shared resource or a competitive advantage. In this context, others’ access to knowledge disincentivizes individual action rather than communicating a norm. This study offers evidence about some possible approaches and pitfalls to effective dissemination of academic evidence.
