Can Referral Improve Targeting? Evidence from a Vocational Training Experiment

We seek to improve the targeting of agricultural extension training by inviting past trainees to select future trainees from a candidate pool. Some referees are rewarded or incentivized. Training increases the adoption of recommended practices and improves performance on average, but not all trainees adopt. Referred trainees are 3.7% more likely to adopt and randomly selected trainees, but rewarding or incentivizing referees does not improve referral quality. When referees receive financial compensation, average adoption increases and referee and referred are more likely to coordinate their adoption behavior. Additional adopters induced by incentivizing referral adopt imperfectly and incur losses from adoption; they also tend to abandon the new practices in the following year.