Watched: how monitoring affects performance during a survey of development practitioners

We ask whether monitoring time during a survey of development practitioners from Latin America and the Caribbean affects performance. Our proxy for monitoring is a digital clock, randomly assigned to appear on the screen of half of the 357 online surveys. Monitoring reduces the average survey time by 23 per cent, lowers the ratio of unanswered questions by over 60 per cent and has no negative impact on other measures of survey quality. These results illustrate how monitoring alone can be a cost-effective tool for improving performance.