Why do people vote? We design a field experiment to estimate a model of voting “because others will ask”. The expectation of being asked motivates turnout if individuals derive pride from telling others that they voted, or feel shame from admitting that they did not vote, provided that lying is costly. In a door-to-door survey about election turnout, we experimentally vary (1) the informational content and use of a flyer pre-announcing the survey, (2) the duration and payment for the survey, and (3) the incentives to lie about past voting. The experimental results indicate significant social image concerns. For the 2010 Congressional election, we estimate a value of voting “to tell others” of about $$$15, contributing 2 percentage points to turnout. Finally, we evaluate a get-out-the-vote intervention in which we tell potential voters that we will ask if they voted.

Policymakers often consider interventions at the scale of the population, or some other large scale. One of the sources of information about the potential effects of such interventions is experimental studies conducted at a significantly smaller scale. A common occurrence is for the treatment effects detected in these small-scale studies to diminish substantially in size when applied at the larger scale that is of interest to policymakers. This paper provides an overview of the main reasons for a breakdown in scalability. Understanding the principal mechanisms represents a first step toward formulating countermeasures that promote scalability.