A Digital Field Experiment Reveals Large Effects of Friend-to-Friend Texting on Voter Turnout

Abstract

Two decades of field experiments on get-out-the-vote tactics suggest that impersonal tactics, like mass emails, have only a modest or negligible effect on voter turnout, while more personal tactics, like door-to-door canvassing, are more effective. However, the COVID-19 pandemic threatens to upend the vast face-to-face voter mobilization efforts that have figured prominently in recent presidential election campaigns. If campaigns can no longer send canvassers to voters' doors, what tactics can they turn to in order to mobilize their supporters? This paper evaluates a promising alternative to face-to-face get-out-the-vote tactics: mobile app technology that enables millions of people to message their friends to urge them to vote. Prior to the most recent US midterm elections in 2018, the mobile app Outvote randomized an aspect of their system, hoping to unobtrusively assess the causal effect of their users' messages on voter turnout. We develop a statistical methodology to address the challenges of such data, and then analyze the Outvote study. Our analysis reveals evidence of very large and statistically significant treatment effects from friend-to-friend mobilization efforts ($\widehat{\textrm{CACE}}$= 8.3, $\textrm{CI}$ = (1.2, 15.3)). Further, the statistical methodology can be used to study other friend-to-friend messaging efforts. These results suggest that friend-to-friend texting, which is a personal voter mobilization effort that does not require face-to-face contact, is an effective alternative to conventional voter mobilization tactics.