Abstract

Although an extensive experimental literature has tested a wide array of voter mobilization tactics, billboard advertisements have seldom been evaluated, and studies to date have been limited to a small number of sites. This essay reports results from a nationwide experiment conducted during the 2020 general election in the United States. Experimental sites were ethnically diverse locations in metro areas, including both presidential battlegrounds as well as places with no closely contested races. A total of 298 billboards were randomly assigned to treatment or control in 155 geographic clusters. Exposure to billboards by residential location is modelled using cell phone usage patterns. Turnout is measured using public records for residents living at various distances from randomly assigned billboards. Using a variety of estimation approaches, we obtain point estimates that are close to zero, with hints of stronger effects among those who reside near treated billboards. On the whole, it appears that signage does little to raise turnout in high-salience elections.