Nudging Voters? Experimental Evidence on the Effects of Election Polls

• In Event: Perceptions on Representatives and Legislative Compromise

Abstract

This paper distinguishes two ways in which published polls influence electoral preferences. On the one hand, there is an electability effect, which updates citizens’ declared support based on the published popularity of the candidates in the race. On the other hand, electoral polls produced a priming effect, which induces respondents to mentally rehearsing arguments for or against the candidates. We unpack these effects by implementing a survey experiment during the 2018 presidential election in Mexico. This design allows us to test the claimed effects in a real-world situation and to pull apart other campaign factors. The findings show that informing respondents about previous poll results does not sway their support among candidates. Rather, the information influences citizens’ behavior in two ways. Among those who declared having an electoral preference, the information increases respondents’ confidence on their choice regardless the candidates’ position in the race. Among those respondents initially undecided, in contrast, the information makes them more likely to declare their support for a candidate later in the survey. The paper talks to the growing concern of governments, scholars, and general public about the effects of publishing pre-election poll results. 

Authors

• Francisco Cantu, University of Houston

• Javier Márquez, Buendia & Laredo