Interrupted Survey Design for the Out-District Effects of Pre-Election Violence

• In Event: Poster Session: Experiments in Political Behavior & Public Administration & Canadian Politics
  In Poster Session: Poster Session: Experiments in Political Behavior

Abstract

This study utilizes Afrobarometer surveys that were accidentally interrupted by election-related violence in order to investigate how political elites use fear of violence to mobilize their own supporters and suppress the opposition vote. While violence may have both mobilizing (Hafner-Burton et al., 2018, 2014) and demobilizing (Blaydes, 2010; Collier & Vicente, 2012; Rauschenbach & Paula, 2019) effects, how political parties cultivate these seemingly contrasting effects is relatively an unexplored area. I argue that viole nce creates suppressing effects on voter turnout in its immediate vicinity but may have differential effects across incumbent and opposition supporters when it is committed at a distance. As a result, violence may suppress opposition turnout even in districts where there is no violence. When security forces are the ones who commit the violence, their presence at polling stations or in the streets on election day will evoke different responses across different voter groups. Opposition supporters may perceive the presence of police- the perpetrators of violence against their co-partisans- as a source of concern for their own personal safety, whereas incumbent supporters may simply feel “safe” with their presence. I test these claims in a natural survey experiment where the election related violence happens to divide Afrobarometer surveys into two comparable (pre- and post-violence) groups and find support for my theoretical claims.

Author

• Tolgahan Dilgin, Princeton University