The US government spends billions per year on democracy and governance assistance in foreign countries. Nevertheless, one in five elections in Africa since 1990 has been afflicted by significant levels of violence, which impedes citizens from freely voting for their preferred candidates. Identifying how and when coercive violence influences voters is critical to effectively reducing its impact on the quality of elections. This study uses the case of Zimbabwe to understand how citizens make decisions about politics when faced with the threat of violence. More generally, it seeks to understand how the emotional impact of the threat of violence affects perceptions of risks and participation in pro-democracy collective action. The results of this study can help improve efforts to monitor elections and encourage participation in democratic politics by US institutions such as the National Democratic Institute, Carter Center, and USAID. Specifically, it implies that efforts to monitor elections should be focused not only on where coercion is most likely to occur but also on where it is most likely to impact behavior. Second, interventions to prevent and mitigate violence should take into account the psychological resilience of voters.
This study combines a historical analysis of the relationship between violence and voting patterns in Zimbabwe, lab-in-the-field experiments that test for the causal effect of emotions on perceptions of political risks, and a field experiment to test whether emotions can be used to mobilize citizen action. The historical analysis uses constituency and individual level data to test whether violence has a bigger impact on the behavior of voters who are more physically and psychologically vulnerable to intimidation. The lab-in-the-field experiment, conducted through a household survey of opposition supporters in Zimbabwe, uses photos of facial expressions and recalled past experiences of emotions to induce anger and fear. It then measures how these induced emotions affect how participants perceive the likelihood that they personally will face repressive violence as well as their willingness to engage in a number of political actions. It induces emotions using both political and apolitical recall-based stimuli to assess the extent to which the context of the emotions matters. Last, the field experiment leverages a unique partnership with an African political party to test whether negative campaign messages are more effective in mobilizing action in riskier political environments.
