Does electoral violence affect voting choice and willingness to vote? Evidence from a vignette experiment

Many new democracies experience electoral violence. Though this form of political violence is common, there is little understanding of how violence affects vote choice and turnout. This article draws on a vignette experiment that is embedded in a nationally representative survey in Kenya, where electoral violence has occurred several times since the 1990s. We show that voters strongly sanction candidates who are rumored to have used violence, even if the candidate is a coethnic, a copartisan, or has performed well in office in the past. This sanctioning effect, however, is not consistent across voters. Victims of past electoral violence and those in poverty are less likely to sanction candidates that use violence. Rumored use of violence also depresses turnout, even among a violent candidate’s core constituents, when voters do not possess countervailing information about the violent candidate’s past performance in office.