Call Us: US - +1 845 478 5244 | UK - +44 20 7193 7850 | AUS - +61 2 8005 4826

‘‘From Violence to Voting: War and Political Participation in Uganda.’’

The idea that episodes of intergroup violence exacerbate existing divisions and erode a sense of common fate across groups within diverse societies has a long history in the literature on ethnic conflict (Kaufmann 1996; Posen 1993), but few studies systematically document these effects at the level of the individual.2 Violence, the argument goes, charges public discourse with antagonistic content, produces more rigid boundaries between ethnic groups (de Waal 2005), and can lead even moderate individuals to support exclusionary policies (Canetti-Nisim et al. 2009) and fear the physical proximity of members of the other group (Fearon and Laitin 2000).3 In their most extreme version, the arguments in this literature suggest that interethnic violence may harden ethnic boundaries almost irreversibly, making future cooperation between groups difficult, if not impossible.4 Little individual-level research has linked political violence to deteriorating intergroup attitudes and behaviors, and existing studies have focused largely on terrorist threats in developed countries such as the United States and Israel. Perceptions of terrorist threat have been associated with Jewish Israeli opposition to civil and political rights for Palestinians (Canetti-Nisim, Ariely, and Halperin, 2008), riskminimizing behaviors in post-9/11 New York (Huddy et al. 2002), as well as negative stereotyping and support for harsh immigration and security policies directed at Arabs in a national sample of Americans (Huddy et al. 2005).5 Security threats appear to fuel intolerance (Wang and Chang 2006) and a willingness to curtail civil liberties (Davis and Silver 2004). At least three mechanisms underpinning these findings have been articulated in the literature. First, security threats may elicit an affective response of fear and enmity toward the outgroup, which translates into heightened support for exclusionary policies (Canetti-Nisim, Ariely, and Halperin 2008; 2 Exceptions include Shayo and Zussman (2011), Hayes and McAllister (2001), Balcells (2012), and Bauer et al. (2011). 3 Two commonly cited examples of this process are the ossification of the Hutu-Tutsi divide due to cycles of ethnic violence and genocide in Rwanda and Burundi (Prunier, 1995) and the hardening of the previously amorphous division between ‘‘Arabs’’ and ‘‘Africans’’ in Darfur as a result of ethnic violence in the 1980s and the genocide in 2003 (de Waal, 2005). 4 This argument features prominently in academic and policy debates about the relative merits of power-sharing and partition as solutions for ethnic conflict (e.g., Kaufmann 1998; Kumar 1997). See Sambanis (2000) and Sambanis and Schulhofer-Wohl (2009) for a critical empirical assessment of the argument for partition. 5 For a classic work on the relationship between threat perception and increased intolerance, xenophobia, and ethnocentrism, see LeVine and Campbell (1972). 2 bernd beber, philip roessler, and alexandra scacco Jarymowicz and Bar-Tal 2006; Marcus 2000). Second, threats can incentivize individuals to support policies that appear to mitigate risk, including policies that impinge on the rights of minorities (Huddy et al., 2005, 2002). Third, exposure to threat may motivate anxious individuals to reduce uncertainty by supporting simple political solutions, which tend to target minorities and avoid more complex and nuanced strategies of accommodation (Bonanno and Jost, 2006). These existing studies argue that concerns about being the target of outgroup violence can prompt individuals to support harsh policies toward outgroup members. We build on and move beyond this literature in three ways. First, we focus on the effects of riot violence, not terrorist threats. Riots are a more common source of insecurity than terrorist attacks, at least in Africa, but their effects on political opinion are virtually unstudied. From 1990 to 2011, African countries with a population greater than one million experienced a total of 2,115 riot events killing at least 34,500 people, compared to 803 terrorist events killing at least 12,500 people, according to data collected by Salehyan et al. (2012).6 Second, we study the effects of violence in a developing country, where physical security is a central subject of political debate.7 Third, we study the effects of past exposure to actual episodes of intergroup violence as op