Do individuals in conflict settings misperceive the extent to which their ingroup supports peace? If so, what drives these misperceptions, and can correcting them build further support for peace? To answer these questions, I conduct survey experiments in South Sudan in the midst of ongoing conflict. I document large misperceptions of ingroup pro-peace attitudes, with the average respondent substantially underestimating support for peace within their group. I also show that underestimation is greatest among individuals who have less access to information. However, providing information about rates of pro-peace ingroup attitudes does not cause individuals to update their perceptions of ingroup attitudes, and it therefore has no impact on individuals' own attitudes. Exploring mechanisms, I provide evidence that the most likely reason individuals do not update their perceptions of ingroup attitudes is that they do not believe the information they are provided. The findings suggest that it is common for individuals in conflict settings to underestimate the extent to which their ingroup supports peace, but correcting these misperceptions is not straightforward. More generally, the results underscore the difficulty of building pro-peace attitudes in conflict-affected settings through information provision.

Civic education is widely used to reduce support for political violence, but more evidence is needed to evaluate its effectiveness. I conducted a field experiment in South Sudan to assess a typical adult civic education intervention designed to reduce support for violence. Implemented amid ongoing civil conflict just a few months prior to South Sudan's first national elections, the intervention convened village residents to listen to an educational radio drama and participate in facilitated discussions. Contrary to pre-registered hypotheses, I find that it increased support for violently rejecting unfavorable election results. I show that the most likely mechanism was a change in beliefs about norms related to rejecting elections. I also show that the group discussion element most likely drove the shifts in beliefs. To explain these results, I present a theory that clarifies how discussion can shape beliefs about norms and demonstrates that the impact of discussion is highly unpredictable. This research emphasizes the prominent role of discussion in shaping the outcomes of civic education and the risks associated with group discussion in these interventions. It also has important implications for the many other interventions studied by political scientists that create opportunities for group discussion.

Can pro-peace messages from elites reduce violent conflict? I explore this question by estimating the impact of the Pope's pro-peace messages during his visit to South Sudan. Leveraging the fact that access to radio was a major determinant of exposure to the Pope's messages, I employ difference-in-differences designs that exploit variation in radio coverage throughout the country. I estimate that the messages caused a 58 percent reduction in fatalities in the three months following the visit. I argue that these results are driven by shifts in beliefs that increased the costs of fighting among militias, and I provide supporting evidence from interviews with militias and leaders in conflict-affected villages. The findings have implications for the potential to reduce violence with messaging, the ability for civilians to influence conflict dynamics, and the role of religious actors in creating peace.

In this chapter we provide an overview of the economics and political science literature on clientelism. We argue that clientelism – characterized by the direct exchange of a citizen’s vote in return for payment or access to other goods and services – undermines electoral accountability and helps explain rent-seeking and the under provision of public goods across the world. We begin with a brief overview of the prevalence of clientelism across countries and over time, based on expert surveys and public opinion data. Next, we discuss the role of poverty as an underlying factor that facilitates clientelism and also argue that clientelism may further reproduce poverty. We then move to study agency problems in clientelistic transactions and discuss agency problems in the broker-voter and politician-broker relationships. We present evidence consistent with the importance of monitoring and targeting (i.e. selection) to overcome agency problems and highlight how networks and social connections help facilitate these strategies. Next, we discuss how clientelistic machines are funded, highlighting how government funds and social programs are often instrumental sources of clientelistic exchanges. We then discuss evidence on how to weaken clientelism and promote the adoption of more programmatic policies. We end with a discussion of open questions and avenues for future work.

Civilians can support the state in countering violent extremism by providing information about extremist activity. However, there is often a lack of civilian-state cooperation. This problem is especially acute among women. Although women often have access to information about extremist activity, they may be especially reluctant to cooperate with the state due to perceptions that security institutions are dominated by men and hostile to women. With this research, we are conducting a field experiment in four counties in coastal Kenya to examine whether fostering connections between women civil society leaders and women in formal security institutions can encourage greater civilian-state cooperation in countering extremism. We hypothesize that building connections can increase cooperation among civilians by expanding knowledge about how to cooperate, strengthening trust, and reducing the perception that security institutions are male-dominated. We have secured funding for this project and are in the process of implementing the experiment. 

Secession and efforts to secede are often associated with political violence. Yet recent surveys reveal that a surprising number of Americans support their state breaking away from the union. What drives this support? We examine the role of four factors: (1) the imposition of objectionable policies by non-copartisans, (2) fiscal concerns, (3) fears that secession could provoke violence, and (4) reluctance to abandon copartisans. We will test these hypotheses through an online survey experiment.

Can adult civic education increase women’s political participation and support for their participation in patriarchal settings?  How does group discussion between men and women influence outcomes? We investigate these questions through a field experiment of a civic and education intervention in South Sudan, just months before the first national elections. We leverage (1) random assignment of the intervention to villages and (2) random assignment of participants within treatment villages to either mixed- or same-gender discussion groups.