Why do some elections in countries recovering from armed conflict promote violence while others remain peaceful? What do United Nations (UN) peace operations do to contain the risk of electoral violence? And how do their activities contribute to electoral peace and the transition from war to democracy? My dissertation addresses these questions. I argue that violent electoral conflict is puzzling because election winners and losers could avoid the costs associated with violence by negotiating a mutually beneficial power-sharing compromise. To explain why negotiations fail, I reject the assumption of completely rational actors and instead build on prospect theory: First, actors evaluate different courses of action in terms of gains and losses based on their private reference point. Second, actors are generally loss-averse and try to avoid a sure loss, even at the risk to incurring an even larger loss. I argue that election winners’ reference point is the value of holding political office. If the value of holding office is higher (e.g. political office comes with exclusive access to state resources), then election winners face a larger loss from negotiating a power-sharing compromise (e.g. foregoing exclusive access to state resources) and election losers face a larger loss when accepting electoral defeat (e.g. exclusion from state resources). To avoid these losses, election winners reject a negotiated compromise and election losers reject electoral defeat. By consequence, bargaining is more likely to break down into costly electoral violence. Using data on 622 national elections in all 85 conflict-affected countries from 1990 to 2012, the analysis supports my argument. UN peace operations recognize the risks of holding elections after armed conflict. Since the end of the Cold War, UN peace operations have engaged in an increasingly wide variety of activities to support peaceful and credible elections. For example, they prepare elections, encourage democratic reforms, strengthen politically independent media outlets and provide security for voters. Yet, variation in what UN peace operations do on the ground and their effectiveness in containing electoral violence remains under-studied. Using new data on the peacebuilding activities by 44 UN peace operations between 1990 an 2012, I show that the UN invests more resources in peacebuilding when threats of electoral violence loom larger. This result is important because it suggests that we underestimate rather than overestimate peacekeepers’ effectiveness due to this selection bias. I then investigate in which ways peacekeeping mitigates (or exacerbates) electoral violence. I argue that the mere deployment of a UN peace operation increases the risk of electoral violence because peacekeepers channel additional resources to elected governments (e.g. foreign aid and military support). Therefore, they unintentionally increase election losers’ losses associated with electoral defeat and election winners’ losses associated with a negotiated compromise. However, if UN peace operations engage in specific activities, they may contain electoral violence in two ways. First, UN peacekeepers can raise the costs of violence by providing security to voters, candidates and election workers and by counteracting violence-inciting propaganda in domestic media. Second, UN peacekeepers can decrease the value of holding office and the losses associated with power-sharing compromise and electoral defeat by strengthening democratic institutions that guarantee the inclusion of election losers in future political decision-making processes. The statistical analysis indeed shows that while deploying UN peacekeepers increases the risk of electoral violence, UN-led efforts to assist electoral security and counteract violence-inducing misinformation in domestic media contribute to peaceful elections. However, I find no significant effects of election and institution-building assistance. To explore the causal mechanisms underpinning the effects of UN peacekeepers’ activities, I rely on fieldwork-based evidence from the 2015 post-conflict elections in Côte d'Ivoire. Interviews with Ivorian political leaders corroborate the hypothesis about peacekeepers’ ability to make violence costlier. Case evidence also provides possible explanations for the null effect regarding election and institution-building assistance: UN peacekeepers may lack sufficient resources to effectively transform the electoral and institutional environment. Consequently, they cannot sooth election losers’ fear of marginalization upon defeat or incentivize election winners to accept a power-sharing compromise. More importantly, peacekeepers’ robust intervention in previous conflict dynamics may lead political competitors to doubt the impartiality of UN peace operations. These doubts decrease peacekeepers’ ability to build trust in elections and democratic institutions. Overall, this dissertation shows that existing studies might have been a bit too over-optimistic about the virtues of peacekeeping. Solely deploying peacekeepers can increase the risk of violent electoral competition. However, peacekeepers can mitigate these detrimental consequences of liberal peacebuilding and facilitate the war-to-democracy transition if they invest in the ‘right’ activities.

Celebrities are a common feature of many campaigns to fight global poverty. However, increasingly, academics, practitioners, and the public express doubts on their ability to promote public engagement. Expert messengers, such as activists or NGO workers, could pose an alternative to celebrities. To contribute to this debate, the thesis asks whether celebrities and experts can be used as effective messengers in global poverty campaigns aimed at engaging the British public. The work integrates insights from the literature on political engagement, theoretical models on persuasive communication from social psychology, and experimental social science research methods. Research on political engagement looks at individuals' attitudes and political behaviours towards a political issue. Effective campaigns are recognised as an important driver of both factors, with current research analysing the effectiveness of different communication strategies aimed at engaging the public. I propose an adapted version of a dual-pathway model of information processing which looks at the persuasive role of experts and celebrities in charity campaigns. I hypothesise that trust in experts and celebrities depends on the recipients' personal characteristics. Furthermore, I hypothesise that a celebrity endorsements of a message can act as heuristic cues to a temporary attitude and behaviour change. Expert endorsements can also work as heuristic cues, but they can also stimulate cognition around the campaign?s message. This promotes stronger changes in attitudes and behaviours. The thesis includes three results. First, I provide evidence of the effectiveness of expert messengers, who can engage both previously uninvolved members of the public and reinforce the engagement of previously involved ones. Secondly, I find only limited evidence on the celebrity endorsements hypothesis, with mixed results on their effectiveness at engaging uninvolved individuals. Thirdly, I find that individuals' characteristics drive their credibility perceptions of experts and celebrities, positing that more targeted messages could increase the campaigns' effectiveness.

Universally, humans have a strong need to feel valued and cared for by their close social relationships (Baumeister & Leary, 1995). The degree to which people can identify such sources of support is strongly correlated with positive emotional and physical health outcomes (Thoits, 2011; Taylor, 2011), as well as academic achievement (Song, Bong, Lee, & Kim, 2015; Wentzel, Russell & Baker, 2016). Yet, few researchers have robustly tested how supportive communication between students and their social networks can be stimulated when it is lacking. This thesis makes a contribution to the academic literature and education policy by developing and testing interventions that motivate, inform, and remind students and their immediate social networks about their learning. The thesis introduces a fresh approach to the design of social support interventions. Rather than introducing new ties or establishing formal mentoring relationships, students' existing relationships are enlisted to provide support. Students' friends and family are, after individual randomization, sent a series of weekly text messages over the full academic year. These messages contain actionable and relevant information about the student's course, inspired by recent information interventions in education (Kraft & Rogers, 2015; Chande, 2017). The results indicate that informing study supporters of students' learning improves student attendance and attainment in maths and English qualifications. The follow-up trial shows that communicating with study supporters and students simultaneously is more effective than communicating with study supporters only. Qualitative evidence provides new insights for the design and implementation of supportive information interventions. Additionally, this thesis provides novel qualitative evidence that support from parents and friends helps students overcome challenges through cognitive appraisal processes. This data therefore offers new support for the popular hypothesis in the published literature (e.g. Feeney & Collins, 2015) that improved coping with emotions is a primary mechanism of social support on psychological well-being and academic success.

No description available.

No description available.

No description available.

No description available.

Virtually all governments are multilevel and where change in power distribution has occurred in recent decades, it has been to increase decentralization. However, not all scholars are sure that decentralization has always been beneficial for the quality of government. In addition, it has reduced the relevance of national government and made regions more important for the daily life of citizens as they became responsible for matters ranging from education to health policies to environmental protection. This has changed how governments are structured, the importance of national and sub-national elections and the nature of party competition. The mixed outcomes of decentralization for government quality and its large impact on the life of citizens highlight the importance of studying the causes and consequences of decentralization; this dissertation adds in this direction. This research looks closely at each key element of the multilevel structure, namely parties, regions and national-level factors, and the micro and macro-level relations that link them. Chapter 2 presents the prospective argument that parties in government decentralize towards their sub-national strongholds when they are at risk of future electoral defeat. This argument is tested by applying multilevel models to a sample of 15 European countries and their 141 regions. Chapter 3 shows that decentralization destabilizes the sub-national party system because it increases the risk of decline to new parties. I test this argument using survival models and applying them to a sample of 1,235 regional branches of political parties in 12 European countries. Chapter 3 analyzes the relationship between decentralization and the quality of elections in 67 countries. I demonstrate that it improves the fairness of national elections because it reduces the benefits of committing fraud. This dissertation represents important contributions to the literature on decentralization, multilevel governments, asymmetric federalism, party competition and electoral integrity for several reasons: a) it provides a comprehensive explanation for heterogeneous decentralization that is not conditioned to the presence of regional and ethnic parties; b) it proposes an inclusive measure for the risk of electoral defeat; c) it provides insights into the elements that affect the regional party system; and d) it shows that parties engage in cost-benefit calculations before deciding to commit electoral fraud.

Political candidates rely on news media to communicate with voters. Existing scholarship shows that racial and gendered patterns of campaign news coverage are unfavourable to minorities and women seeking elected office. Yet, the intersectional effects of race and gender have rarely been considered in this context. Recent elections have seen stark rises in the racial diversity of female candidates for the UK House of Commons and US House of Representatives. Responding to these developments, this thesis asks: What are the intersectional effects of race and gender on news coverage of political campaigns by minority women? Employing an intersectional theoretical framework, I formulate hypotheses regarding the effects of candidates’ racial and gendered identity on the amount, overall tone and content of campaign coverage they receive. Collectively, these hypotheses anticipate that most aspects of coverage will be least favourable for minority women, compared to similar candidates from other intersectional groups. A quantitative and qualitative content analysis is performed on local US, and national US and British newspaper coverage of matched samples of minority female, minority male, white female and white male candidates. The matching strategy and a series of explanatory models control for additional campaign, candidate and media factors which may affect coverage outcomes. The results show that minority women occupy a paradoxical position of hypervisibility and invisibility in the national press: a few individuals are singled out for exceptional attention while most candidates from this group struggle to receive recognition. In the local press, minority women receive less coverage than comparable white women, and less positive coverage than comparable candidates from all other groups. However, several of the hypotheses are unsupported: there little evidence of variation in the amount of viability or issue coverage candidates receive, and although stark differences emerge between the explicit and latent foregrounding of candidates’ race and/or gender, many of the relevant news frames are surprisingly positive. The qualitative analysis does however, show continued scepticism and hostility to the progressive measures which are necessary for minority women’s descriptive representation. More broadly, I argue that by considering only the effects of a single axis of identity, research on women or minorities in politics may run the risk of making claims that obscure the experiences of all but the most privileged within each group.

No description available.