Date of Award

Summer 8-15-2019

Author's School

Graduate School of Arts and Sciences

Author's Department

Political Science

Degree Name

Doctor of Liberal Arts (DLA)

Degree Type



This dissertation explores the behavioral determinants and legacies of conflict from a socio-psychological perspective. The three papers that compose this dissertation investigate use tools of causal inference to explore, on the one hand, the socialization processes that influence leaders' decision to pursue war-like policies and, on the other hand, the longue durée legacy of wartime violence on preferences, as well as on people's proneness to being swayed by messages of collective threat during conflict. The first paper of the dissertation examines the long-term determinants of interstate conflict by exploring the impact of leader's socialization processes. Based on the soft power theory of international experiences and the impressionable-years socialization hypothesis, I theorize that leaders with the experience of attending a university in a Western democratic country should be less likely than non-Western-educated leaders to initiate militarized interstate disputes. I test this proposition by employing a new dataset, building on Archigos and LEAD, that includes background attributes of more than 900 leaders from 147 non-Western countries between 1947 and 2001. The results support the hypothesis, even when accounting for leader selection, time-variant country and leader-level controls, other leaders' background characteristics, and country and year fixed effects. This finding lends credence to the soft power thesis of institutions on international sojourners, and highlights the value of considering leaders' background and experiences in conflict research. The second paper of the dissertation shifts the focus from the long-term behavioral causes of wartime violence to its consequences. Specifically, this chapter studies the long-term legacies of wartime violence by exploring its impact on civic engagement. My argument is that exposure to wartime violence transforms people's psychological makeup in a way that increases their long-term civic engagement. I test this proposition using evidence from the Vietnam War by exploiting the distance to the arbitrarily drawn border at the seventeenth parallel as an instrument for conflict intensity. The results show that individuals who lived during the war in a province that was heavily affected by the conflict are more likely to be engaged in civic organizations in 2001, 26 years after the end of the conflict. I further find that this effect occurs because war exposure increases people's participatory values, not because of postconflict development. The third chapter of the dissertation investigates the past-present nexus in the reignition of conflicts in divided societies. To begin with, this chapter offers the first estimate of the effect of priming collective/group threat on actual electoral outcomes. This is done through a region-wide field experiment embedded in an organization's leafleting campaign in the 2017 Catalan regional elections in Spain. Unlike most field experiments that randomize at the level of individuals, randomization takes places at the level of the precinct, equivalent to a small village. The cluster randomization allows me to evaluate the effect of the treatment on actual voting behavior and, thus, alleviates concerns related to using self-reported behaviors. The results show that in precincts assigned to receive a group threat prime, the vote share of Catalan pro-secessionist political parties increased by between 1% and 1.7% compared to the no-increase of precincts that received a vote encouragement with either no prime of group threat or no prime at all. Building on research suggesting that historical conflict strongly influences the formation of preferences across generations, I propose a resonance theory of persuasion by which communities more severely affected by historical political violence should be more likely to be influenced by primes of collective/group threat. To test this hypothesis, I combine the field experiment with a novel dataset of local-level historical wartime violence and state repression in Spain based on the geographic location and size of about 2,650 mass graves throughout Spain. In line with the resonance hypothesis, the results show that primes of collective threat are particularly effective in those communities that were more heavily affected by the Spanish Civil War and the ensuing repression, with an increase by between 3.6% and 5.8% in the support for Catalan pro-secessionist parties---compared to the increase of between 1% and 1.7% of precincts with an average exposure to the historical conflict. These findings have far-reaching implications for our understanding of how conflict history combined with rhetoric of group threat enhances secessionist mobilization.


English (en)

Chair and Committee

James L. Gibson

Committee Members

Brian Crisp, Raymond Duch, Guillermo Rosas, Margit Tavits,


Permanent URL: https://doi.org/10.7936/x62t-da20