Doctor of Philosophy (PhD)


Political Science

Document Type



By distinguishing the mode of warfare, the focus of this dissertation is to examine the spatial and temporal changes in the mode of warfare, from terrorism, or guerrilla warfare, to conventional warfare and suggests that each type of warfare represents how insurgents effectively challenge incumbents across space and over time. In so doing, this dissertation places a particular emphasis on combatants’ relative capacity, since insurgents’ strategies are mainly driven more by calculations regarding their strength and weakness relative to government forces. The results of the analyses are mixed in terms of the theoretical expectations of the dissertation. Nonetheless, there were indeed spatial and temporal variations of the mode of warfare during the Vietnam War. By employing Firth’s logistic analyses, I first find that (1) political and physical geography, (2) popular mobilization, and (3) external support are associated with a specific mode of warfare at the sub-national level. In locations in which the presence and control of government forces are strongest, and thus maximum asymmetry of relative capacity between insurgents and incumbents occurs, insurgents favor terrorism. When rebels’ capacity changes from maximum asymmetry to asymmetry with popular support and control of territory, they are likely to choose guerrilla attacks. When rebels’ capacity is strengthened further by outside support—approaching the relative capacity to parity—they resort to conventional warfare. Second, the type of warfare evolves as rebels’ relative capacity accumulates strength over time, from terrorism to guerrilla warfare to conventional warfare. Third, the findings underscore the importance of the relationship between previous conflict outcomes and the dynamics of warfare: The success of strategies and iii battle wins can endogenously reinforce rebels’ relative capacity, thus prompting changes in the type of warfare; conversely, battle losses undermine rebels’ relative capacity, thus adversely causing changes in rebels’ strategies. Besides theoretical claims, this dissertation highlights the need for a multi-method approach to better explain conflict dynamics. I combine large-N statistical analyses with semi-controlled data across and within case analyses and suggest that the two methods together offset each method’s limits, thus contributing to more valid causal inferences in conflict studies. Overall, this dissertation demonstrates that analyzing different conflict patterns at the micro level lends valuable insight to the nature of conflict process.



Committee Chair

Clare, Joseph

Available for download on Thursday, November 01, 2029