Collaborative Research: Mapping Relational Data in Civil Conflicts

近几十年来，国际体系中有一半以上的国家经历了内战，越来越多的国家面临暴力内部冲突。自1946年以来，大多数国内冲突至少有三个活跃的武装组织，有些甚至有多达十个武装组织，但大多数分析只研究了政府与假定的同质反对派组织之间的相互作用。在测量和建模技术方面的研究不足，无法充分捕捉对了解国内冲突所需的数据生成过程至关重要的网络和关系特征。该项目通过将参与者和战斗概念化为一组节点和纽带，构成了一个网络，揭示了国内冲突的相互依存动态。该项目将生成一个用户友好的统计软件包和网站，称为冲突网，这将使研究人员能够更容易地概念化和研究内战。越来越多的文献表明，控制关系数据中的相互依赖性产生较少的偏差参数估计，并使我们能够更好地理解冲突将如何演变。冲突网将对该领域做出宝贵贡献，并使对当前政治事件和国内冲突感兴趣的学者能够从两个关键方面进行分析。首先，从建模的角度来看，创建一个能够轻松生成次国家级冲突网络数据的工具，将提高对国内冲突驱动因素进行推断的能力，并且使用网络视角对武装行为体之间的冲突进行操作，将提高预测未来冲突发生的准确性。其次，该项目不仅解决了文献中的关键空白，而且为国内冲突研究的新方向铺平了道路：在细粒度分析单元中预测战斗级事件；关于暴力事件如何取决于群体特征、动机和社会政治联系的涨落的新假设；更深入地了解武装行为体之间的相互依赖如何加剧了针对平民的暴力，并给政府行为体带来了新的挑战。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

More than half of the countries in the international system have experienced a civil war in recent decades and a growing number of countries face violent internal conflicts. The majority of civil conflicts since 1946 have had at least three active armed groups and some have had up to ten groups, yet most analyses only examine interactions between the government and an assumed homogeneous opposition group. Research has fallen short in measuring and modeling techniques that adequately capture network and relational characteristics essential to the data-generating process needed for understanding civil conflict. This project unmasks the interdependent dynamics of civil conflict by conceptualizing actors and battles as a set of nodes and ties that constitute a network. The project will generate a user-friendly statistical package and website, called conflictNet, that will allow researchers to more easily conceptualize and study civil war. A growing body of literature demonstrates that controlling for interdependence in relational data yields less biased parameter estimates and allows us to better understand how conflict will evolve. ConflictNet will be a valuable contribution to the field and enable scholars interested in the analysis of current political events and civil conflict in two keys ways. First, from a modeling perspective, the creation of a tool that allows for the easy generation of network data on subnational-level conflict will improve the ability to conduct inference on the drivers of civil conflict, and operationalizing conflict between armed actors using a network perspective will allow for greater accuracy in forecasting future occurrences of conflict. Second, this project not only addresses critical gaps in the literature but also paves the way for new directions in intrastate conflict research: the prediction of battle-level incidents at fine-grained unit of analyses; new hypotheses on how violent events are conditional on the ebb and flow of group level characteristics, motivations, and socio-political connections; and a deeper understanding of how interdependence across armed actors escalates violence against civilians and poses new challenges for government actors.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.