Collaborative Research: Substantive and Methodological Innovations for the Study of Temporal Dynamics in International Conflict

这个奖项资助了一个项目，该项目从两个方面促进了我们对国际关系的理解。首先，从理论的角度，它从经验的角度阐述了学者们几十年来原则上理解的问题：历史、序列和对未来的看法等时间因素对于完全理解国际互动至关重要。研究人员的工作检验了许多关于冲突的核心理论，包括将民主化、经济相互依存和不确定性与战争联系起来的命题。他们的建议强调，我们对政治动态做出的推论必然取决于对模型假设的认可和满足，以及该模型与生成所研究数据的过程之间的密切对应。关于研究现象的模型假设(例如协变效应中的时间稳定性)或基本实质性事实(例如国际舞台上冲突的反复性质)都不是无害的；例如，对先前争端的不同方法处理意味着对这些争端对当前冲突的影响的根本不同理解。在这种背景下，这个项目研究了一系列相关的问题：1.冲突会导致未来的冲突吗？他们在这个问题上提出了两个假设，一个是从理性主义的角度出发，另一个是从植根于新现实主义和社会心理学的模型中得出的。他们使用创新技术测试这些相互竞争的理论，这些技术解释了过去两个国家之间的冲突(即重复发生的事件)对未来发生争端的可能性的影响。2.它们审查了所有现有的国际冲突定量研究中隐含的一种断言，即观察结果是可以交换的。研究人员断言，这一假设在理论和经验上都是不正确的。文化和外源性冲击显然会影响冲突的可能性。这种异质性是模型错误说明的一种形式，使得使用标准统计方法得出的结论令人怀疑。他们使用混合模型来解决不可测量的异质性，并表明这些模型如何密切符合我们对国际争端的理解，并对这些争端的原因提供更好的评估。3.最后，它们确定某些影响对冲突的影响是否随着时间的推移而产生不同的影响。许多理论工作表明，这些因素的影响随着时间的推移而不同，这些因素已被证明是促进或抑制国际冲突的。其实质性影响从根本上关系到国际关系中不同理论视角的有效性，并对国际舞台上的决策者具有重大的政策影响。传统的分析不能也不能允许这些因素对争端的影响随着时间的推移而改变。总而言之，调查员将进一步改进国际关系研究，更加明确地注意一些关于时间和冲突的一般理论预期与用于模拟和检验这些预期的统计工具之间的联系。

This award funds a project that contributes to our understanding of international relations in two broad ways. First, from a theoretical standpoint, it addresses empirically what scholars have understood in principle for decades: that temporal factors such as history, sequence, and perceptions of the future are critical to any complete understanding of international interactions. The investigators' work tests a number of central theories of conflict, including propositions linking democratization, economic interdependence, and uncertainty to war. Their proposal emphasizes that the inferences we draw about political dynamics are necessarily conditioned by both the recognition and satisfaction of one's model's assumptions, and on a close correspondence between that model and the process generating the data examined. Neither model assumptions (e.g. that of temporal stability in covariate effects) nor basic substantive facts about the phenomenon of study (such as the repeated nature of conflict in the international arena) are innocuous; different methodological treatments of prior disputes, for example, imply fundamentally different understandings of the effects of those disputes on current conflicts. In this vein, this project examines a series of related issues:1. Do conflicts beget future conflicts? They develop two hypotheses on this question, one from a rationalist perspective and the other from a model rooted in neorealism and social psychology. They test these competing theories, using innovative techniques that account for the influence of past conflicts between two states (that is, repeated events) on the likelihood of a future dispute. 2. They examine an assertion implicit in all existing quantitative studies of international conflict, that observations are exchangeable. The investigators assert that this assumption is incorrect on both theoretical and empirical grounds. Culture and exogenous shocks clearly affect the probability of conflict. Such heterogeneity is a form of model misspecification, making the conclusions drawn using standard statistical approaches suspect. They use mixture models to address unmeasured heterogeneity, and show how such models conform closely to our understanding of international disputes and provide superior assessments of the causes of those disputes. 3. Finally, they determine whether the effects of certain influences on conflict have varying effects over time. A host of theoretical work suggests that the affect of those factors, which have been shown to foster or inhibit international conflict, vary over time. The substantive implications bear fundamentally on the validity of different theoretical perspectives in international relations, and have substantial policy implications for decision makers in the international arena. Conventional analyses do not and cannot allow for the effects of such factors on disputes to change over time.In sum, the investigators will further improve the study of international relations by paying greater and more explicit attention to the connection between a number of general theoretical expectations regarding time and conflict and the kinds of statistical tools used to model and test those expectations.