Doctoral Dissertation Research: Novel Graph Models for Studying Mobilization

长期以来，历史学家、社会学家和政治学家一直在关注民众的集体政治行动问题。英国废除奴隶贸易的运动（1788 - 1807年）也许是第一个这样的例子。在研究这个问题的社会科学学科中，现在有一个共识，即社会关系的模式对动员水平有影响。本文研究了这一案例，使用档案数据来模拟个人和定居点的社交网络，通过这些网络，运动得以组织和传播。新的图模型将被用来估计这些结构，并调查哪些功能有助于运动的成功。这项研究提供的见解将有助于为这一时期的历史叙述和运动本身提供信息，此外，这项研究将有助于更全面地了解网络结构的类型，使或限制在其他社会和历史背景下的集体动员。这项研究介绍了一种创新的方法-指数随机图建模（ERGM）的自我中心抽样数据，作为一种解决方案，将基于数据和模拟的实证研究更紧密地联系在一起。这些模型是为了研究传染病的传播而开发的，因此它们很适合研究英国的废奴问题，因为：（1）很容易看到（虽然不是等价的）疾病的传播和思想之间的关系;（2）传染病建模者经常面临的难以找到的当代人口的数据问题与历史社会学家面临的问题惊人地相似。曼彻斯特案例研究的主要源文件，然后嵌入在英格兰的全国范围内，为这些模型提供数据：（1）回答为什么废除死刑的动员在这些不同规模上取得成功的问题，（2）提供一个框架，以研究类似历史背景下的其他动员事件，（3）带来ERGM，一种新的分析工具，这项工作结合了高度技术性的网络分析和对数字档案数据集的细致汇编。这些数据源在历史研究中尚未得到广泛利用，因此这项研究将为历史、社会和计算机科学的跨学科学者开展类似工作提供一个模板。 此外，这项工作将指出社会运动成功的一些决定因素。该项目产生的数据将有助于研究人员继续研究方法技术，理论建设和研究这个特殊的案例。

The problem of popular, collectively organized political action has long had a grip on the attention of historians, sociologists and political scientists. The British campaign for the abolition of the slave trade (1788-1807) is perhaps the first of such cases. Across social science disciplines that engage with this problem, there is now a consensus that the patterning of social relations has a bearing on the level of mobilization. This dissertation examines this case, using archival data to model the social networks of individuals and settlements through which the movement was organized and diffused. Novel graph models will be used to estimate these structures and investigate which features contributed to movement success. The insights that this research provides will help inform the historical narrative of the period and the movement itself, and, moreover, this research will contribute to a more complete understanding of the type of network structures that enable or constrain collective mobilization in other social and historical contexts.This research introduces an innovative methodology - exponential random graph modeling (ERGM) for ego-centrically sampled data - as a solution to move empirical investigations grounded in data and simulation closer. These models have been developed to study the spread of infectious diseases, and thus they lend themselves well for the study of British abolition because (1) it is easy to see the parallel (though not equivalence) between the diffusion of diseases and ideas; and (2) the data problems of hard to find contemporary populations that often confront infectious disease modelers are strikingly similar to those that historical sociologists face. Primary source documents for a case study of Manchester which then is embedded at the national scale in England provide the data for these models to: (1) answer the question of why abolitionist mobilization was successful at these different scales, (2) provide a framework to study other mobilization events in similar historical con- texts, and (3) bring ERGMs, a new analytic tool, into the mainstream sociological research on mobilization.This work is a combination of highly technical network analysis and the meticulous compilation of datasets from digital archives. These data sources are not yet widely exploited in historical research, and thus this research will provide a template for conducting similar work by interdisciplinary scholars from history, the social and computer sciences. In addition, this work will point to some of the determinants of social movement success. The data produced by this project will be useful to researchers for continued investigation of methodological techniques, theory building, and the study of this particular case.