Doctoral Dissertation Research: The Strength of Long Ties

40年前，社会学家马克·格拉诺维特在《弱关系的力量》一书中做出了开创性的工作。40年前，许多研究证实了他违反直觉的观点，即人们更有可能从跨越网络距离较远的熟人那里获得新信息，而不是从亲密的朋友那里。尽管桥梁纽带对扩散的重要性已经有了很好的证明，但重点主要是比较嵌入在紧密联系的圈子中的纽带的关系强度与连接长距离网络的纽带的关系强度，而不是检查跨越不同网络距离的桥梁纽带的纽带强度。利用推特和其他社会/通信网络的人口规模数据，该项目将(A)检验桥梁联系往往随着网络距离的增加而变弱这一命题，(B)利用统计和模拟方法解释桥梁联系的强度和跨度长度之间的关系，以及(C)进一步探讨不同文化和经济发展的国家之间这些关系的差异。这项研究将有助于跨社会科学的研究，包括对扩散和传染、网络进化、社会资本和劳动力市场过程的研究。开发和传播有效的方法来收集、分析和比较由数亿人组成的多个网络，将有助于主要依赖基于调查的自我中心网络数据的社会科学社区。该项目还将促进关于使用分布式和并行云计算工具研究人口规模的社会网络的研究生培训

40 years since the pioneering work of sociologist, Mark Granovetter, on the "strength of weak ties," numerous studies have confirmed his counter-intuitive thesis that individuals are more likely to acquire novel information from an acquaintance who bridges long network distances than from a close friend. Although the importance of bridge ties for diffusion has been well documented, the focus has primarily been on comparing the relational strength of ties embedded in tightly knit circles vs. the relational strength of ties that bridge long network distances, rather than examining tie strengths of bridge ties that span varying network distances. Drawing on population-scale data from Twitter and other social/communication networks, this project will (a) test the proposition that bridging ties tend to be weaker with the network distance they span, (b) develop an explanation for the relationship between the strength and span-length of bridging ties using statistical and simulation approaches, and (c) further explore the variation in these relationships across countries of different cultures and economic development. This study will contribute to research across the social sciences, including studies of diffusion and contagion, network evolution, social capital, and labor market processes. The development and dissemination of effective methods to collect, analyze, and compare multiple networks constituted by hundreds of millions of individuals will contribute to the social science community that primarily relied on survey-based egocentric network data. The project will also contribute to graduate training on the use of distributed and parallel cloud-computing tools for studying population-scale social networks