What are statistical models for social networks (and why do we need them)?
Speakers:
Bio: I joined the University of Manchester as a Lecturer in October 2019. Before that, I was a postdoctoral researcher at the Social Networks Lab at ETH Zürich between 2016-2019. I got my Sociology DPhil from Oxford in 2016, and earlier my MA in Sociology and Economics from the Corvinus University of Budapest (Hungary) in 2010. Throughout these career stages, I have been involved in exciting research projects linking sociology, social psychology, education research, statistics, and quantiative social network analysis.
The analysis of social networks helps us to learn about how different actors (such as people; companies; countries) in society are interrelated. A distinctive feature of network datasets is that their data points are not independent. For example: whether Alice is friends with Bob may depend on whether they are both friends with Cloe. This violates one of the key assumptions of “standard”; non-network statistical models. In this session; we explore what could go wrong if we applied standard models to network data. We then take a quick tour of statistical models that have been developed specifically for social networks over the past decades. A few examples will highlight how these methods may help us to understand and tackle complex social issues.