Statistical methods for the analysis of social network data
Do high school students make their friends start smoking? Is a bully a central actor in the classroom? Is the ‘old boy network’ of chief executive officers just a myth, or can social network analysis help to prove its existence through interlocking directorates? ‘It is our job to develop statistical methods, models and software necessary for sociological research’, says Marijtje van Duijn, leading the small group of researchers of statistical methods for social network analysis, statisticians and mathematicians.
Social network analysis.
Social network analysis has a prominent place in sociology. Advances in computational technology have facilitated methodological development. ‘A growing number of researchers discover the added value of social network analysis for answering their research questions’, Van Duijn says.
The group’s statistical models are applied in the analysis of social networks in school classrooms, for example in the research on bullying carried out in the department. Other examples are studies on social relations between employees in a company, or relations between organizations.
The statisticians are available for consultation and collaboration with their colleagues on the application of statistical methods. Sometimes simple, sometimes advanced methods are required, or even to be developed. ‘It is our role to accommodate the research in the department’, says Van Duijn, ‘statistics is a great auxiliary discipline for the social sciences’.
Groningen is renowned for its contribution to statistical methods for social network analysis. One of the pioneers, professor Tom Snijders, is the active founder of the group. The stochastic actor oriented model for changes in social networks and individual behavior, developed by Snijders and colleagues, is used by researchers all over the world. It is implemented in the software RSiena. Other areas of expertise of the group include missing data handling and multilevel analysis.
|Last modified:||17 April 2020 2.18 p.m.|