coordinators: Marijtje van Duijn, Christian Steglich en Tom Snijders

Social network analysis is an interdisciplinary field of research with a long history of input from sociology, anthropology, statistics, mathematics, information sciences, education, psychology, and other disciplines. The large interest in social networks can be understood in view of the important theoretical and intuitively appealing research questions connected with social networks and the challenging methodological problems associated with the collection and analysis of social network data.

The research group consists of methodologists/statisticians and carries out research in which development and elaboration of new statistical methodology, software implementation and dissemination through publications and workshops, and collaboration with substantive researchers in sociology are integrated.

The research is aimed at development of methods for the analysis of network data, focusing on methods of statistical inference for the analysis of network dynamics, and of the joint dynamics of networks and actor attributes ("behavior"), as well as on the analysis of cross-sectional network data. The models developed by the research group are implemented in publicly available computer programs (Stocnet, Siena, p2). (http://stat.gamma.rug.nl/stocnet/) 

From 2002 until 2006 Tom Snijders led a NWO research program ‘Dynamics of Networks and Behavior' in collaboration with researchers from ICS Utrecht and health researchers from the University of Maastricht to investigate process of social influence and partner selection, with applications in adolescent networks and smoking and drinking behavior, and also a methodological Ph.D. thesis by Michael Schweinberger. In 2006 a European research collaboration started, ‘Dynamics of Actors and Networks across Levels: Individuals, Groups, Organizations, and Social Settings', which is funded by ESF. Christian Steglich has the role of coordinating post-doc in both projects. (http://stat.gamma.rug.nl/siena.html) 

Various developments have taken place in the analysis of cross-sectional network data. The greatest flexibility in the representation of network dependencies currently is given by the so-called exponential random graph model (‘ERGM'), or p* model. Another model, which is useful especially to study effects of covariates on the existence of ties in a network, is the p2 model, developed by Bonne Zijlstra, Marijtje van Duijn en Tom Snijders.