Models for the analysis of social networks with multiple sources of variation

Bonne Zijlstra PhD

Project: University of Groningen

Project financed by: The Netherlands Organization for Scientific Research (NWO)

Supervisors: Prof. dr T.A.B. Snijders, dr M.A.J. van Duijn

Project running from: 1 September 2001 – 22 May 2006

Summary:
For the analysis of social network data, models are proposed that take into account variation coming from individuals, dyads (i.e. pairs of individuals), groups of individuals, as well as multiplex relations and multiple questions pertaining to one relation. These models extend an existing model for the analysis of a single network which accomodates variation connected to the individuals and the dyads in the network.

Several estimation techniques will be investigated, and the most suitable will be implemented in user-friendly software. The models will be used in the analysis of data on friendship networks from several national and international research projects.

Date of defence: 11 September 2008

Title of thesis: Random effects models for directed graphs with covariates

ISBN: 978-90-367-3497-4

ISBN: 978-90-367-3498-1 (electronic version)