Methods for Social Network Analysis, TT 2009.

Course provider: Tom A.B. Snijders (tom.snijders "at" nuffield.ox.ac.uk; http://stat.gamma.rug.nl/snijders).

This course is part of the methodology courses offered at the Department of Politics and International Relations of the University of Oxford, and also available to students from other departments.
The course is part of the teaching within the Centre for Research Methods in the Social Sciences.

Overview

Social networks are structures of relations between social actors - e.g., friendship between individuals, collaboration between companies, trade between countries, etc. They are important both in themselves and for explaining a variety of actor-dependent characteristics (behavioural tendencies, attitudes, performance, etc.) Particularly interesting is the two-way influence between relational networks and individual actions and outcomes. The data structure of social networks, where the usual statistical assumption of independent cases is totally implausible, poses special requirements for data analysis.

This course presents a number of important data analysis methods for social networks.
The material presented at this website is too extensive for the course. Only a selection of this will be treated.
The course is about five main topics:

  1. Centrality and other positional characteristics of actors.
  2. Concepts of positional equivalence.
  3. The exponential random graph model (ERGM), a statistical model for single observations of networks.
  4. Stochastic actor-based models for network dynamics (i.e., for analysing longitudinal network data).
  5. Stochastic actor-based models for the simultaneous dynamics of networks and behaviour (which here is a term referring generally to changeable actor characteristics such as behavioural tendencies or performance).
The focus is on the last three topics. They are treated in the framework of statistical inference - which is a usual framework for most data analysis but less so for social network analysis.

The course includes computer classes for hands-on data analysis using the StOCNET and SIENA computer programs.
From these websites you can download the programs. You can also download the StOCNET manual and the the most recent version of the Siena 3.2 manual.

The course takes place in Trinity Term 2009 in weeks 2 and 4 (see below for meeting schedule).

All papers to which links are included in this page, are intended only for personal use by participants in this course.

Background: introductory literature

It should be noted that this course is not a general introduction to Social Network Analysis, but a specific introduction to statistical methods for Social Network Analysis. For a general introduction, have a look at

Meeting Schedule

Location: Manor Road Building.
Time schedule for TT 2009.

Lecture Notes

For printing the handouts from Acrobat reader, if it does not come out righ, choose the scaling option "Fit to printer margins".

  1. Centrality and positional characteristics.
  2. Positional equivalence.
  3. The exponential random graph model (ERGM).
  4. Stochastic actor-based models for network dynamics.
  5. Stochastic actor-based models for the simultaneous dynamics of networks and behaviour.
Reading list Background literature