Advanced Courses on Social Network Analysis: 21 – 25 June 2010
Two parallel courses:
- A) Advanced methods for one mode, two mode and egonetworks
- B) Statistical analysis of social networks
Duration: 5 days (9.30am — 5pm), 21 – 25 June 2010.
Level: Intermediate
Course Fee: £1000 (£750 for academic, £500 for students)
Course Leaders
- Martin Everett, Nick Crossley, Elisa Bellotti
- Mark Tranmer, Johan Koskinen
Course A - Advanced methods for one mode, two mode and egonetworks
Martin Everett, Nick Crossley, Elisa Bellotti (University of Manchester)
This course assumes a basic knowledge of social network analysis and familiarity with the software package UCINET. We will cover advanced topics in centrality (Eg Bonacich power beta centrality), and cohesive subgroups (Eg advanced secondary analysis and techniques such as markov clustering) together with methods for blockmodelling using both structural and regular equivalence. We shall examine particular data types, e.g. valued data and two mode data, consider issues such as missing data and cover more advanced topics for ego networks, including structural holes and brokerage roles. In addition we shall look at the matrix algebra routine and show how users can use UCINET in a more sophisticated way to run analysis and data manipulations that are not in the standard menu. The following is indicative and depends on participants’ interests and backgrounds.
The course will
- Develop the users skill in the software package UCINET, with some insights into Keyplayer, Mage and Pajek.
- Explain the implementation and theory of blockmodelling and positional analysis techniques
- Describe some advanced network concepts involving centrality, subgroups and other relevant areas
- Explain methods and techniques for handling 2-mode, valued and missing data.
- Develop the participants knowledge and understanding for methods and techniques for ego networks.
- Develop the participants knowledge and skills for mixing methods in social network research, particularly how to combine qualitative methods with network data.
Preliminary reading
- Stanley Wasserman, Katherine Faust (1994), Social Network Analysis: Methods and Applications, Cambridge University Press, Cambridge.
- David Knoke, Song Yang (2008) Social Network Analysis (2ND edition), Sage, London.
- Robert Hanneman and Mark Riddle (2005) Introduction to social network methods
Course B - Statistical analysis of social networks
Mark Tranmer (University of Manchester), Johan Koskinen (University of Oxford)
21 June 2010
An introduction to R
One day
This course is aimed at people who wish to familiarise themselves with R. It serves as a pre-requisite for the statistical analysis of social networks course, but taken as a one-day course, it is also suitable for anyone interested in using R more generally. R is a command language that can be used to carry out standard statistical analyses but also has powerful facilities to enable users to create their own routines or implement methods designed by other researchers.
The course will:
- Introduce participants to the R environment
- Explain how to enter data and run simple descriptive statistical methods
- Describe how to run standard procedures
- Show how to run commands designed by other researchers and how to develop commands for non-standard analyses.
22 June -25 June 2010
Statistical analysis of social networks
4 days
This course assumes basic statistical knowledge such as regression and familiarity with R. Participants not familiar with R take the 1 day course on 21 June 2010, which will prepare them for this course. Anyone who already knows R takes Day 1 of the A. course on 21 June 2010. The first two days will look at ERGMs (Exponential Random Graph Models) using the software package PNET and also statnet in R. This allows us to answer questions such as: Are there more triads in my network than I would expect by chance? And more complex questions involving attributes such as am I more likely to be friends with someone who is a similar age to me? The last two days are devoted to the examination of longitudinal data using the R version of the SIENA package. This looks at network formation over time and is an actor based model that allows for endogenous network effects (such as transitivity and popularity) as well actor attributes (such as homophily) to be included in the model. A brief review of standard regression models, such as logistic regression, will be given during the first part of the course.
The course will
- Introduce the theory and terminology of the Exponential Random Graph Model (ERGM) and show how it can be applied to network data using PNET and statnet and discuss issues such as convergence, degeneracy and goodness of fit.
- Extend the ERGM to deal with attribute data and show how the model can be used in practice.
- Describe the actor based model implemented in RSiena
- Show how the model can be extended by using a variety of practical examples with an emphasis on interpretation of the output.
Preliminary reading
- Robins, G L, Pattison, P E, Kalish, Y, Lusher, D (2007) An introduction to exponential random graph (p * ) models for social networks Social Networks. 29:173-191.
- Snijders, T.A.B., Doreian, P. (2010). Introduction to the special issue on network dynamics. Social Networks, 32, 1-3.
- Snijders, T.A.B., Steglich, C.E.G., and van de Bunt, G.G. (2010). Introduction to actor-based models for network dynamics. Social Networks, 32, 44-60.
Further reading
- Robins G, Pattison P, Wang, P (2009) Closure, connectivity and degree distributions: Exponential random graph (p*) models for directed social networks. Social Networks 31:105-117.
- Robins G, Snijders T, Wang P, Handcock M, Pattison P (2007) Recent developments in exponential random graph (p*) models for social networks a Social Networks 29 192–215