MultiLevel Network Modelling Group

What is Social Network Analysis (SNA)?

Social Network Research is broadly concerned with the way in which individuals are relationally tied to each other and what consequences these relational patterns have for the individuals, and the social groups comprising these individuals. In social network research a graph is a commonly used analytical tool for describing a network. In such a graph, the nodes represent the individuals in the network, and ties between individuals that are connected to each other are represented by lines. These may be undirected edges, or directed arcs. The latter include arrowheads to show the direction of the relationship. When the individuals are people, the ties may represent giving advice, receiving support, friendship, having sex, etc. Nodes may also represent different social units such as countries or boards of directors, with appropriately defined relations on these units. Since Moreno’s foundational work in the 1930s, in what was then called sociometry, networks have proved to be of great use in explaining, for example, how innovations and opinions spread through social interaction, what consequences differential embeddedness (different position within the network) have on the power and well-being of individuals, how substance abuse and behaviour co-evolve (Freeman, 2004; Borgatti et al., 2009). There has also been an explosion of interest in social networks in disciplines such as physics and biology.

What are Exponential Random Graph Models (ERGMs)?

Exponential family random graph models (ERGM) (Frank and Strauss, 1986) are statistical models for the ties in a network that not only take exogenously defined characteristics of individuals into consideration (such as gender, organisation size, gross-domestic product), but also recognise the complicated interdependencies between tie variables. ERGMs are derived out of principled assumptions for the dependencies between tie variables and simple dependency assumptions give rise to a collection of configurations that are themselves interpretable in terms of theories of how ties self-organise. Much work has gone into model development, inference and the development of software for fitting ERGMs - e.g. PNet and statnet. The application of ERGMs to investigate social network structure is becoming increasingly popular in the social sciences.

References:

Stephen P. Borgatti, Ajay Mehra, Daniel J. Brass, Giuseppe Labianca, 2009, Network Analysis in the Social Sciences, Science, 323, 892-895.

Ove Frank and David Strauss, 1986, Markov Graphs, Journal of the American Statistical Association, 81, 832–842.

Linton Freeman, 2004, The Development of Social Network Analysis: A Study in the Sociology of Science. Vancouver: Empirical Press.

Other resources:

International Network for Social Network Analysis (INSNA)

Social Networks (academic journal) http://www.elsevier.com/wps/find/journaldescription.cws_home/505596/description]

Tom Snijders’ network page

MultiLevel Network Modelling Group
Home Team SNA & ERGMs Multilevel Networks Activities Links Events CCSR	What is Social Network Analysis (SNA)? Social Network Research is broadly concerned with the way in which individuals are relationally tied to each other and what consequences these relational patterns have for the individuals, and the social groups comprising these individuals. In social network research a graph is a commonly used analytical tool for describing a network. In such a graph, the nodes represent the individuals in the network, and ties between individuals that are connected to each other are represented by lines. These may be undirected edges, or directed arcs. The latter include arrowheads to show the direction of the relationship. When the individuals are people, the ties may represent giving advice, receiving support, friendship, having sex, etc. Nodes may also represent different social units such as countries or boards of directors, with appropriately defined relations on these units. Since Moreno’s foundational work in the 1930s, in what was then called sociometry, networks have proved to be of great use in explaining, for example, how innovations and opinions spread through social interaction, what consequences differential embeddedness (different position within the network) have on the power and well-being of individuals, how substance abuse and behaviour co-evolve (Freeman, 2004; Borgatti et al., 2009). There has also been an explosion of interest in social networks in disciplines such as physics and biology. What are Exponential Random Graph Models (ERGMs)? Exponential family random graph models (ERGM) (Frank and Strauss, 1986) are statistical models for the ties in a network that not only take exogenously defined characteristics of individuals into consideration (such as gender, organisation size, gross-domestic product), but also recognise the complicated interdependencies between tie variables. ERGMs are derived out of principled assumptions for the dependencies between tie variables and simple dependency assumptions give rise to a collection of configurations that are themselves interpretable in terms of theories of how ties self-organise. Much work has gone into model development, inference and the development of software for fitting ERGMs - e.g. PNet and statnet. The application of ERGMs to investigate social network structure is becoming increasingly popular in the social sciences. References: Stephen P. Borgatti, Ajay Mehra, Daniel J. Brass, Giuseppe Labianca, 2009, Network Analysis in the Social Sciences, Science, 323, 892-895. Ove Frank and David Strauss, 1986, Markov Graphs, Journal of the American Statistical Association, 81, 832–842. Linton Freeman, 2004, The Development of Social Network Analysis: A Study in the Sociology of Science. Vancouver: Empirical Press. Other resources: International Network for Social Network Analysis (INSNA) Social Networks (academic journal) http://www.elsevier.com/wps/find/journaldescription.cws_home/505596/description] Tom Snijders’ network page