Social networks are dynamic by nature. Network dynamics is important for domains ranging from friendship networks (e.g., van Duijn et al., 2003; Burk et al., 2007) to, for example, interorganisational networks (Borgatti and Foster, 2003; Berardo and Scholz, 2010). Ties can be established and can be terminated; also there may be changes in the node set. Changes in ties may be considered the result of the structural positions of the actors within the network - e.g., when friends of friends become friends -, characteristics of the actors ('actor covariates'), characteristics of pairs of actors ('dyadic covariates'), and residual random influences representing unexplained influences. The study of network dynamics should shed light on the underlying theoretical micro-mechanisms that induce the evolution of social network structures on the macro-level.

This summer school treats stochastic actor-based models for network dynamics (Snijders, van de Bunt, and Steglich, 2010), which are a type of models that have the purpose to represent network dynamics on the basis of observed longitudinal data, and evaluate these according to the paradigm of statistical inference. This means that the models represent network dynamics as being driven by many different tendencies, such as the micro-mechanisms alluded to above, which may have been theoretically derived and/or empirically established in earlier research, and which may well operate simultaneously. Some examples of such tendencies are reciprocity, transitivity ('friends of my friends are my friends'), homophily (choice of network ties based on similarity of salient attributes), and assortative matching (choice of network ties based on similarity of network position). In this way, the models aim to give a good representation of the stochastic dependence between the creation, and possibly termination, of different network ties. These stochastic actor-based models allow to test hypotheses about these tendencies, and to estimate parameters expressing their strengths, while controlling for other tendencies (which in statistical terminology might be called 'confounders'). The actor-oriented nature means that changes in the network are modelled as representing from choices by the actors who are represented by the nodes in the network. This leads to a model combining agency and structure, and is well suited for expressing theories based on purposeful behaviour by social actors. With respect to data requirements, the focus will be on the analysis of network panel data, i.e., data where the dependent variables are constituted by a network, and possibly one or more individual variables, that have been observed repeatedly for a given group of actors. The number of repeated measurements varies in practice mostly from 2 to 10, but may be larger.

In the basic model (Snijders, 2001) the changing network is the dependent variable. Many important scientific questions, however, can be framed as questions about the mutually dependent dynamics of networks and attributes (behaviour, attitudes, performance, etc.) of the individual actors in the network. This here is called co-evolution of networks and behaviour, where the term 'behaviour' also stands for performance, attitudes, etc. The choice of network ties may depend on the attributes and network embeddedness of the actors and of their possible candidates towards whom to extend a tie; this is called social selection. Also, the behaviour of actors may depend on not only their own attributes but also on their network position and on the behaviour and other attributes of those to whom they are directly or indirectly tied in the network; this is called social influence. Models for the co-evolution of networks and behaviour allow the joint representation of social selection and social influence, as elaborated in Steglich, Snijders, and Pearson (2010).

Statistical methods for parameter estimation and hypothesis testing based on these models have been implemented in the R package RSiena (Ripley and Snijders, 2011). R is a general statistical software system (R Development Core Team, 2011), and the incorporation of these methods in an R package allows combining them with many other statistical methods. The summer school will teach the basic methods of actor-oriented modelling for network dynamics as well as their implementation in R.