Micro-Level Interpretation of Exponential Random Graph Models with Application to Estuary Networks

A comprehensive understanding of the relational processes that constitute policy networks requires both an analysis of exogenous (i.e. covariates) determinants of the link structures in the network and a study of how the relationships that form the network depend upon each other, the endogenous determinants of the network. The Exponential Random Graph Model (ERGM) is an increasingly popular method for the statistical analysis of networks that can be used for both of these inferential tasks. Conventional interpretation of ERGM results is conducted at the network level, such that effects are related to overall frequencies of network structures (e.g., the number of closed triangles in a network). This limits the utility of the ERGM because there is often interest, particularly in the social sciences, in network dynamics at the actor or relationship levels. We present a comprehensive framework for interpretation of the ERGM at these lower levels, which casts network formation as block-wise updating of a network. These blocks can represent, for example, each potential link, each dyad, and the out or in-going ties of each actor. We contrast this interpretive framework with the popular actor-oriented model of network dynamics. The alternative models we discuss and the interpretation methods we propose are illustrated on estuary network data introduced in Berardo and Scholz (2010).