There is growing interest in signed networks across disciplines, yet statistical methods for analyzing such networks remain limited. Signed networks extend standard network analysis by distinguishing between positive and negative ties and by modeling their interdependence over time.
This survey focuses on two major approaches for analyzing dynamic signed networks: exponential random graph models (ERGMs) and stochastic actor-oriented models (SAOMs). ERGMs treat networks as realizations from probability distributions defined by structural network statistics, whereas SAOMs model network evolution as a sequence of actor-driven tie changes in continuous time. Both frameworks have recently been extended to signed relations but differ in their assumptions about network change and the representation of network dependence.
The paper reviews the main concepts, strengths, and limitations of these approaches and discusses their connections to structural balance theory. An empirical application to a dynamic network from the Syrian Civil War illustrates how these models can be used to study the coevolution of cooperation and conflict among armed groups.