Studying network change in education with stochastic actor-oriented models

This chapter provides an intuitive introduction to Stochastic Actor-oriented Models (SAOMs) in the context of educational research. SAOMs are multivariate statistical models that can be used to analyse changes in a social network observed at discrete time points. The model allows inference to various social mechanisms that may have shaped the evolution of the network between its observed states. While SAOMs are frequently applied in educational research, they are complex and may be difficult to use and interpret without prior experience. This chapter is intended for researchers who wish to gain a basic understanding of the model from a theoretical and applied perspective. First, we discuss the theoretical background of the model with a brief overview of key relational mechanisms and the rationale behind SAOMs. Second, we provide a step-by-step guide to fitting simple models. Third, we present and interpret a model about the effects of alcohol consumption on friendship dynamics in a high-school classroom. Lastly, we provide pointers to further literature on various SAOM extensions and applications. We hope that readers will find this chapter a useful step in learning this increasingly popular network analysis method.