Stochastic Actor-Oriented Models for the Co-Evolution of Networks and Behavior: An Introduction and Tutorial

Abstract

Stochastic actor-oriented (SAO) models are a family of models for network dynamics that enable researchers to test multiple, often competing explanations for network change and estimate the extent and relative power of various influences on network evolution. SAO models for the co-evolution of network ties and actor behavior, the most comprehensive category of SAO models, examine how networks and actor attributes—their behavior, performance, or attitudes—influence each other over time. While these models have been widely used in the social sciences, and particularly in educational settings, their use in organizational scholarship has been extremely limited. This paper provides a layperson introduction to SAO models for the co-evolution of networks and behavior and the types of research questions they can address. The models and their underpinnings are explained in nonmathematical terms, and theoretical explanations are supported by a concrete, detailed example that includes step-by-step model building and hypothesis testing, alongside an R script.

Multilevel Methods and Statistics: The Next Frontier

“…The field of network science is progressing at an incredibly fast pace, with many of the new developments focused on multilevel and temporal factors (Lazega & Snijders, 2016). There are currently techniques that allow for examining the multilevel and temporal factors involved with dyadic ties in networks (e.g., Krivitsky & Handcock, 2014) and changes to network structure (e.g., Kalish, 2020; Snijders, Lomi, & Torló, 2013; Snijders & Steglich, 2015), identifying communities (e.g., Fortunato & Hric, 2016) and important influencers (e.g., Lü et al, 2016) in networks, considering networks involving multimode (i.e., human and nonhuman factors) nodes (Contractor, Monge, & Leonardi, 2011), and a wide variety of other methodological approaches and tools. As is apparent from the above discussion of future research needs associated with bottom-up effects and processes, these and other network methodological developments have great potential to shed light on a number of important multilevel research questions and address remaining method-oriented challenges.…”

The Power, Accuracy, and Precision of the Relational Event Model

“…REM departs from a typical assumption of other network statistical frameworks, such as ERGMs or SAOMs, for which the issue of dependence between observations is critical (Kalish, 2020;Lusher et al, 2013). In REM, each event is considered to be conditionally independent of all other events in the sequence.…”

The turnover contagion process: An integrative review of theoretical and empirical research

“…Finally, researchers may adopt stochastic actor‐oriented (SAO) models to test ideas regarding how coworker DoM and EoM cues impact employee DoM and EoM evaluations (see Kalish, 2020). SAO models allow researchers to examine whether coworkers (alters) and an employee (ego) develop similar levels of DoM (e.g., job satisfaction) and EoM (e.g., job search behavior) over time (i.e., the covariate‐similarity effect), which would provide evidence of contagion.…”