Incorporating Ensembles of Social Networks in Structural Equation Models: A Finite Mixture Approach

• In Event: Marriott Roundtable Session 29
  In Roundtable Session: Advances in Clustering and Classification

Abstract

Social network methodology has grown substantially in recent years. However, despite advances in analyzing single networks (e.g., Liu, Jin, & Zhang, 2018), little methodological work has tackled data from multiple, independent networks. As education researchers increasingly gain the capacity to collect data from ensembles of networks—e.g., peer-support networks within classrooms (Gest & Rodkin, 2011)—methods that accommodate within-network dependence while simultaneously estimating complex, multivariate structural effects are needed for flexible theory-based modeling. The current study utilizes a finite mixture approach to cluster networks and provide for estimation of predictors and distal outcomes within a structural equation modeling framework. The model formulation and estimation are presented and evaluated with a simulation, and a real data analysis demonstrates the model in action.

Authors

• Tessa Johnson, University of Maryland - College Park

• Tracy Sweet, University of Maryland - College Park