Same but different A comparison of estimation approaches for exponential random graph models for multiple networks

The Exponential Random Graph family of models (ERGM) is a powerful tool for social science research as it allows for the simultaneous modeling of endogenous network characteristics and exogenous variables such as gender, age, and socioeconomic status. However, a major limitation of ERGM is that it is mainly used for descriptive analysis of a single network. This paper examines two methods for estimating multiple networks: hierarchical and integrated. We contrast the two approaches, evaluate their accuracy and discuss the advantages and drawbacks of each. Furthermore, we make recommendations for future researchers on how to proceed with multiple network analysis depending on various factors such as the number of networks and the hierarchical structure of the data. This research is important as it highlights the need for the analysis of multiple networks in order to gain a more comprehensive understanding of social phenomena and the potential for new discoveries.

Automated summary

The Exponential Random Graph family of models (ERGM) is a powerful tool for simultaneous modeling of endogenous network characteristics and exogenous variables. This paper examines two methods for estimating multiple networks, hierarchical and integrated, and evaluates their accuracy and advantages. Recommendations are provided for future researchers on how to proceed with multiple network analysis. This research highlights the importance of analyzing multiple networks to gain a comprehensive understanding of social phenomena.