Power and Multicollinearity in Small Networks: A Discussion of "Tale of Two Datasets: Representativeness and Generalisability of Inference for Samples of Networks" by Krivitsky, Coletti, and Hens.

This article discusses the work of Krivitsky, Coletti, and Hens on Exponential-Family Random Graph Models (ERGMs) and their application to multi-network ERGMs. The authors explain how to build, estimate, and analyze multi-ERGMs using different data sources. The article focuses on two important issues that the authors did not cover: sample size requirements and multicollinearity. It emphasizes the need for power analysis in ERGMs, especially for small networks, and suggests using simulation studies to determine the number of networks needed for research projects. The article also discusses the concept of network size and the relevance of projectivity in statistical inference. The supplementary materials provide access to the code used for generating figures and analyses, which can be useful for library patrons conducting research on these topics. [Extracted from the article]