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
Journal of the American Statistical Association · 2023
Published2023Cited by5
Citation counts from OpenAlex; downloads from CRAN. Updated 2026-08-15.
Abstract
The recent work by Krivitsky, Coletti & Hens [KCH] provides an important new contribution to the Exponential-Family Random Graph Models [ERGMs], a start-to-finish approach to dealing with multi-network ERGMs. Although multi-network ERGMs have been around for a while (mostly in the form of block-diagonal models and multi-level ERGMs, see Duxbury and Wertsching (2023), Wang et al. (2013), Slaughter and Koehly (2016)), not much care has been given to the estimation and post-estimation steps. In their paper, Krivitsky, Coletti & Hens give a detailed layout of how to build, estimate, and analyze multi-ERGMs with heterogeneous data sources. In this comment, I will focus on two issues the authors did not discuss, namely, sample size requirements and multicollinearity.
Cite
@article{vegayonPowerMulticollinearitySmall2023a,
title = {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},
author = {{Vega Yon}, {George G.}},
year = {2023},
month = {10},
journal = {Journal of the American Statistical Association},
volume = {118},
number = {544},
pages = {2228–2231},
doi = {10.1080/01621459.2023.2252041},
url = {https://doi.org/10.1080/01621459.2023.2252041},
issn = {0162-1459, 1537-274X},
}