R Package iglm: Regression under Interference in Connected Populations

2026-04-13

AI Summary

This study addresses the challenge of modeling interference effects—such as spillovers—among individuals in interconnected populations, which conventional regression methods fail to capture accurately. The authors propose a scalable interference regression framework, implemented for the first time in R, that offers both theoretical guarantees and computational efficiency while supporting customizable model components and flexible representation of complex dependence structures. The approach formulates a convex optimization objective based on pseudolikelihood and employs a combination of Minorization-Maximization and Quasi-Newton algorithms for efficient estimation across datasets of varying scales. Empirical analyses on real-world data concerning hate speech in social media and student communication networks demonstrate the method’s effectiveness and practical utility.

Abstract

We introduce R package iglm, which implements a comprehensive framework for studying relationships among predictors and outcomes under interference. The implemented regression framework facilitates the study of spillover and other phenomena in connected populations and has important advantages over existing packages, among them scalability and provable theoretical guarantees. On the computational side, the regression framework relies on scalable methods that can be applied to small and large data sets, by solving a convex optimization program based on pseudo-likelihoods using Minorization-Maximization and Quasi-Newton algorithms. On the statistical side, the regression framework comes with provable theoretical guarantees. To increase the versatility of iglm, users can add custom-built model terms. We showcase iglm using two data sets, including hate speech on the social media platform X and communications among students.

Problem

• interference
• spillover
• connected populations
• regression
• pseudo-likelihood

Innovation

• interference
• pseudo-likelihood
• convex optimization
• scalable regression
• connected populations