Improving control over unobservables with network data

2025/11/01 by Vincent Starck, Starck, Vincent · 1 citation
Mathematics · Social Sciences · #Advanced Causal Inference Techniques #Intergenerational and Educational Inequality Studies #School Choice and Performance #econ.EM

paper · pdf · doi:10.48550/arxiv.2511.00612

openalex publication_date 2025/11/01 · openalex created_date 2025/11/05 · openalex updated_date 2026/07/28

Abstract

This paper develops a method to conduct causal inference in the presence of unobserved confounders by leveraging networks with homophily, a frequently observed tendency to form edges with similar nodes. I introduce a concept of asymptotic homophily, according to which individuals' selectivity scales with the size of the potential connection pool. The resulting network formation model accommodates common empirical features such as homophily, degree heterogeneity, sparsity, and clustering, and provides a framework to obtain consistent estimators of treatment effects that are robust to selection on unobservables. In an application, I recover an estimate of the effect of parental involvement on students' test scores that is greater than that of OLS, arguably due to the estimator's ability to account for unobserved ability.

Citations

• Efficient Estimation of Average Treatment Effects Using the Estimated Propensity Score
• Estimating causal effects of treatments in randomized and nonrandomized studies.

Cited by

• Peer effect analysis with latent processes

Related

• The Network Propensity Score: Spillovers, Homophily, and Selection into Treatment
• From Acquaintances to Friends: Homophily and Learning in Networks
• Network homophily via tail inequalities
• Statistical Tests for Contagion in Observational Social Network Studies
• Graph Neural Networks for Causal Inference Under Network Confounding