Heterogeneous Treatment Effects under Network Interference: A Nonparametric Approach Based on Node Connectivity

Event Date: 

Wednesday, February 5, 2025 - 3:30pm to 4:45pm

Event Date Details: 

Wednesday, February 5th, 2025 

Event Location: 

  • HSSB 1173

Event Price: 

FREE

Event Contact: 

Dr. Heejong Bong 

Research Fellow 

Department of Statistics 

University of Michigan, Ann Arbor 

  • Department Seminar

Abstract:

In network settings, interference between units complicates causal inference as outcomes may depend on the treatments received by others in the network. Traditional estimands in such settings often aggregate treatment effects across individuals in the population. In contrast, we propose a framework for estimating node-wise counterfactual means, enabling granular insights into how network structure influences treatment effect heterogeneity. Our approach introduces KECENI (Kernel Estimation of Causal Effect under Network Interference), a doubly robust and nonparametric estimation procedure that ensures consistency and asymptotic normality under network dependence. The method’s practical utility is illustrated through an application to microfinance data, showcasing the role of network characteristics in shaping treatment effects.

Bio: 

Dr. Heejong Bong is currently a Research Fellow in the Department of Statistics at the University of Michigan, Ann Arbor. He received his PhD from Carnegie Mellon University. His current and past research interests include casual inference with network data, matrix-variate graphical model, casual inference in epidimiology, and latent factor model.