Join the Goergen Institute for Data Science  and the Department of Political Science for Identification and Estimation of Treatment Effects from Social Network Data with Licheng Liu, PhD candidate in political science and statistics at Massachusetts Institute of Technology (MIT).

Abstract: I introduce a generalized propensity score (GPS) based approach to the identification and estimation of treatment effects from observational social network data, where formation of social tie between pair of units depends on individual level characteristics. Ignoring the tie formation process, its interaction with the treatment assignment mechanism and interference induced by the social network can lead to biased estimation of treatment effects. I propose a unified framework that addresses these challenges by jointly modeling treatment assignment and network formation, incorporating their complex interactions in observational social network data. Generalized propensity score can be semi-parametrically estimated given probabilistic models for these two processes and functional form of exposure mapping (Aronow and Samii, 2017) for effective treatment. Average potential outcomes and treatment effects are estimated with inverse probability weighting estimators. I illustrate the proposed method in several Monte Carlo studies and an empirical analysis that investigates the effect of adopting a new political communication technology on political participation in Uganda.

Bio: Licheng Liu is a Ph.D. candidate in political science and statistics at Massachusetts Institute of Technology. He works in political methodology and political economy. His research interests include causal inference, network analysis, and applications in international and comparative political economy. His work has appeared in American Journal of Political Science, Political Analysis, and Political Science Research and Methods.

This seminar is part of a tenure-track, Assistant Professor of Computational Social Science faculty search led by the Goergen Institute for Data Science, in collaboration with the Departments of Political Science, Linguistics, and Economics.