Feature Selection for Causal Effect Estimation


We define the notion of a minimal control function, on the basis of which a novel regression penalty is devised that is unbiased for average treatment effects. The development of the new approach combines insights from three distinct methodological traditions for studying causal effect estimation: potential outcomes, causal diagrams, and structural models with additive errors. It is demonstrated that traditional feature selection and/or regularization approaches to treatment effect estimation can exhibit severe bias for average and conditional average treatment effects. This is joint work with Drew Herren.


DoMSS Seminar
Monday, January 24        
1:30 pm
Zoom meeting room link:  https://asu.zoom.us/j/6871076660

Note: This meeting will be via Zoom. This semester, we anticipate some talks will be in person but most will be by Zoom.


Richard Hahn
Associate Professor of Statistics
Arizona State University

Virtual via Zoom