Causal Bandits Podcast Causal Inference & the "Bayesian-Frequentist War" | Richard Hahn S2E8 | CausalBanditsPodcast.com
75 snips
Dec 27, 2025 In this enlightening discussion, Professor Richard Hahn from Arizona State University delves into the ongoing debate between Bayesians and frequentists in statistics. He shares insights on why Bayesian Additive Regression Trees (BART) are effective and how they compare to models like XGBoost. The conversation uncovers the significance of heterogeneous treatment effects and the challenges in generalizing RCT results. Richard emphasizes the importance of realistic simulation studies for understanding causal inference, while coining the term "feature-level selection bias"—a must-listen for stats enthusiasts!
AI Snips
Chapters
Books
Transcript
Episode notes
Regularize Treatment Effects Directly
- Regularize treatment effects directly when estimating heterogeneous effects to avoid overfitting.
- Reparameterize models (e.g., causal BART) so priors act on the treatment effect, not separate potential outcomes.
Feature Choice Hides Heterogeneity
- Heterogeneous treatment effects exist but often go undetected because measured features are convenient, not scientifically chosen.
- Measure theory-driven covariates to reveal meaningful heterogeneity instead of accepting handed features.
Target The Population Effect You Care About
- Analyze population average effects if that's your real goal instead of only sample ATE from randomization inference.
- Prefer approximate answers to the right question over exact answers to the wrong one.
