Chad Hazlett
Chad Hazlett
Professor of Political Science and Statistics, UCLA
Causal inference methods for hard questions in politics, policy, and medicine.
Featured work
Making sense of sensitivity: Extending omitted variable bias
A general, interpretable framework for asking how much unobserved confounding would have to exist to overturn a result — now the standard tool for honest observational-studies reporting.
Cinelli & Hazlett · JRSS B · 2020
Read →kpop: A kernel balancing approach for survey weighting
Weights non-representative survey samples to match population targets in a reproducing-kernel space — reducing the specification assumptions practitioners usually have to defend.
Hartman, Hazlett & Sterbenz · JRSS A · 2025
Read →Wildfire exposure increases pro-environment voting
A natural-experiment study of how climate-related hazards reshape political behavior — and where partisanship still dominates.
Hazlett & Mildenberger · APSR · 2020
Read →From "is it unconfounded?" to "how much confounding would it take?"
A sensitivity-based alternative to binary debates over causal identification in observational social science — applied to support for peace in Colombia.
Hazlett & Parente · Journal of Politics · 2023
Read →Safe learning outside randomized trials: SCQE for COVID-19 therapies
Bounds the effects of remdesivir, hydroxychloroquine, and dexamethasone using only a plausible assumption about baseline trends — demonstrating how SCQE could have informed safe decisions before RCTs completed.
Hazlett & Wulf et al. · Observational Studies · 2025
Read →Seeing like a district: What close-election designs for leader characteristics can and cannot tell us
Clarifies what the popular close-election regression discontinuity can — and cannot — identify when researchers use it to study how leader traits shape policy and conflict.
Bertoli & Hazlett · Political Analysis · 2025
Read →Inference at the data's edge: Gaussian processes for extrapolation uncertainty
Uses Gaussian-process regression to make model-dependency visible — quantifying rather than masking the uncertainty that arises when we extrapolate beyond well-supported regions of covariate space.
Cho, Kim & Hazlett · Political Analysis · 2026
Read →