MODE: Minimax Optimal Deterministic Experiments for Causal Inference in the Presence of Covariates
Shaohua Xu, Songnan Liu, Yongdao Zhou, Boris Ryabko · 2024
Read MODE: Minimax Optimal Deterministic Experiments for Causal Inference in the Presence of Covariates by Shaohua Xu, Songnan Liu, Yongdao Zhou, Boris Ryabko (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11727206.txt.
Description
Shaohua Xu and co-authors present a 2024 method for designing fairer experiments to measure cause and effect. In trials with many background traits, random assignment can leave the treated and control groups uneven on those traits. The team measures that unevenness as a gap score and asks how to make the worst-case error of the standard effect estimate as small as possible.
Their design, called MODE, picks the assignment by a fixed rule built from low-clump number methods rather than by chance. It shrinks the gap score in a way that does not depend on any outcome model. Proofs show its effect estimate settles faster than those from random or re-drawn assignments. Tests on fake data back the claim, with smaller trait gaps and tighter estimates than rival designs.
Catalog record
| CREATOR | Shaohua Xu, Songnan Liu, Yongdao Zhou, Boris Ryabko |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |