Fast Rerandomization for Balancing Covariates in Randomized Experiments: A Metropolis-Hastings Framework

Jiuyao Lu, Tianruo Zhang, Ke Zhu · 2026

Read Fast Rerandomization for Balancing Covariates in Randomized Experiments: A Metropolis-Hastings Framework by Jiuyao Lu, Tianruo Zhang, Ke Zhu (2026) 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/arxiv-2602-07613.txt.

Description

Rerandomization lifts randomized tests by drawing new treatment assignments until covariate balance meets a threshold, but rejection sampling becomes prohibitively slow under strict thresholds. Jiuyao Lu and two co-authors build a fast method on Metropolis-Hastings sampling.

The method preserves uniformity over the allowed assignment space, keeping the theoretical guarantees of the classic method, while dodging the inefficiency that forces users into suboptimal thresholds. Experimental builders get strong balance without the wait, and results that remain valid under the accelerated scheme. The method keeps the balance that justifies inference by resampling, rather than trading theory for speed.

Catalog record

CREATORJiuyao Lu, Tianruo Zhang, Ke Zhu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE103 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-07613

Browse the full library · About the Tower of Babel