Rerandomization in Stratified Randomized Experiments

Stratification and rerandomization are two well-known methods used in randomized experiments for balancing the baseline covariates. Renowned scholars experimental design have recommended combining these methods; however, limited studies addressed statistical properties of this combination. This article proposes to be stratified experiments, based on overall stratum-specific Mahalanobis distances. The first method is applicable nearly arbitrary numbers strata, strata sizes, proportions treated units. second method, which generally more efficient than suitable situations number fixed with their sizes tending infinity. Under randomization inference framework, we obtain asymptotic distributions estimators formulas variance reduction when compared randomization. Our analysis does not require any modeling assumption regarding potential outcomes. Moreover, provide asymptotically conservative confidence intervals average treatment effect. advantages proposed exhibited through an extensive simulation study a real-data example.