WITHDRANW: Rejective Sampling, Rerandomization, and Regression Adjustment in Survey Experiments

• Activity Number: 285 - New Advances in Sample Design and Adjusting for Survey Nonresponse
• Type: Contributed
• Date/Time: Wednesday, August 11, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Survey Research Methods Section
• Author(s): Zihao Yang and Tianyi Qu and Xinran Li
• Companies: University of Illinois at Urbana-Champaign and University of Illinois at Urbana-Champaign and University of Illinois
• Keywords: causal inference; potential outcome; randomization-based inference; covariate balance; Mahalanobis distance; randomization test

Abstract

As pointed out by Morgan and Rubin (2012), chance imbalances often exist in covariate distributions between treatment groups. Not surprisingly, such covariate imbalances also occur in randomized survey experiments. More importantly, the covariate imbalances happen not only between different treatment groups, but also between the sampled experimental units and the overall population of interest. We propose a two-stage rerandomization design that can actively avoid these undesirable covariate imbalances at both the sampling and treatment assignment stages. We further develop asymptotic theory for rerandomized survey experiments, demonstrating that rerandomization provides better covariate balance, more precise estimates of treatment effects, and shorter confidence intervals that are still asymptotically valid. We also propose covariate adjustment to deal with the remaining covariate imbalance after rerandomization, show that it can further improve both the sampling and estimated efficiency, and connect it to usual regression models with least squares estimates.