A0689
Title: Kernel discrepancy-based rerandomization for controlled experiments
Authors: Yiou Li - DePaul University (United States) [presenting]
Lulu Kang - University of Massachusetts Amherst (United States)
Abstract: Controlled experiments have been widely used in various disciplines for causal inference. Rerandomization has been proposed and advocated to improve the covariates balance. One key component in rerandomization is the balance criterion. The kernel discrepancy is introduced between the empirical distributions of the covariates in different treatment groups, and the upper bound of the variance of the difference-in-mean estimator of the treatment effects is shown to be regulated. Accordingly, using kernel discrepancy is proposed as a balance criterion. Using the linear kernel function, the distribution of the kernel discrepancy is obtained for finite samples, which provides the critical value for an acceptable rerandomization. For more complicated kernel functions, empirical distributions are proposed using the kernel discrepancy to obtain the critical value. The discrepancy-based criterion is model-free and thus makes the estimation of the treatment effect(s) robust to the model assumptions. More importantly, the proposed design is applicable to both continuous and categorical response measurements. Through simulation study and a real example, it is shown that the proposed design approach achieves accurate estimation even if the model assumption is not correct.
