Optimal A Priori Balance in the Design of Controlled Experiments

Nathan Kallus

We develop a unified theory of designs for controlled experiments that balance baseline covariates a priori (before treatment and before randomization) using the framework of minim…