Covariate balance under Bayesian decision theory
By: André A. F. Fumis, Victor Fossaluza, Rafael B. Stern
Potential Business Impact:
Helps doctors pick the best patient groups for tests.
We study optimal sample allocation between treatment and control groups under Bayesian linear models. We derive an analytic expression for the Bayes risk, which depends jointly on sample size and covariate mean balance across groups. Under a flat conditional prior, the covariate mean balance term simplifies to the Mahalanobis distance. Our results reveal that the optimal allocation does not always correspond to equal sample sizes, and we provide sufficient conditions under which equal allocation is optimal. Finally, we extend the analysis to sequential settings with groups of patients arriving over time.
Similar Papers
How Balanced Should Causal Covariates Be?
Methodology
Makes sure science studies are fair and true.
Cross-Balancing for Data-Informed Design and Efficient Analysis of Observational Studies
Methodology
Makes studies fairer by using results to pick groups.
Optimal Treatment Rules under Missing Predictive Covariates: A Covariate-Balancing Doubly Robust Approach
Methodology
Helps doctors pick best treatments even with missing info.