Matching on-the-fly in Sequential Experiments for Higher Power and Efficiency

TL;DR

This paper introduces a dynamic matching procedure for sequential experiments that enhances statistical power and efficiency by adaptively pairing subjects and combining data for more accurate treatment effect estimation.

Contribution

It presents a novel on-the-fly matching method for sequential trials, improving power and efficiency over existing allocation procedures.

Findings

Higher statistical power demonstrated in simulations

Increased efficiency compared to traditional methods

Effective in both controlled and real-world data

Abstract

We propose a dynamic allocation procedure that increases power and efficiency when measuring an average treatment effect in sequential randomized trials. Subjects arrive iteratively and are either randomized or paired via a matching criterion to a previously randomized subject and administered the alternate treatment. We develop estimators for the average treatment effect that combine information from both the matched pairs and unmatched subjects as well as an exact test. Simulations illustrate the method's higher efficiency and power over competing allocation procedures in both controlled scenarios and historical experimental data.