Optimality of Matched-Pair Designs in Randomized Controlled Trials

32 Pages Posted: 15 Dec 2019 Last revised: 9 May 2022

See all articles by Yuehao Bai

Yuehao Bai

University of Michigan at Ann Arbor - Department of Economics

Date Written: April 11, 2022

Abstract

In randomized controlled trials, treatment is often assigned by stratified randomization. I show that among all stratified randomization schemes, a certain matched-pair design achieves the maximum statistical precision for estimating the average treatment effect. In an important special case, the optimal design pairs units according to the baseline values of the primary outcome variable of interest. In a simulation study, such a design lowers the standard error by an average of 10% relative to the status-quo methods and 34% in some cases. Therefore, I recommend that researchers pair units according to the baseline outcome if it is available.

Keywords: Matched-pair design, baseline outcome, stratified randomization, experiment, randomized controlled trial

JEL Classification: C12, C13, C14, C90

Suggested Citation

Bai, Yuehao, Optimality of Matched-Pair Designs in Randomized Controlled Trials (April 11, 2022). Available at SSRN: https://ssrn.com/abstract=3483834 or http://dx.doi.org/10.2139/ssrn.3483834

Yuehao Bai (Contact Author)

University of Michigan at Ann Arbor - Department of Economics ( email )

611 Tappan Street
Ann Arbor, MI 48109-1220
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
341
Abstract Views
1,669
Rank
161,446
PlumX Metrics