clusterRCT - code for analyzing blocked, cluster-randomized trials

Introduction

clusterRCT provides an easy means to run a collection of different estimators to estimate average treatment effects in blocked, cluster randomized trials (trials where you have randomization blocks, and within each block clusters are randomized into treatment and control). It also provides some methods for exploring data with a blocked, clustered structure. The package also can be used for simple cluster-randomized trials with no blocking.

Core methods are compare_methods() which compares the estimates from a collection of different estimation strategies on a given dataset, and describe_clusterRCT() which describes overall characteristics of a given dataset.

The package was designed to support a paper, “An Applied Researcher’s Guide to Estimating Effects from Multisite Cluster Randomized Trials: Estimands, Estimators, and Estimates,” accepted to JREE in 2026. For the replication files of the paper, see the replication GitHub. That set of code heavily uses this package in the implemented simulations and empirical data analyses.

Installation

Installation is simple:

devtools::install_github("lmiratrix/clusterRCT")

Replicating the paper

The paper’s results were generated with a slightly older, less cleaned-up version of the package than what is on the main branch today. Ongoing cleanup (dependency fixes, bug fixes, documentation) is not intended to change any reported estimates, but to install the exact version used to produce the paper’s results, use the tagged release instead of the latest version:

devtools::install_github("lmiratrix/clusterRCT",
            ref = "jree-submission-snapshot-2025-11-20")

Acknowledgements

The research reported here was partially supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R305D220046. The opinions expressed are those of the authors and do not represent views of the Institute or the U.S. Department of Education.