Prefer a chat interface with context about you and your work?
Small sample performance of bias‐corrected sandwich estimators for cluster‐randomized trials with binary outcomes
The sandwich estimator in generalized estimating equations (GEE) approach underestimates the true variance in small samples and consequently results in inflated type I error rates in hypothesis testing. This fact limits the application of the GEE in cluster‐randomized trials (CRTs) with few clusters. Under various CRT scenarios with correlated binary …