Cluster-Robust Estimators for Bivariate Mixed-Effects Meta-Regression

TL;DR

This paper introduces two new cluster-robust variance estimators for bivariate meta-regression, designed to improve small sample performance and reduce inflated Type 1 errors in dependent effect size analyses.

Contribution

The paper proposes novel estimators that transform residual variances using the hat matrix diagonal, enhancing robustness in small samples for dependent effect sizes.

Findings

New estimators outperform existing methods in simulations

Improved control of Type 1 error rates in small samples

Effective application demonstrated on real-world data

Abstract

Meta-analyses frequently include trials that report multiple effect sizes based on a common set of study participants. These effect sizes will generally be correlated. Cluster-robust variance-covariance estimators are a fruitful approach for synthesizing dependent effects. However, when the number of studies is small, state-of-the-art robust estimators can yield inflated Type 1 errors. We present two new cluster-robust estimators, in order to improve small sample performance. For both new estimators the idea is to transform the estimated variances of the residuals using only the diagonal entries of the hat matrix. Our proposals are asymptotically equivalent to previously suggested cluster-robust estimators such as the bias reduced linearization approach. We apply the methods to real world data and compare and contrast their performance in an extensive simulation study. We focus on bivariate meta-regression, although the approaches can be applied more generally.