p-Uniform* is an estimator of the three-parameter selection model: Commentary on van Aert and van Assen (2026)

Abstract

Van Aert and Van Assen (2026) introduced a novel method called p-uniform* for estimating average effect size in meta-analyses subject to publication bias. They characterized the method as a distinct, more parsimonious model than the well-known three-parameter selection model (3PSM) of Hedges (1992). I argue that p-uniform* is not a distinct model, but rather a different estimator of the average effect size and heterogeneity parameters of the 3PSM. The p-uniform* estimator coincides with the maximum likelihood estimator of the 3PSM when all included effects have identical sampling variances; otherwise, the estimators differ. Before p-uniform* is used in applications, its theoretical rationale should be clarified, and its empirical performance should be more thoroughly explored through simulations under realistic data-generating processes, especially ones involving primary studies that vary in sample size.