The use of weighted averages of Hedges' <i>d</i> in meta?analysis: Is it worth it?

Hedges' d is a measure of effect size widely used to standardize results across different studies in ecological and evolutionary meta-analyses. When summarizing or other sizes multiple studies, it considered best practice use weighted averages, with weights inversely proportional the variances estimated sizes. Importantly, within-study variance d, ? 2 , function sample size, n, ?, itself. Since true are unknown, also unknown needs be estimated, for which numerous approaches have been proposed. We examined behaviour inverse-variance weights, specifically performance 14 unweighted variants estimator, test conditions under weighting improved performance. found that when (per group) were >10, there was little difference between estimators did not notably improve performance; bias <5% differences root mean squared error <10% all estimators, except between-study ? relatively small compared ( / < 0.5). propose new estimator (a) most efficient subset minimally biased (b) had least estimating ?2. Contrary current guidance, we contend simpler generally acceptable. Exceptions relative 0.5) large disparity (>3X) proportion meta-analysis. Using averages may enable more robust inferences by allowing cannot extracted (a common practical limitation) included. That said, averaging necessary where estimation itself interest.