Dear Meta-Analysis Experts,
I'm running a multilevel model with metafor's rma.mv().
Does the QE statistic (Test for Residual Heterogeneity) reported by rma.mv(),
in addition to the variance explained by moderators, also account for the
variance components explained by random-effects (Sigmas, Taus, & Gammas)?
In other words, a significant QE in such models suggests that the true
effects are heterogeneous beyond what has been explained by moderators AND
random components OR just moderators?
Thank you,
Simon