RoBMA: Robust Bayesian Meta-Analyses

A framework for estimating ensembles of meta-analytic models (assuming either presence or absence of the effect, heterogeneity, and publication bias) and using Bayesian model averaging to combine them. The ensembles use Bayes Factors to test for the presence or absence of the individual components (e.g., effect vs. no effect) and model-averages parameter estimates based on posterior model probabilities (Maier, Bartoš & Wagenmakers, 2020, <doi:10.31234/>). The user can define a wide range of non-informative or informative priors for the effect size, heterogeneity, and weight functions. The package provides convenient functions for summary, visualizations, and fit diagnostics.

Version: 1.0.4
Imports: runjags, bridgesampling, rjags, coda, psych, stats, graphics, extraDistr, scales, DPQ, callr, Rdpack
LinkingTo: BH
Suggests: ggplot2, rstan, metaBMA, testthat, vdiffr, knitr, rmarkdown
Published: 2020-08-07
Author: František Bartoš ORCID iD [aut, cre], Maximilian Maier ORCID iD [aut], Eric-Jan Wagenmakers ORCID iD [ths], Joris Goosen [ctb]
Maintainer: František Bartoš <f.bartos96 at>
License: GPL-3
NeedsCompilation: yes
Citation: RoBMA citation info
Materials: README NEWS
CRAN checks: RoBMA results


Reference manual: RoBMA.pdf
Vignettes: Fitting custom meta-analytic ensembles
Reproducing BMA
Common warnings and errors
Package source: RoBMA_1.0.4.tar.gz
Windows binaries: r-devel:, r-release:, r-oldrel:
macOS binaries: r-release: RoBMA_1.0.4.tgz, r-oldrel: not available
Old sources: RoBMA archive


Please use the canonical form to link to this page.