method.RoBMA: Robust Bayesian Meta-Analysis (RoBMA) Method

In PublicationBiasBenchmark: Benchmark for Publication Bias Correction Methods

Robust Bayesian Meta-Analysis (RoBMA) Method

Description

Implements the robust Bayesian meta-analysis (RoBMA) method that uses Bayesian model-averaging to combine results across several complementary publication bias adjustment methods. See \insertCitemaier2023robust;textualPublicationBiasBenchmark and \insertCitebartos2023robust;textualPublicationBiasBenchmark for details. If "study_id" column is included in the data input, the method uses multilevel parameterization as described in \insertCitebartos2025robust;textualPublicationBiasBenchmark.

Note that the prior settings is dispatched based on "es_type" column attached to the dataset. The resulting estimates are then summarized on the same scale as was the dataset input (for "r", heterogeneity is summarized on Fisher's z).

Important: This method requires JAGS (Just Another Gibbs Sampler) to be installed on your system. Please download and install JAGS from https://mcmc-jags.sourceforge.io/ before using this method.

Usage

## S3 method for class 'RoBMA'
method(method_name, data, settings)

Arguments

method_name	Method name (automatically passed)
data	Data frame with yi (effect sizes), sei (standard errors), es_type (either "SMD" for Cohen's d / Hedge's g, "logOR" for log odds ratio, "z" for Fisher's z, or "r" for correlations. Defaults to "none" which re-scales the default priors to unit-information width based on total sample size supplied "ni".)
settings	List of method settings (see Details.)

Details

The following settings are implemented

"default"

RoBMA-PSMA with publication bias adjustment as described in \insertCitebartos2023robust;textualPublicationBiasBenchmark. (the MCMC settings was reduced to speed-up the simulations) with the three-level specification whenever "study_ids" are supplied with the data

"PSMA"

RoBMA-PSMA with publication bias adjustment as described in \insertCitebartos2023robust;textualPublicationBiasBenchmark. (the MCMC settings was reduced to speed-up the simulations) with the three-level specification whenever "study_ids" are supplied with the data

Value

Data frame with RoBMA results

Author(s)

František Bartoš [e-mail]

References

\insertAllCited

Examples

# Generate some example data
data <- data.frame(
  yi      = c(0.2, 0.3, 0.1, 0.4, 0.25),
  sei     = c(0.1, 0.15, 0.08, 0.12, 0.09),
  es_type = "SMD"
)

# Apply RoBMA method
result <- run_method("RoBMA", data)
print(result)

PublicationBiasBenchmark documentation built on March 16, 2026, 5:07 p.m.