The RoBMA package provides a comprehensive set of
vignettes to help users navigate different aspects of Robust Bayesian
Meta-Analysis. This guide outlines the available vignettes and their
specific focus to help you find the relevant information for your
analysis.
Introductory Vignettes
Tutorial: Adjusting for Publication Bias in JASP and R
This is the main introduction to the RoBMA framework. It covers the basics of adjusting for publication bias using selection models, PET-PEESE, and Robust Bayesian Meta-Analysis. It is the recommended starting point for new users.
Reproducing Bayesian Model-Averaged Meta-Analysis
This vignette demonstrates how to perform a classic Bayesian model-averaged meta-analysis. It focuses on reproducing standard BMA results and understanding the core components of the method.
Advanced Modeling Features
Robust Bayesian Model-Averaged Meta-Regression
Learn how to incorporate moderators into your meta-analysis using
RoBMA.reg(). This vignette explains how to fit
meta-regression models to account for heterogeneity explained by
study-level covariates.
Multilevel Robust Bayesian Meta-Analysis
This vignette demonstrates how to perform multilevel meta-analysis to
account for dependent effect sizes (e.g., multiple estimates from the
same study). It uses the spike-and-slab algorithm
(algorithm = "ss") to efficiently estimate models with
within-study and between-study heterogeneity while adjusting for
publication bias.
Multilevel Robust Bayesian Model-Averaged Meta-Regression
This vignette demonstrates how to perform multilevel meta-regression. In addition, it illustrates how to rescale default prior distributions to work with non-standardized effect sizes.
Z-Curve Publication Bias Diagnostics
This vignette details the use of meta-analytic z-curves for diagnosing publication bias. It explains how to interpret z-curve plots and statistics provided by the package.
Specialized Applications
Informed Bayesian Model-Averaged Meta-Analysis in Medicine
This vignette focuses on applying RoBMA in medical contexts. It discusses the use of informed priors tailored for medical research questions and continuous outcomes.
Informed Bayesian Model-Averaged Meta-Analysis with Binary Outcomes
Similar to the Medicine BMA vignette, but specifically for binary
outcomes. It covers the BiBMA models (Binomial-Normal) and
appropriate prior settings for medical meta-analysis with binary
data.
Customization and Performance
Fitting Custom Meta-Analytic Ensembles
For advanced users who need to go beyond the default model ensembles. This vignette demonstrates how to customize the ensemble of models, including specifying custom priors and model combinations.
Fast Robust Bayesian Meta-Analysis via Spike and Slab Algorithm
For computationally intensive problems or quick approximations, the
“spike-and-slab” algorithm (algorithm = "ss") can be used.
This vignette explains how to use this faster alternative to the default
bridge sampling approach.
Hierarchical Bayesian Model-Averaged Meta-Analysis
This vignette introduces multilevel models. It shows how to handle
dependencies in the data (e.g., multiple effect sizes from the same
study) using the study_ids argument to specify a
hierarchical structure. Note that this vignette relies on multivariate
parameterization that is relevant only for the bridge sampling
algorithm. However, it is still helpful for describing the
parameterization.