version 4.0.0

Breaking changes

• rewrites the package around the unified brma class hierarchy. Single-model fits now use brma(), brma.glmm(), bselmodel(), bPET(), and bPEESE(); model-averaged fits use BMA(), BMA.glmm(), and RoBMA().
• removes the legacy RoBMA.reg(), NoBMA(), NoBMA.reg(), BiBMA(), and BiBMA.reg() constructors. Use mods, scale, and cluster in the new constructors, BMA() for no-bias normal-likelihood model averaging, and BMA.glmm() for GLMM model averaging.
• replaces old input aliases such as d, r, logOR, OR, z, y, se, v, n, study_names, study_ids, weight, and transformation with yi, vi/sei, ni, slab, cluster, weights, measure, output_measure, and transform.
• removes legacy helper APIs including combine_data(), check_setup(), extract_posterior(), marginal_summary(), marginal_plot(), plot_models(), adjusted_effect(), as_zcurve(), and the old z-curve plotting methods.
• normal-likelihood fitting functions now require an explicit measure for fitted models. Use measure = "GEN" for generic effect sizes without a known unit-information scale.
• update() for brma objects now focuses on extending MCMC samples, updating labels, and refreshing cached quantities, not changing model structure.
• set_convergence_checks() no longer accepts the old remove_failed and balance_probability arguments.

Features

• adds brma() / brma.norm() for single normal-likelihood Bayesian meta-analysis, including random-effects, meta-regression, multilevel, and location-scale models.
• adds brma.glmm() for binomial-normal and Poisson-normal GLMM meta-analysis from raw two-arm counts (measure = "OR" and "IRR").
• adds single-model publication-bias constructors bselmodel(), bPET(), and bPEESE().
• adds BMA() / BMA.norm() for Bayesian model averaging without publication-bias adjustment.
• adds BMA.glmm() for Bayesian model averaging of GLMM meta-analyses without publication-bias adjustment.
• rewrites RoBMA() as a product-space model-averaged ensemble over effect, heterogeneity, moderator, scale, and publication-bias components.
• adds formula/data-frame input handling for effect sizes, moderators, scale predictors, clusters, labels, subsets, likelihood weights, and raw GLMM counts.
• adds default prior construction from standardized effect-size measures, estimated or manually supplied unit-information standard deviations, and informed empirical priors.
• adds prior_weightfunction(), wf_cumulative(), wf_fixed(), and wf_independent() for BayesTools-backed selection-weightfunction priors.
• adds prior_PET(), prior_PEESE(), prior_none(), prior_factor(), prior_informed(), and BayesTools contrast helpers as package-level prior utilities.
• adds posterior package interfaces via as_draws(), as_draws_array(), as_draws_df(), as_draws_list(), as_draws_matrix(), and as_draws_rvars() for fitted models and brma_samples.
• adds the brma_samples posterior-sample class with print, summary, matrix, and posterior conversion methods.
• adds predict.brma() for posterior predictions of fixed terms, cluster effects, latent true effects, observed responses, and scale terms, with newdata, conditional, bias_adjusted, output_measure, and transform support.
• adds convenience wrappers fitted(), pooled_effect(), pooled_heterogeneity(), blup(), true_effects(), and ranef() for brma objects.
• adds model-comparison helpers add_loo(), loo(), loo_compare(), loo_weights(), check_loo(), add_waic(), waic(), and logLik() using the loo package.
• adds bridge-sampling marginal likelihood support for single-model brma fits via add_marglik(), bridge_sampler(), logml(), bf(), bayes_factor(), and post_prob().
• adds residual and influence diagnostics: residuals(), rstandard(), rstudent() / LOO-PIT, hatvalues(), influence(), dfbetas(), dffits(), cooks.distance(), covratio(), and vif().
• adds plotting methods for brma objects: posterior/prior plots, funnel(), regplot(), qqnorm(), radial() / galbraith(), MCMC diagnostic plots, weightfunction plots, and PET-PEESE plots.
• adds marginal_means() with summary and plotting methods for moderator models.
• adds summary_models() for marginal and individual model-weight summaries of product-space RoBMA, BMA, and BMA.glmm objects.
• adds interpret() for concise textual interpretation of fitted brma and model-averaged objects.
• renames the zplot diagnostic API to as_zplot() and adds the direct plotting wrapper zplot(), with plot(), hist(), lines(), summary(), and print methods for zplot objects.
• adds RoBMA.options() and RoBMA.get_option() package options for defaults such as core count, automatic LOO/WAIC/marginal-likelihood computation, prior scaling defaults, and selection-bias defaults.

Changes

• renames the multilevel clustering argument to cluster.
• renames study labels to slab, matching metafor naming.
• renames likelihood weights to weights and applies them consistently to posterior fitting, log-likelihoods, LOO, WAIC, and diagnostics.
• uses measure, output_measure, and transform for effect-size scale handling. Supported conversions include SMD, COR, ZCOR, and OR; transform = "EXP" exponentiates log ratio measures for display.
• standardizes continuous predictors by default and transforms reported coefficients back to the original scale unless standardized coefficients are requested.
• uses treatment contrasts by default for single-model constructors and mean-difference contrasts by default for model-averaged constructors.
• changes predict.brma() default to type = "terms". GLMM type = "response" predictions return continuity-corrected effect-size estimators by default via as_measure = TRUE.
• separates output unit from conditioning_depth for residuals, fitted values, LOO, WAIC, and related diagnostics.
• supports estimate-level and, for multilevel models, cluster-level LOO/WAIC targets with target metadata to prevent invalid comparisons.
• keeps bridge-sampling marginal likelihoods for single-model brma objects; product-space RoBMA, BMA, and BMA.glmm objects relly on product-space only.
• routes selection-weightfunction priors through the BayesTools selection backend and selected-normal kernel, removing legacy weighted-normal mapping paths.
• uses bias_indicator and branch-aware selected-normal contexts for RoBMA publication-bias mixtures instead of inferring selection branches from omega.
• increases zplot default posterior thinning controls to 10000 samples and accepts Inf where full posterior evaluation is requested.
• adds max_samples controls to expensive funnel, regplot, and zplot summaries.
• updates the package startup message to point users to vignette("v00-introduction", package = "RoBMA").
• requires BayesTools 0.3.0 for forward API and selection-backend support.
• adds bridgesampling, loo, MASS, and parallel as imports and posterior as a suggested package.

Fixes

• fixes loading and runtime checks for the RoBMA JAGS module and native R routines.

Performance and internals

• moves fitting to JAGS product-space models with mixture-prior indicators for model averaging.
• replaces legacy weighted-normal and multivariate-normal native code with selected-normal kernels shared by JAGS and R-native calls.
• adds native selected-normal routines for log likelihoods, normalizers, CDFs, moments, RNG, weighted summaries, funnel contours, regplot intervals, and zplot densities/threshold summaries.
• adds native GLMM marginal and cluster log-likelihood helpers for binomial and Poisson models.
• caches selected-normal normalizers and uses telescoping selection probabilities with log-space fallbacks for better numerical stability.
• relocates selected-normal C++ code to src/selnorm/ and updates Makevars*, native registration, cleanup rules, and JAGS distribution registration.
• removes unused native matrix/LAPACK helper sources and older source-level transformation helpers.

Documentation and tests

• reorganizes vignettes into numbered workflows covering introduction, prior distributions, baseline Bayesian meta-analysis, feature coverage, metafor parity, model averaging, RoBMA, multilevel models, medicine examples, and zplot diagnostics.
• regenerates roxygen documentation for the new constructors, priors, predictions, summaries, diagnostics, plots, model-comparison methods, and datasets.
• refreshes the README and pkgdown site for the 4.0.0 API.
• adds cached model fits under numbered vignette/model directories.
• refactors tests into ordered input, fitting, prediction, plotting, diagnostics, model-comparison, selected-normal kernel, and vignette-cache coverage.
• adds regression tests for selected-normal telescope probabilities, native/R fallback parity, posterior-row alignment, GLMM response conversion, LOO/WAIC targets, bridge sampling, and visual outputs.

version 3.6.1

Features

• Explanation vignette that helps navigate users through the vignettes
• two vignettes demonstrating robust Bayesian meta-analysis and meta-regressions
• summary() function now provides publication bias model type summary (type = "models") for models fitted using algorithm = "ss"
• improves control over zplot diagnostics (i.e., specifying col, border, etc for the individual elements)

version 3.6

Features

• funnel() plot to visualize residuals vs the expected sampling distribution for RoBMA() and RoBMA.reg() models when using the algorithm = "ss"
• residuals() method for RoBMA() and RoBMA.reg() models when using the algorithm = "ss"
• as_zplot() function to transform meta-analytic models into a zplot object, only available for RoBMA() and RoBMA.reg() fitted using the algorithm = "ss"
• plot(), summary(), and print() functions for the as_zplot objects

version 3.5.1

Features

• summary() function now supports a standardized_coefficients argument to report either standardized (default) or raw meta-regression coefficients
• extract() function to extract the posterior samples of the model parameters
• true_effects() function to summarize the true effect size estimates of RoBMA() and RoBMA.reg() models when using the algorithm = "ss"
• predict() method for RoBMA() and RoBMA.reg() models when using the algorithm = "ss"

Fixes

• fitting a meta-regression using predictors with missing values result in a clear error message

Changes

• improving the speed of unit tests

version 3.5

Features

• approximate and computationally feasibly 3lvl selection models via the RoBMA() and RoBMA.reg() functions with the cluster argument when using algorithm = "ss"
• 3lvl binomial-normal models for binary data via the BiBMA and BiBMA.reg functions with the cluster argument when using algorithm = "ss"
• pooled_effect() function to compute the pooled effect size from the RoBMA.reg, NoBMA.reg, and BiBMA.reg models
• adjusted_effect() function to compute the adjusted effect size from the RoBMA.reg, NoBMA.reg, and BiBMA.reg models
• enables summary_heterogeneity() for BiBMA models

Fixes

• passing and checks of the cluster and study_labels arguments
• PEESE prior distribution now scale as 1/scale instead of 1/scale^2 with the rescale_priors argument
  
• the conditional prediction interval based on summary_heterogeneity() is now conditional on the presence of the effect
• additional minor prior handling fixes (i.e., missing marginal estimates when only alternative prior distributions were specified etc)
• diagnostics with mixture baseline priors when using algorithm = "ss"
• summary_heterogeneity() with only a single study does not produce relative heterogeneity instead of crashing

version 3.4

Features

• adding binomial-normal meta-regression models for binary data via the BiBMA.reg function
• the spike and slab algorithm for faster model estimation via the algorithm = "ss" argument for BiBMA models
• default prior distributions for all parameters of BiBMA models are now set via the set_default_binomial_priors() function

version 3.3

Features

• the spike and slab algorithm for faster model estimation via the algorithm = "ss" argument (see a new vignette for more details)
• refactoring of the JAGS C++ code of weighted distributions and exporting of the lpdfs into JAGS (maintenance)
• weights_mix JAGS prior distribution to sample a mixture of weight functions directly

Fixes

• incorrectly omitting models with more than one predictor when computing conditional marginal summary

version 3.2.1

Features

• default prior distributions for all parameters are now set via the set_default_priors() function
• rescale_priors argument allows to conveniently re-scale the prior distributions for the effect, heterogeneity, and bias simultaneously

version 3.2

Features

• summary_heterogeneity() function to summarize the heterogeneity of the RoBMA models (prediction interval, tau, tau^2, I^2, and H^2)
• check_RoBMA_convergence() function to check the convergence of the RoBMA models
• adds informed prior distributions for binary and time-to-event outcomes via BayesTools 0.2.17

Fixes

• checking and fixing the number of available cores upon loading the package (hopefully fixes some parallelization issues)
• update() function re-evaluates convergence checks of individual models (https://github.com/FBartos/RoBMA/issues/34)
• typos and minor issues in the vignettes

version 3.1

Features

• binomial-normal models for binary data via the BiBMA function
• NoBMA and NoBMA.reg() functions as wrappers around RoBMA RoBMA.reg() functions for simpler specification of publication bias unadjusted Bayesian model-averaged meta-analysis
• adding odds ratios output transformation`
• extending (instead of a complete refitting) of models via the update.RoBMA() function (only non-converged models by default or all by setting extend_all = TRUE)

Fixes

• handling of non-converged models

version 3.0.1

Fixes (thanks to Don & Rens)

• compilation issues with Clang (https://github.com/FBartos/RoBMA/issues/28)
• lapack path specifications (https://github.com/FBartos/RoBMA/issues/24)

version 3.0

Features

• meta-regression with RoBMA.reg() function
• posterior marginal summary and plots for the RoBMA.reg models with summary_marginal() and plot_marginal() functions
• new vignette on hierarchical Bayesian model-averaged meta-analysis
• new vignette on robust Bayesian model-averaged meta-regression
• adding vignette from AMPPS tutorial
• faster implementation of JAGS multivariate normal distribution (based on the BUGS JAGS module)
• incorporating weight argument in the RoBMA and combine_data functions in order to pass custom likelihood weights
• ability to use inverse square weights in the weighted meta-analysis by setting a weighted_type = "inverse_sqrt" argument

Changes

• reworked interface for the hierarchical models. Prior distributions are now specified via the priors_hierarchical and priors_hierarchical_null arguments instead of priors_rho and priors_rho_null. The model summary now shows Hierarchical component summary.

version 2.3.2

Fixes

• suppressing start-up message
• cleaning up imports

version 2.3.1

Fixes

• fixing weighted meta-analysis parameterization

version 2.3

Features

• weighted meta-analysis by specifying cluster argument in RoBMA() and setting weighted = TRUE. The likelihood contribution of estimates from each study is down-weighted proportionally to the number of estimates in that study. Note that this experimental feature is supposed to provide a conservative alternative for estimating RoBMA in cases with multiple estimates from a study where the multivariate option is not computationally feasible.

version 2.2.3

Fixes

• updating the Makevars to install with R 4.2 and JAGS 4.3.1

version 2.2.2

Fixes

• updating the C++ to compile on M1 Mac

version 2.2.1

Changes

• message about the effect size scale of parameter estimates is always shown
• compatibility with BayesTools 0.2.0+

version 2.2

Features

• three-level meta-analysis by specifying cluster argument in RoBMA. However, note that this is (1) an experimental feature and (2) the computational expense of fitting selection models with clustering is extreme. As of now, it is almost impossible to have more than 2-3 estimates clustered within a single study).

version 2.1.2

Fixes

• adding Windows ucrt patch (thanks to Tomas Kalibera)
• adding BayesTools version check

version 2.1.1

Fixes

• incorrectly formatted citations in vignettes and capitalization

Features

• adding informed_prior() function (from the BayesTools package) that allows specification of various informed prior distributions from the field of medicine and psychology
• adding a vignette reproducing the example of dentine sensitivity with the informed Bayesian model-averaged meta-analysis from Bartoš et al., 2021 (open-access),
• further reductions of fitted object size when setting save = "min"

version 2.1

Fixes

• more informative error message when the JAGS module fails to load
• correcting wrong PEESE transformation for the individual models summaries (issue #12)
• fixing error message for missing conditional PET-PEESE
• fixing incorrect lower bound check for log(OR)

Features

• adding interpret() function (issue #11)
• adding effect size transformation via output_scale argument to plot() and plot_models() functions
• better handling of effect size transformations and scaling - BayesTools style back-end functions with Jacobian transformations

version 2.0

Please notice that this is a major release that breaks backwards compatibility.

Changes

• naming of the arguments specifying prior distributions for the different parameters/components of the models changed (priors_mu -> priors_effect, priors_tau -> priors_heterogeneity, and priors_omega -> priors_bias),
• prior distributions for specifying weight functions now use a dedicated function (prior(distribution = "two.sided", parameters = ...) -> prior_weightfunction(distribution = "two.sided", parameters = ...)),
• new dedicated function for specifying no publication bias adjustment component / no heterogeneity component (prior_none()),
• new dedicated functions for specifying models with the PET and PEESE publication bias adjustments (prior_PET(distribution = "Cauchy", parameters = ...) and prior_PEESE(distribution = "Cauchy", parameters = ...)),
• new default prior distribution specification for the publication bias adjustment part of the models (corresponding to the RoBMA-PSMA model from Bartoš et al., 2021 manuscript),
• new model_type argument allowing to specify different “pre-canned” models ("PSMA" = RoBMA-PSMA, "PP" = RoBMA-PP, "2w" = corresponding to Maier et al., in press , manuscript),
• combine_data function allows combination of different effect sizes / variability measures into a common effect size measure (also used from within the RoBMA function),
• better and improved automatic fitting procedure now enabled by default (can be turned of with autofit = FALSE)
• prior distributions can be specified on the different scale than the supplied effect sizes (the package fits the model on Fisher’s z scale and back transforms the results back to the scale that was used for prior distributions specification, Cohen’s d by default, but both of them can be overwritten with the prior_scale and transformation arguments),
• new prior distributions, e.g., beta or fixed weight functions,
• estimates from individual models are now plotted with the plot_models() function and the forest plot can be obtained with the forest() function,
• the posterior distribution plots for the individual weights are no able supported, however, the weightfunction and the PET-PEESE publication bias adjustments can be visualized with the plot.RoBMA() function and parameter = "weightfunction" and parameter = "PET-PEESE".

version 1.2.1

Fixes

• check_setup function not working at all

version 1.2.0

Changes

• the studies’s true effects are now marginalized out of the random effects models and are no longer estimated (see Appendix A of our manuscript for more details). As a results, arguments referring to the true effects are now disabled.
• all models are now being estimated using the likelihood of effect sizes (instead of test-statistics as usually defined). We reproduced the simulation study that we used to evaluate the method performance and it achieved identical results (up to MCMC error, before marginalizing out the true effects). A big advantage of using the normal likelihood for effect sizes is a considerable speed up of the whole estimation process.
• as a results of these two changes, the results of the models will differ to those of pre 1.2.0 version

Fixes

• autofit being turn on if any control argument was specified

version 1.1.2

Fixes

• vdiffr not being used conditionally in unit tests

version 1.1.1

Fixes

• inability to fit a model without specifying a seed
• inability to produce individual model plots due to incompatibility with the newer versions of ggplot2

version 1.1.0

Features

• parallel within and between model fitting using the parallel package with ‘parallel = TRUE’ argument

version 1.0.5

Fixes:

• models being fitted automatically until reaching R-hat lower than 1.05 without setting max_rhat and autofit control parameters
• bug preventing to draw a bivariate plot of mu and tau
• range for parameter estimates from individual models no containing 0 (or 1 in case of OR measured effect sizes)
• inability to fit a model with only null mu distributions if correlation or OR measured effect sizes were specified
• ordering of the estimated and observed effects when both of them are requested simultaneously
• formatting of this file (NEWS.md)

Improvements:

• priors plot: parameter specification, default plotting range, clearer x-axis labels in cases when the parameter is defined on transformed scale
• parameters plots: probability scale always ends at the same spot as is the last tick on the density scale
• adding warnings if any of the specified models has Rhat higher than 1.05 or the specified value
• grouping the same warnings messages together

version 1.0.4

Fixes:

• inability to run models without the silent = TRUE control

version 1.0.3

Features:

• x-axis rescaling for the weight function plot (by setting ‘rescale_x = TRUE’ in the ‘plot.RoBMA’ function)
• setting expected direction of the effect in for RoBMA function

Fixes:

• marginal likelihood calculation for models with spike prior distribution on mean parameter which location was not set to 0
• some additional error messages

CRAM requested changes:

• changing information messages from ‘cat’ to ‘message’ from plot related functions
• saving and returning the ‘par’ settings to the user defined one in the base plot functions

version 1.0.2

Fixes:

• the summary and plot function now shows quantile based confidence intervals for individual models instead of the HPD provided before (this affects only ‘summary’/‘plot’ with ‘type = “individual”’, all other confidence intervals were quantile based before)

version 1.0.1

Fixes:

• summary function returning median instead of mean

version 1.0.0 (vs the osf version)

Fixes:

• incorrectly weighted theta estimates
• models with non-zero point prior distribution incorrectly plotted using when “models” option in case that the mu parameter was transformed

Additional features:

• analyzing OR
• distributions implemented using boost library (helps with convergence issues)
• ability to mute the non-suppressible “precision not achieved” warning messages by using “silent” = TRUE inside of the control argument
• vignettes

Notable changes:

• the way how the seed is set before model fitting (the simulation study will not be reproducible with the new version of the package)