Modular, Extensible, Transparent, Accessible, Bootstrapped Extraction For Systematic Reviews

The pkgdown website for this project is located at

In systematic reviews, extracting data from primary sources is a crucial step with a high potential for introduction of biases. This package facilitates specification of R extraction scripts that are simultaneously human- and machine-readable. In addition, they are modular and extensible, lending themselves well to living reviews where insights as to optimal extraction evolve over time. By being R Markdown script files, they are optimally transparant, and their structure was designed to also be accessible to readers without R. Finally, each extraction script contains the original specification, enabling bootstrapping new specifications from each single extraction script.

Originally, the package was intended to help with everything that needs to be done before the metafor package can be used, hence the name. Since then, the metaverse was born, with dedicated packages for specific stages, such as developing search strategies, working with bibliographic data, and visualisation. Presently, therefore, metabefor mostly covers extraction.

Getting started

Depending on your situation there are three places to get started.

If you’re already familiar with systematic reviews and evidence synthesis, you may want to start with NITRO, the Narrated Illustration of a Transparent Review Outline. This is somewhat of a worked example, available at

Second, if you’re relatively new to systematic reviews, you may want to start with the SysRevving book, available at This living open access book is intended as a general resource on conducting systematic reviews, built around metabefor.

Third, if you’re already using metabefor, you may just want to consult the manual as included with the package. From within R, each function’s manual page can be displayed by prepending the function name with a question mark (e.g. ?metabefor::rxs_fromSpecifications. Alternatively, the index of functions is available at the metabefor PkgDown site

Finally, in any case you may want to keep the glossary of terms ready. This is located at


You can install the released version of metabefor from CRAN with:


You can install the development version of metabefor from GitLab with:


(assuming you have remotes installed; otherwise, install that first using the install.packages function)

Other useful resources

For an extensive resource about effect size computation and conversion, see