Penalized Meta-Analysis

Conduct penalized meta-analysis, see Van Lissa, Van Erp, & Clapper (2023) . In meta-analysis, there are often between-study differences. These can be coded as moderator variables, and controlled for using meta-regression. However, if the number of moderators is large relative to the number of studies, such an analysis may be overfit. Penalized meta-regression is useful in these cases, because it shrinks the regression slopes of irrelevant moderators towards zero.

pema: Penalized Meta-Analysis

Conduct penalized meta-analysis (“pema”) In meta-analysis, there are often between-study differences. These can be coded as moderator variables, and controlled for using meta-regression. However, if the number of moderators is large relative to the number of studies, such an analysis may be overfitted. Penalized meta-regression is useful in these cases, because it shrinks the regression slopes of irrelevant moderators towards zero.

Where do I start?

For most users, the recommended starting point is to read the paper published in Research Synthesis Methods, which introduces the method, validates it, and provides a tutorial example.

Installing the package

Use CRAN to install the latest release of pema:

install.packages("pema")

Alternatively, use R-universe to install the development version of pema by running the following code:

options(repos = c(
    cjvanlissa = 'https://cjvanlissa.r-universe.dev',
    CRAN = 'https://cloud.r-project.org'))

install.packages('pema')

Citing pema

You can cite pema using the following citation (please use the same citation for either the package, or the paper):

Van Lissa, C. J., van Erp, S., & Clapper, E. B. (2023). Selecting relevant moderators with Bayesian regularized meta-regression. Research Synthesis Methods. https://doi.org/10.31234/osf.io/6phs5

About this repository

This repository contains the source code for the R-package called pema.

Contributing and Contact Information

We are always eager to receive user feedback and contributions to help us improve both the workflow and the software. Major contributions warrant coauthorship to the package. Please contact the lead author at [email protected], or:

• File a GitHub issue for feedback, bug reports or feature requests
• Make a pull request to contribute your code or prose

By participating in this project, you agree to abide by the Contributor Code of Conduct v2.0. Contributions to the package must adhere to the tidyverse style guide. When contributing code, please add tests for that contribution to the tests/testthat folder, and ensure that these tests pass in the GitHub Actions panel.