Introduction

The R-sig-meta-analysis mailing list is a mailing list for discussing the use of the statistical software package R for meta-analysis.

Discussions on the list often revolve around repeating topics. One can browse the mailing list archives or use search engines like Google or DuckDuckGo to find posts in the archives relevant to oneโ€™s interest. This still requires some digging.

The purpose of this page is to provide pointers to relevant posts pertaining to various topics. Note that links to relevant posts from the archives may take you to the first post in a thread, which typically raises a question/issue. Pertinent answers can then be found in the responses to this post.

Topics

What R packages are there for meta-analysis?

Take a look at the CRAN Task View: Meta-Analysis which provides a very detailed overview of R packages for meta-analysis and their purposes.

What to do when one cannot compute the sampling variances for some studies

Standard meta-analytic models use as input the effect size estimates and the corresponding sampling variances (or the standard errors). At times, it may be difficult to extract the information needed to compute the sampling variances / standard errors for some of the studies. Some possible approaches one can consider in this case are discussed here:

Meta-analysis of survival data (KM estimates)

The Freeman-Tukey double arcsine transformation and its back-transformation

Definition / computation of \(I^2\), \(R^2\), and \(I^2\) in meta-regression models