The R-package metagen

Introduction

This is an R-package for inference in the random effects meta regression model. The purpose of the package is two fold. On the one hand, it allows the user to run different inferential methods on any given data that follow a meta analysis or meta regression model. The output is a neat summary of all kinds of different methods, many of which are commonly used in the meta analysis or meta regression context and some which are not as common, too. The big feature are two methods based on generalised inference principles.

On the other hand, the package allows to run rather sophisticated large scale computer experiments that evaluate the performance of all implemented methods with respect to coverage probability (and, well, average length). The idea is to use the above procedures to trace out the possible magnitude of the unknown parameters of the model, and, then, to use the possibility of running computer experiments in a possible sub-space of the parameter space to get an idea of which method yield, in fact, trustworthy results.

Installation from CRAN

The easiest way to install the package in a running R environment is:

install.packages("metagen")
library("metagen")

Installation using devtools GitHub-functionality

If you have devtools installed on your system, you can install the latest up-to-date version of metagen directly from within R.

install.packages("devtools")
devtools::install_github("metagen")
library("metagen")

Installation manually from GitHub

After cloning the repository, you may use the following magic in the build directory to build and install the package on your system:

make package
make install

Then open up an R-session and type:

library("metagen")

And you are good to go.

Where to start

Start to get a feel for the functionality of the package by having a look at the help files of the functions metareg and metagen.

?metareg
?metagen

This should get you started very quickly.

Author

Thomas W. D. Möbius