Welcome to lme4 - Mixed-effects models project!

An R package for fitting and analyzing linear, nonlinear and generalized linear mixed models.

Resources

• A guide to changes in the upcoming release of lme4 (Markdown source, downstream package compatibility report)
• The project summary page, including links to the source code browser, bug and feature request trackers, compiled package files, etc.
• Doxygen documentation of the underlying C functions
• References to articles and other research using nlme or lme4, or the corresponding BibTeX file. (If you would like to add your work to this database, please email vasishth.shravan at gmail dot com.)
• Slides from short courses on lme4
• Chapter drafts of the book lme4: Mixed-effects Modeling with R

Installation

Binary and source packages of recent versions of the development version (lme4, previously lme4Eigen) are available here, as well as MacOS binaries for RcppEigen. The current status for the various versions is as follows:

• The stable release of lme4 on CRAN (version 0.999375-42) is the last version of this branch.
• R-forge has lme4.0 (version 0.9999-1), which is the continuation of the old branch. It should be perfectly compatible with the current CRAN version, except for its package name.
• R-forge also has a new version of lme4 (version 0.999902344-0), which is as far as possible externally compatible with the old branch, but its internal structures have changed, so packages dependent on the old structures will need to be updated to work with the new release. The new version also produces objects of class merMod rather than class mer.
• An older development branch, lme4a, is also available on R-forge, but we believe that it is now dominated by the new development branch (i.e. the r-forge version of lme4), and we strongly request that you work with the new version. If you encounter a situation where lme4 can't do something that lme4a can, please inform the maintainers.
• Although installing packages into older versions of R is sometimes possible by downloading and manually installing the package files, installation will generally work much more smoothly if you are running the latest version of R. If you want help installing lme4 to an older version of R you will have to provide a very good reason ...
• You may have to install dependencies (e.g. RcppEigen and minqa for lme4) from other repositories first.
1. If possible, you should install the most recent version of the development version built on r-forge, e.g.

   install.packages("lme4",repos="http://r-forge.r-project.org")
   

   .
2. If the appropriate version for your OS is not available (e.g. binaries for MacOS), the next best option is to install the most recent version available at the lme4 repository, e.g.

   install.packages("lme4",repos="http://lme4.r-forge.r-project.org/repos")
   

   These versions will generally be slightly less up-to-date than the ones built by R-forge, because they must be updated by hand.
3. Alternatively (e.g. if you have network problems), you should be able to install these versions by descending through the directory hierarchy from here, downloading a copy and using

   install.packages("[name_of_file_here]",repos=NULL)
   

   or a menu item such as "install from local zip file"
4. The most reliable way to install the package is to check out the most recent copy and build it from source on your machine; however, this may require extra tools (Subversion client, compilers, etc.). Alternatively, if you have the compiler tools but don't want to mess with Subversion, you will often get farther with install.packages(...,type="source") than trying to find a binary package compiled for your OS.
5. If all else fails, please contact the maintainers.