Installing R and JAGS on Windows

HOME I've previously posted on compatibility issues encountered when installing R and JAGS on Macs or Ubuntu, but Windows seemed immune to these problems ... until now!

The issue arises because current R versions, 3.3.0 or later, are not compatible with the current default installer for JAGS. Note that the problem is with the Windows installer, not JAGS source code, and doesn't affect other platforms. A compatible installer is available, but it's not the default. See Martyn Plummer's post for more details.

Install R

Download and install R statistical software from the CRAN web site.

Install JAGS

If you already have JAGS 4.2.0 installed, it's a good idea to uninstall the old JAGS 4.2.0 using Control Panel > Programs and features (Windows 7) or Settings > System > Apps & features (Windows 8, 10) and delete the old JAGS 4.2.0 folder in Program Files before installing the new binaries.

Download and run the new installer, "JAGS-4.2.0-Rtools33.exe". This is not the default on Sourceforge, so make sure you get the right file - you may have to search a bit. The default installer, "JAGS-4.2.0.exe", is not compatible with R 3.3.0 or later.

Install the rjags package

After installing R and JAGS, install rjags by opening R and typing


Boxes may pop up asking about personal libraries: click Yes. You will need to select a mirror web site near you, or choose 0-cloud, which is a world-wide network.

Once installed, check that it works by running


You should see a message telling you that rjags has linked to JAGS 4.2.0.


Starting with R version 3.3.0, a newer compiler is used to create the code actually used by the Windows operating system, the "binaries", the files you get when you download and run "R-3.3.x-win.exe". Specifically it uses the gcc 4.9.3 compiler for C and C++ code instead of gcc 4.6.3. Some packages contain C or C++ code, and they need to be compiled with the same compiler that was used for R. For packages on CRAN, this is taken care of automatically: when you run install.packages() or update.packages(), you get the package binaries that match your R version.

This applies to the rjags package too, it has to match the R version. But it also has to match with JAGS. JAGS source code is also in C++, and running the "JAGS-4.2.0.exe" installer gives you binaries from the old compiler. Binaries created with the new compiler are in "JAGS-4.2.0-Rtools33.exe".

If you are just using JAGS from the command line, not from R, none of this matters.

I have actually (inadvertently!) thoroughly tested R 3.3.0 with the old JAGS binaries on 64-bit installations and had no problems. But when I installed R 3.3.0 on a 32-bit machine together with the old JAGS binaries, calling library(rjags) caused R to crash: "R for Windows GUI front-end has stopped working."

Updated 10 June 2016 by Mike Meredith