[R-sig-ME] convert mer to merMod?

Ben Bolker bbolker at gmail.com
Fri Jan 31 15:01:39 CET 2014


On 14-01-31 07:12 AM, Matthew Husband wrote:
> Dear Ben Bolker-
> 
> Sorry to trouble you with this – it looks like the new lme4 package has
> moved on to a new class for lmer().  However, I have some older analyses
> saved that are class mer.  Is there a way to convert them so that I can
> use them with the the standard functions (I.e. summary(), etc.).  It
> will be a great pain to try to rerun the analyses…
> 
> Best,
> -Matt
> 
  [cc'ing back to r-sig-mixed-models]

  In principle this could be done (and it would be handy to have), but I
think in practice the easiest thing to do is to install lme4.0, the
backward-compatibility version of lme4; this is available from the
repository at http://lme4.r-forge.r-project.org/repos, i.e.

  install.packages("lme4.0",repos="http://lme4.r-forge.r-project.org/repos")

(if you don't have the dependencies installed already you have to do
something like repos=c("...",getOption("repos")), but lme4.0 has a
fairly short dependency list and you should probably have them already
if you've already been using lme4).

  This also has the advantage of signalling that you're using an older
version of lme4, causing less confusion if someone else gets slightly
different results with a newer version.

  You will also need to have a look at the convert_old_lme4() function
in the lme4.0 package.

  Ben Bolker



More information about the R-sig-mixed-models mailing list