Re: [R] [R-sig-ME] lme nesting/interaction advice

From: Doran, Harold <HDoran_at_air.org>
Date: Mon, 12 May 2008 09:37:55 -0400


> But that avoids the question as to *why* it isn't very well
> set up for crossed random effects? What's the problem?
> What are the issues? The model is indeed bog-standard.
> It would seem not unreasonable to expect that it could be
> fitted in a straightforward manner, and it is irritating to
> find that it cannot be. If SAS and Minitab can do it at
> the touch of a button, why can't R do it?

I haven't followed this thread carefully, so apologies if I'm too off base. But, in response to Rolf's questions/issues. First, SAS cannot handle models with crossed random effects (at least well at all). SAS is horribly incapable of handling even the simplest of models (especially generalized linear mixed models). I can cite numerous (recent) examples of SAS coming to a complete halt (proc nlmixed) for an analyses we were recently working on. R (and Ubuntu) was the only solution to our problem.

Now, lme is not optimized for crossed random effects, but lmer is. That is why lmer is supported and lme is not really supported much. lmer is optimized for models with nested random effects and crossed random effects.

When working with models with nested random effects, and software optimized for those problems (e.g., HLM, SAS, mlWin) the variance/covariance matrix forms a special, and simple structure that can be easily worked with. This is not the case for models with crossed random effects.

Software packages designed for nested random effects can be tricked into handling models with crossed random effects, but this kludge is slow and really inefficient.

If you want complete transparency into the why and how, here is a citation for your review.

Best
Harold

@article{Doran:Bates:Bliese:Dowling:2007:JSSOBK:v20i02,   author = "Harold Doran and Douglas Bates and Paul Bliese and Maritza Dowling",
  title = "Estimating the Multilevel Rasch Model: With the lme4 Package",

  journal =	"Journal of Statistical Software",
  volume =	"20",
  number =	"2",
  pages =	"1--18",
  day =  	"22",
  month =	"2",
  year = 	"2007",
  CODEN =	"JSSOBK",
  ISSN = 	"1548-7660",
  bibdate =	"2007-02-22",
  URL =  	"http://www.jstatsoft.org/v20/i02",
  accepted =	"2007-02-22",
  acknowledgement = "",
  keywords =	"",
  submitted =	"2006-10-01",

}

R-help_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Received on Mon 12 May 2008 - 13:39:41 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Mon 12 May 2008 - 14:30:38 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive