[R-sig-ME] question about linear mixed models

Douglas Bates bates at stat.wisc.edu
Thu Jul 8 14:59:18 CEST 2010


On Thu, Jul 8, 2010 at 6:56 AM, Kevin Wright <kw.stat at gmail.com> wrote:
>> lme4 can handle crossed and nested random effects whereas nlme can
>> only do nested random effects.
>>
> Actually, nlme *can* do crossed random effects.
>
> In the book by Pinheiro and Bates, see the Cell Culture
> Bioassay example starting on pg 163 for a demonstration
> of fitting crossed random effects using pdIdent and pdBlocked objects.

Yes, but that is a horrible kludge and would not be suitable for cases
with more than a few levels for each of the grouping factors.

It is much better to use lme4 when fitting models with crossed random effects.




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