[R-sig-ME] lme4 heteroscedasticity???

[Douglas Bates]

On Mon, Oct 13, 2014 at 10:24 AM, Paul Johnson <paul.johnson at glasgow.ac.uk>
wrote:

> You can model both heteroscedasticity and crossed random effects in nlme.
> Crossed random effects can be handled in nlme using pdBlocked and pdIdent.
> See p163-6 of Pinheiro, J.C. & Bates, D.M. (2000). Mixed-Effects Models in
> S and S-PLUS. Springer, New York.
>

But it will be a very slow and memory-intensive process unless one of the
grouping factors has very few levels.  The nlme package does not use sparse
matrix methods, nor does it use a profiled objective criterion.

In other words, yes it is possible but no, it is not easy to do so this way.


> On 13 Oct 2014, at 16:12, Douglas Bates <bates at stat.wisc.edu> wrote:
>
> > It is best to send questions like this to the
> > R-SIG-Mixed-Models at R-Project.org mailing list, which I am cc:ing on this
> > reply.  Several of those who read that list can respond more quickly
> than I
> > am able to.
> >
> > As far as I know there is not yet the capability in lme4 to model
> > heteroscedasticity in the distribution of the response given the random
> > effects.
> >
> > On Mon, Oct 13, 2014 at 6:01 AM, Košmelj, Katarina <
> > Katarina.Kosmelj at bf.uni-lj.si> wrote:
> >
> >> Hello,
> >>
> >> I am analyzing a mixed model with three crossed factors, two random
> >> (sample, laboratory) and one fixed (method); the response variable is
> the
> >> number of somatic cells in milk.  The main question is: is the
> precision of
> >> the means of the three method is comparable? Therefore, I would like to
> >> compare a model with different variances for the methods with the model
> >> considering the same variance for the methods.
> >>
> >>
> >>
> >> In nlme, this is feasible, however, two crossed random factors can not
> be
> >> tackled, this can be analyzed with lme4.
> >>
> >>
> >>
> >> In nlme, the problem of heteroscedasticity if solved. Is this problem
> >> solved in lme4 yet?
> >>
> >> Do you have any suggestion how to deal with this problem?
> >>
> >>
> >>
> >> Regards,
> >>
> >> Katarina
> >>
> >>
> >>
> >
> >       [[alternative HTML version deleted]]
> >
> > _______________________________________________
> > R-sig-mixed-models at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
>
>

	[[alternative HTML version deleted]]