[R-sig-ME] lmer and autocorrelation structures

Douglas Bates bates at stat.wisc.edu
Fri Sep 7 02:35:34 CEST 2007


On 9/6/07, Douglas Bates <bates at stat.wisc.edu> wrote:
> On 9/6/07, Douglas Bates <bates at stat.wisc.edu> wrote:
> > On 9/6/07, Chris O'Brien <obrienc at email.arizona.edu> wrote:
> > > Dear mixed-modelers,
> >
> > > I'm interested in modelling a dataset using generalized linear mixed models
> > > (specifying "family=binomial" with random effects) in which there is
> > > temporal autocorrelation in the response. Is there a way to do this in the
> > > new lme4 package?  I've used the "correlation" argument in the nlme library
> > > to specify a correlation structure for least-square means regression, but
> > > this option appears not to be implemented in the new lmer call in the lme4
> > > package.   Is such a thing possible (and appropriate) at this time?
> > > thanks,
> >
> > Possible, but only in the fortune("Yoda") sense.  In other words that
> > capability is not built into the lme4 package.
> >
> > I don't think we can comment on the appropriateness of the model
> > without more information and perhaps a look at the data.  However,
> > until someone writes the code to incorporate an autoregressive
> > structure into lmer, the question is moot.
>
> P.S. If the reference to fortune("Yoda") is too obscure, try
>
> install.packages("fortunes"); library(fortunes); fortune("Yoda")

P.P.S.  I have always felt that to qualify as a Yodaism, Simon's first
sentence should have been "R this is.", not "This is R."




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