[R] prediction for linear mixed model

[Ben Bolker]

array chip <arrayprofile <at> yahoo.com> writes:


> Hi, I am wondering if we can make prediction on a linear mixed model
>  by lmer() from lme4 package?

> Specifically I am fitting a very simple glmer() with binomial family
> distribution, and want to see if I can get the predicted probability like
> that in regular logistic regression?

> fit<-glmer(y~x+(1|id),dat,family=binomial)

> where y is the response variable (0, 1), and x is a continuous variable.
> I would like to get a predicted probability at a given value of x.

  See (1) code on http://glmm.wikidot.com/faq for this precise task,
and (2) the 'predict' method in the bleeding-edge development version
(lme4Eigen). 

  This question is probably slightly better suited for the
r-sig-mixed-models <at> r-project.org mailing list.