[R-sig-ME] Lme4 query

[Douglas Bates]

On Mon, Aug 3, 2009 at 10:57 AM, D'Arrigo J.<J.DArrigo at soton.ac.uk> wrote:
> Dear Douglas and Martin,

> I am using the 'lmer' function to fit a random-effect model for a binary outcome. I aim to use the estimated random effects (ranef) from the model to make some weight adjustments. However, I am first interested in understanding how your function in R compute these random effects estimates. I have been trying to find some literature but unsuccessful. I would appreciate your help here.

Technically the values returned by ranef are not parameter estimates
because the random effects are not parameters.  They are an unobserved
set of random variables and the values returned by ranef are the
conditional modes of these random variables given the observed data
and the estimated values of the parameters in the model.

I'm not sure of what to suggest as a reference for this.  Perhaps you
could send a message to the R-SIG-Mixed-Models at R-project.org mailing
list, which I have cc:d on this reply, to see if someone reading that
list is able to provide a reference.