[R] lmer function.

[Douglas Bates]

On Mon, Mar 31, 2008 at 7:13 AM, Boikanyo Makubate
<boikanyo at stats.gla.ac.uk> wrote:
> I am using the lmer function from the lme4 package.  I wrote the
>  following statement, specifying the method to be adaptive Gaussian
>  quadrature. I am getting an error saying "method = "AGQ" not yet
>  implemented for supernodal representation".  Please help.
>
>   > fit <-
>  lmer(response~beta1+(1|patient),family=binomial,method="AGQ",data=data.2)

As the message indicates, the AGQ method has not yet been implemented.

It is very tricky to implement the AGQ method.  For one thing it would
be infeasible for more than one grouping factor for the random effects
if these were non-nested.  The Laplacian method is the default method
in the development version of the lme4 package, available as

install.packages("lme4", repos = "http://r-forge.r-project.org")