Dear readers,
First of all, let me thank everybody who was kind enough to try and help
me with my question.
My original question was:
> I would like to fit a mixed-effects model but with (for instance)
> binomial response and logit link function as in the glm-case.
> This does not seem to be possible with lme or nlme. Or is it?
> I have been told that this kind of thing is possible in other packages.
> Can anybody tell me how/whether this can be done in S-PLUS?
As several people remarked, there is no standard function in the current
version of S-plus that does fit such a model.
Bjarke Klein pointed me to the script reglm by Gordon Smith which does
seem to do the trick.
It can be found at: http://www.statdem.sdu.dk/~gks/s/index.html
Dennis J. Murphy pointed me to some DOS/Windows-programs bij Don Hedeker
that also perform special kinds of mixed-effects regression. They can be
found at: http://www.uic.edu/~hedeker/mixwin.html.
Regards,
Tobias Mueller
|