[R-sig-ME] Ordered Probit/Logit with random coefficients

[Joshua Wiley]

Hi Andreas,

I agree with David---you may have best results with a Bayesian
approach, such as in
MCMCglmm.  The package author, Jarrod Hadfield, has a great set of
course note available online:
http://cran.r-project.org/web/packages/MCMCglmm/vignettes/CourseNotes.pdf

You may also find this package or at least some of the examples useful:

https://github.com/JWiley/postMCMCglmm

I wrote it specifically after doing a bunch of work with ordinal
models in MCMCglmm.  If you install it, you can run:

demo("randomMCMCglmm", package="postMCMCglmm")

to see an example of a three level ordered outcome with a random
intercept.  The (not run) examples in

?predict2.MCMCglmm

may also be useful to you.

If you do not want to install the package, you can get the compiled
PDF of the manual here:
https://github.com/JWiley/postMCMCglmm/blob/master/inst/doc/postMCMCglmm-manual.pdf
with the examples I mentioned, and the demonstrations in the "demo"
folder: https://github.com/JWiley/postMCMCglmm/tree/master/demo


More examples and a vignette should be coming at some point.

Cheers,

Josh


On Mon, Jan 21, 2013 at 12:41 PM, David Duffy <David.Duffy at qimr.edu.au> wrote:
> On Mon, 21 Jan 2013, Andreas Karpf wrote:
>
>> I searched everywhere but I didn't find what I want, that is why I as the
>> question here. Does anybody know of a function in R which allows to
>> estimate ordered probit/logit model with random coefficients.
>
>
> Eventually, someone will mention MCMCglmm with family="ordinal"
>
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models



--
Joshua Wiley
Ph.D. Student, Health Psychology
Programmer Analyst II, Statistical Consulting Group
University of California, Los Angeles
https://joshuawiley.com/