[R-sig-ME] Mixed Effects Model for Ordinal Data

[Rick Bilonick]

On Thu, 2008-07-31 at 08:27 -0400, Doran, Harold wrote:
> But, just in case it's not obvious, the current link functions for lmer
> don't support ordinal data. 
> 
> > -----Original Message-----
> > From: r-sig-mixed-models-bounces at r-project.org 
> > [mailto:r-sig-mixed-models-bounces at r-project.org] On Behalf 
> > Of Gregor Gorjanc
> > Sent: Thursday, July 31, 2008 4:33 AM
> > To: r-sig-mixed-models at r-project.org
> > Subject: Re: [R-sig-ME] Mixed Effects Model for Ordinal Data
> > 
> > Rick Bilonick <rab at ...> writes:
> > 
> > > 
> > > Is there an mixed effects R package that can be used to 
> > model ordinal 
> > > data?
> > > 
> > 
> > Hint: take a look at the help page for lmer i.e. type
> > 
> > ?lmer
> > 
> > and read about the family argument
> > 
> > gg
> > 
> > 

I had never noticed anything in lmer specifically for ordinal logistic
regression (like lrm in Design). I created separate logistic regressions
using lmer but it would be nice to be able to make it proportional like
lrm does.

I also looked at logitord from repeated but it is not very flexible
(only one level of nesting and one observation per time point per
subject as far as I can tell) and difficult to get to work.

Rick B.