[R] Ordinal data - Regression Trees & Proportional Odds

[Volker Schlecht]

Dear Dr. Fieberg,

you used a regression tree approach to explore ordinal data set in addition to 
the proportinal odds model. I find this very interesting. I would like to know, how good the 
results of the regression tree approach turned out in comparison to the 
proportional odds model. Since people very often treat ordinal data as continuous, I would 
like to know how successfull this strategy is. I mean, I can see, what may be a reason
to do this: One could argue that ordinal data an a sufficiently large scale (say 1 to 10) are just 
coarsely measured continuous data. 

I would be delighted if you could tell me your experiences about the treatment of ordinal data with 
models for continuous data. Also I would be delighted, if you could tell me, if your results are published somewhere.

Best wishes,

Volker Schlecht