I have two ordinal dependent variables, each having three response levels. You can use an ordered logit or probit model for such data if you have one dependent variable. I've seen some papers about multivariate ordered regression, and wonder if there are prepackaged functions in any of the usual stats software environments to do this. I am most proficient in R and Stata. Thank you.
closed as off-topic by mdewey, kjetil b halvorsen, gung♦, John, whuber♦ May 5 '17 at 18:08
This question appears to be off-topic. The users who voted to close gave this specific reason:
- "This question appears to be off-topic because EITHER it is not about statistics, machine learning, data analysis, data mining, or data visualization, OR it focuses on programming, debugging, or performing routine operations within a statistical computing platform. If the latter, you could try the support links we maintain." – mdewey, kjetil b halvorsen, gung, John, whuber
I am not aware of any default functions in R that run cumulative link models. Also, to the best of my knowledge, there is no subtle difference in extending univariate models to multivariate ones as far as the details of computation or interpretation of effects; everything is exactly analogous to other R linear models: use a formula function, interpret effects, and call it good. Unless you provide examples of which code you're using, I'm not too sure your problem is really a problem!
MASS package has
polr which is very nice. I am fond of the
lrm function in Frank Harrel's
rms package as it conveniently calculates sandwich based robust errors.
It's an old question but I bumped onto it.
Packages in R are "ordinal" by Christensen. There used to be polr but that does not work with recent R versions.
I'd use ordinal which is quite convenient with good documentation.