Confidence intervals not included for polr about ggeffects

[strengejacke]

The predict()-function in MASS does not calculate CI. For some objects, like merMod, I use this code - however, this does not seem to work for polr-models, at least not always reliable.

If you have any ideas, I'm happy to implement CI for polr as well, but currently I don't know how to do this.

[stefan-mueller]

Thanks for the quick response!

I had a look at your overview of ordinal regressions in R. Does clm() (ordinal), lrm() (rms) or vglm() (VGAM package) include an option to calculate CI? As far as I see these models are not implemented in ggeffects, so I do not want to cause too much work. If I can help, please let me know. Thanks again for your great work on the package(s)!

[strengejacke]

Support for clm is implemented - I hope this is sufficient? Will commit tonight, but note that you need to install sjstats from GitHub as well in order to use ggpredict() with clm-objects.

lrm-objects are also implemented, but only for binary outcomes.