Chapter 10: Ordinal Logistic
Regression
If there is no inherent ordering of the outcome
categories, a polytomous regression model is
appropriate. If there is an inherent ordering of
the outcome categories, then an ordinal logis-
tic regression model may also be appropriate.
The proportional odds model is one such ordi-
nal model, which may be used if the propor-
tional odds assumption is met. This model is
discussed in Chap. 10.454 12. Polytomous Logistic Regression