These tutorial examples illustrate
a variety of methods involved in logistic regression analysis.
A few additional examples relate to other aspects of categorical
data analysis.
Arthritis treatment data
The response has 3 categories: No improvement, some improvement,
marked improvement. A first analysis treats the response as
binary (Improved vs. Not Improved); a second analysis
uses the proportional odds model for the 3-category response.
Women's labour force participation, Canada 1977
Fits several logistic models to predict whether a woman
works full-time, part-time, or does not work outside the home.
Compares the proportional odds model to the fitting of nested
dichotomies for the polytomous response.