2

I have run an ordinal regression in R using polr from the MASS package. I'm trying to interpret an interaction. Both predictors are categorical. Here is an example:

Predicting success in the job market (Low, medium or high) from exam performance (not passed, passed) and sex (male, female).

The baseline categories are exam performance (not passed) and sex (male). There are main effects of both exam performance and sex. The interaction odds ratio is 2.04. This is displayed as:

  • exam_performancePassed:sexFemale: 2.04

How should this interaction be interpreted?

Amy
  • 95
  • 6
  • Women who passed their exam perform better than the sum of coefficients of being a woman and of passing the exam would suggest. Is that an answer or is the question more complex than I understand it? – Bernhard Oct 28 '19 at 09:39

0 Answers0