I have a multivariate unordered logistic regression with four independent variables (bureau, democracy, africa, wartype) and a categorical, unordered dependent variable with
