Dear Statalists,

I have 4 alternatives ( let's say A B C and D ). I thought earlier a multinomial logit model would be more sensible since it is just one model. However, the various alternatives 1-4 are not mutually exclusive. If possible I should test for equality of the impacts of key variables across the 4 types. It’s more complicated when they are from separate models. So, my question is what are the other solutions if my alternatives are not mutually exclusive but I want to run them in the same model ??

Thank you for your time and consideration.