Multiple comparison test for qualitative variable

[#1]

Hello everybody,

I would like to know if it exists an equivalent of the Tukey's test but for a qualitative response (i.e . after running a nominal or ordinal logistic regression).

Thanks

[#2]

Hi @Raaed 

Thanks for your answer but actually I am looking for something similar with a nominal or ordinal response, instead of a continuous one, like in your example.

Any idea ?

[#3]

Please look at this portion of the JMP Help to see if this approach is useful.

[#4]

Hi @Mark_Bailey ,

I didn't know this option which can indeed be very useful!

However if I understand correctly it works only if the response has to levels. Moreover in my case,  what I have in mind is to compare pairwise the levels, but I don't know if it is feasible :)

[#5]

Yes, this method depends on a dichotomous response. Yes, it compares each level to the overall proportion, not one level to another level.