Normally, we carry out chi-square test for NOMINAL data while T-TEST for numerical data. But I happened to meet the problem when doing a paired sample test:
I have to carry out a test on NOMINAL data, etc Y or N on something , but It is weighted. For example, for the occurrence of A and B, the weight for A is 1.5 and for B is 1.0.
So I should do a chi test or a t-test instead?
Advertise on Talk Stats