I've been giving the task of working on data that is not mine.
I have to compare two independent populations for two different continuous variables using a classification variable. The continuous variables both have to be weighted by a real number weighting variable.
Here is the problem: I cannot use the FREQ statement on PROC NPAR1WAY.
Can anybody suggest another method for testing these two populations. They are not normally distributed. I thought of transforming the variables to normality (if that's even possible), but I wasn't sure what to do about the weight variable then.