I want to check for associations between categorical variables before using some of them as predictors in a logistic regression. Normally I would simply use chi-square, however I have about 15 variables.
What is the best way to test for associations between pairs of all 15 variables?
If you really want to test pairwise associations than you should perform any test in pairwise manner. In such situation there are no problem with chi-squared test. However, I suppose you want to score and compare sizes of association but not their significances. So there are several different measures for that purpose. Here is a useful discussion about them.
