I've got two types of variables - nominal (mode choice (e.g. car, bike, etc...) and interval (income range (e.g. 10,000-15,000, 15,001-20,000, etc...). I am trying to figure out the best test to see if the variables are correlated, and then I want to test to see if the correlation is significantly different two years later.

Do I use a chi-squared test for the correlation, and then do a Fisher z-transformation and t-test to test for difference????

And what should I do if a third variable (nominal - area (e.g. marina district, downtown, etc...) is thrown into the mix?

Any help would be great.

Thanks!