This is sorta a homework question, but i have been trying myslef to understand it and don't have anyone around to help me through it.
So i've got a data set with x amount of variables and i want to do a chi squared test to determine if 2 of those variables are related using the other variables to control for confounding.
The question is whether i need to do regression analysis on the other confounding variables given that so far all those variables are very similar (statistically so, P=0.07 is the most significant; P=0.7 the least).
I would imagine that if all the variables are pretty similar between the two groups, regression wouldn't be necessary, but i'm afraid i'm missing something.
Thanks for any help.