Equal Odds

#1
I'm trying to figure out how I can use logistic regression to test the hypothesis of equal odds in a 2 x 2 table of counts? Any ideas? I know I have to use an indicator variable, would I use this for the levels of the explanatory group? And what about the response category? Thanks!:wave:
 
#2
If you've got a 2x2 table, both your predictor and response will be indicator variables. It's very straightforward. A simple logistic regression. One categorical predictor.

You'll have a single odds ratio with a Wald statistic. The p-value will test the null hypothesis of equal odds. That corresponds to an odds ratio of 1.

Karen