List Members,

I sent this message a few days ago and received a response that said it was not distributed, so I thought I would try again.

I have a sample size of 73 organizations and I am using logistic regression. Dependent variable (1) n=22 and (0) =n=51

My question is how one determines the significance of an interaction. I am wondering about having trouble interpreting the significance of an interaction between a dichotomous variable G and a nominal one (Group membership A, B, C).

When I stratify by group I get these results:

Group A OR=0.60 (0.53, 6.80) p=0.680 Group B OR=2.00 (0.78, 51.59) p=0.676 Group C OR=13.50 (1.10, 165.97) p=0.042

However, the significance of the interaction term in the SPSS for Group C*variable G is 0.081

Is this interaction "significant?"

Thanks! Chris Cutshaw