grateful for any thoughts about this.
I'm doing a generalized linear model analysis of a binomial dependent variable (whether the participant gets the correct answer or not).
I have three independent variables A (two condition groups), B (two different categories), and C (a nine-level interval scale variable).
When testing for a three-way interaction of A+B+C, I get a significant result with an odds ratio of 0.5. However, how do I write this up to account for the scale variable? I'm tempted to say something like "When A is positive and B is positive, participants were fifty per cent less likely to respond with the correct answer for every level of C"
Or perhaps "For participants in the A=1 condition with the B=1 category, every additional level of C makes them fifty per cent less likely to respond with the correct answer."
But I'm not sure if i'm writing this up correctly? Thoughts?