Logistic regression significant interaction with categorical predictor


New Member
Hello everyone,

I am doing a binary logistic regression and a have some questions about presenting my data.

Depedent variable: Extensive rot (trees in groups will be coded 1 for rot or 0 for no rot)
Predictor variable: Group (1,2,3)
Covariates: Years since death, tree density (both continuous)

I found a significant interaction bewteen group and density. How do I find out what kind of an effect density has on the proportions in each group?

There is an option for post hocs to find where differences are (as groups are significant in this case), but I don't think it is valid seeing as my covariates are not homoscedastic. How else would I find out if there is significant differences between groups, while controlling for my covariates?