Hi all,

I'm conducting an ROI-ROI analysis and would like to compare 2 groups. Data was collected from 4 different study sites, so I would like to include the study site as a covariate, and see if there is a group x site interaction term. I have 2 questions about how to do this:

To include study site as a covariate, should I create one "Site" variable with 4 levels and enter [1 -1 0] for Group 1, Group 2, and Site, respectively? Or, should I create 4 separate site variables (Site 1, Site 2, Site 3, Site 4) and enter [1 -1 0 0 0 0] for Group 1, Group 2, and Sites 1-4 respectively in the second-level model? Is there a difference between these?

To look at the interaction between group and site, am I correct in thinking I will need to create a new variable which multiples Group * Site? I see I can do  that by selecting "Creation interaction(s) of selected covariates).". In this case, do I need to create 1 variable for Group (with 2 levels), or keep variables for Group 1 and Group 2 separately? And do I then enter [1 -1 0 0] to look at effects of Group 1 vs. Group 2 while covarying for site and site*group interaction?

Thanks for any help,
Alex