Hello again,
I have some additional questions after going through the MBA script from AFNI. I suppose this question might be more adequate for the stan forums, but if you could help that would be much appreciated. I came up with the following model for my analysis:
Y ~ 1 + group_dummy1 + group_dummy2 + (1 | ROI) + (1 + group_dummy1 + group_dummy2 | mm(Subj1, Subj2, weights = cbind(1, 1), scale = FALSE))
Would this be correct for my situation? My understanding is that I could not run the model with a single grouping variable, but I don't understand why not (example below).
Y ~ 1 + group + (1 | ROI) + (1 + group | mm(Subj1, Subj2, weights = cbind(1, 1), scale = FALSE))
Similarly, if I were to include a variable like age and wanted to look at the age by group interaction (let's assume two groups for simplicity here), from the MBA documentation I see that I should code a variable for that (group_age, corresponding to the first model below), but I don't understand why it could not be coded as the second model below:
First model:
Y ~ 1 + group + age + group_age + (1 | ROI) +
(1 + group + age+ group_age| mm(Subj1, Subj2, weights = cbind(1, 1), scale = FALSE))
Second model:
Y ~ 1 + group*age + (1 | ROI) +
(1 + group*age | mm(Subj1, Subj2, weights = cbind(1, 1), scale = FALSE))
Finally, if I wanted to include a within-subject factor of socialness (social and non-social videos), where I want to look at the ISC within each condition, the difference between conditions within each group, and the interaction of group by socialness, would the model below be the correct one (once again assuming 2 groups only)? Or should I include socialness by itself as well (i.e. group*socialness)? Or should I really stick with running separate models for each of the within-subject contrasts?
Y ~ 1 + group + group:socialness + (1 | ROI) +
(1 + group + group:socialness | mm(Subj1, Subj2, weights = cbind(1, 1), scale = FALSE))
Thank you so much for your help again!
Best,
Vasco