What do you want to do exactly? Test whether adding the age*group
interaction is significant? If so, then you can just run model2 with a
contrast looking at that term (or those terms). BTW, it would seem that
the model they describe would be ill-conditioned because age*b2 and
group*age*b3 woul
Hi,
i read the following in a paper and would like to do the same:
"At each vertex in each hemisphere, the proportion of total variance in
cortical thickness (CT) accounted for by 2 linear regression models was
compared: One that did not include an age-by-group interaction term, Model
1: [CT = in
