Hi,
Now I am using trait rating for face data analysis and our data has 12
face categories. First I classified the 12 face categories into three
groups. If we focused on one group only,to test if some regions are
correlated with trait ratings in this specified group, we used two
regressors for each group (one for mean of ratings, one for variation of
ratings, 6 regressors in all). Is it right? Or should I set zero for the
other two variational regressors of other two groups, in other words, only
4 regressors now?
Thank you!
Xiaoyun Liang