Dear FSL experts,
Sorry for asking this again.
Should the age variable be demeaned within groups or across all subjects before splitting into two EVs, if I want to test the interaction between groups and age (i.e. slope1>slope2 or slope1<slope2)? I'm rather confused by different descriptions.
when testing the correlation between fa values and behavior scores regardless of groups, should i add one EV with 1s in the design matrix? I got different results with and without an extra EV, so i really need a final decision.
Any help is needed and appreciated! Thanks.
Sincerely,
Qin
|