Doing the t-tests outside the ANOVA, is it similar to doing post-hoc tests within an ANOVA? How does it affect the error terms?
As I understand, it is not an easy fix to implement the ability to do contrasts within the 3dANOVA3. I guess what I am getting at is the statistical implications, if there are any, of doing t-tests outside the ANOVA. I assume there are none???
Gang was implying that the 3dANOVA2 type 3 would afford me more power? My understanding is that by using a mixed model, it accounts for or controls for subject variability?
Could you please comment
Thanks
Jen
