Dear list,
After reading many threads I am still having some questions about the tbss analysis.
I have 3 groups (control, disease 1, disease 2)
The aim is to see the differences in FA between
Control v.s. Disease 1
Control v.s. Disease 2
Disease 1 v.s. Disease 2
I used design_ttest2 to generate the design.mat and .con for randomise, the results seem to be as expected.
Now I want to regress out the age effect for all comparison, and as described in the list, I plan to include the demeaned age as the 3rd EV in the design.mat (the age is unmatched for all). Is it a must to run the permutation test again for all? (sorry for the silly question, but it takes pretty much time to run a iteration of 5000 times). After regressing out the age effect, I would also like to correlate with the neuropsycological scores for the disease group. Is it possible to set up the design.mat which includes all the scores (5 different EV) as well as regressing out the age influence?
The second question is about the difference between running ANOVA & 2sample unpaired t-test.
If there a difference if running one time ANOVA for all , and running 3 times the two sample unpaired t-test?
Thanks in advance
Natalia
|