Hi,
I would really appreciate it if someone could help.
I have a question related to the simple main effect in fMRI spm.
My experiment is a clinical intervention study. I used *2(patient vs. healthy control)2 (pre vs. post) ANOVA and found an interaction. I thus want to do the simple main effect. However, I found three ways to do so, with different answers for each way. I am very confused.
The first way to do a simple main effect is:
I select the data from two cells (e.g. patients in pre and patients in the post) and do a pair-t test on the 2-level. Then the result comes.
The second way is:
I use the 2-level result as the same as interaction and set the contrast. e.g. the contrast would be [1 -1 0 0] when comparing patients in post and patients in pre, the other two 0 would be control pre and control post.
The third way is:
Since I have a peak (or peaks) within the interaction. I extract each beta value for 4 cells within that peak (e.g. right superior parietal lobe in patient pre, patient post, control pre, control post). Then I have these beta values and do ANOVA to see where the simple main effect lies.
My question is, I see three different answers.
For example, in control post-control pre, I see the inferior temporal lobe in the first way, see inferior parietal lobe in the second way, and both in the third way. (all with p=0.01, k=10, the same as the interaction)
What is the correct way to do a simple main effect? Are the any references I can refer to?
Thank you very much for your reading.
Ruby