 |
 |
Hi Anderson, Thanks for the response. Now let me complicate things even more: If I only want to do three levels of analysis, and I have 100 subjects, I COULD on the third level have 200 inputs (100 copes for the single task, and 100 copes for the dual task), make EV1 and EV2 appropriately to separate Single and Dual Copes, and then define the D>S contrasts with these EV's. But I also would like to make comparisons of Old Vs Young. So, in this scenario, I could set up the third level with two groups, one for young and one for old. Each group would have a dual EV and a single EV. In total I'd have 4 EV's and I'd make the contrasts accordingly again, though that could get a little complicated. The other option, that I'm leaning towards for ease of use, is to break this up into 4 levels. The third level would be a 2 group model, running on each second level cope directory, one for the Single Task and one for the Dual Task. There would only be 2 EV's, one for each age group. Then for the contrasts, I could make an "All" and "Young Greater than Old". Then, I could do a 4th level, with the "single" and "dual" gfeat directories as inputs. It would automatically run the analysis on the copes within, so I could make the EV weight 1,-1 and a single contrast. This should give me the dual greater than single for all subjects and for young greater than old, correct? Anyways, do you think both of these methods would produce the same (or very similar) results?