Hi Kaitlyn,

I'm not sure that that is the best approach. My first suggestion is that you first ask whomever planned the experiment; that person should have a clue on what the hypothesis was, and the best way to assess it.

In any case, I don't think you need a repeated measurements ANOVA. Instead, model the 5 conditions from on each session at the 1st level for each subject, including the interaction EVs (see the options in FEAT when assembling the design), and add the contrasts there. At the higher level, test the resulting COPEs from the 1st level, without the need to consider repeated measures. A simple one-sample t-test should do the trick then.

All the best,

Anderson

On 6 May 2014 23:38, Kaitlyn Breiner <[log in to unmask]> wrote:

Hi FSL Users,

I'm currently trying to model two IVs, IV1 has three levels (positive, same, negative); IV2 has two levels (social, control). Each participant (there are 19 so far) experiences all levels of both IVs in one scan. I'm trying to determine the best way to set up the design file/s for comparing the main effects of each IV plus the interactions, seeing as there is no between subjects factor. So far, I've done the following:

Level 1s include 5 design files for each subject, one for each level
Level 3s include 19 inputs (every subject) for a higher-level group analysis (Mixed effects: Flame 1) for each level

I'm wondering at this point how to conduct group contrasts to yield main effects and interactions. I'm unsure whether: 1) I conducted the Level 1s or 3s correctly (would a tripled t-test or something akin to the 3 factor, 2 level design be more appropriate?); 2) whether the next step would be to enter every subjects' input for each level independently to do a group comparison; 3) or whether this isn't possible at all.

Any feedback or direction would be greatly appreciated!

Thank you,

Kaitlyn

--

Kaitlyn S. Breiner, M.A.
Ph.D. Student
Developmental Psychology
University of California, Los Angeles

1285 Franz Hall, Box 951563
Los Angeles, CA 90095

[log in to unmask]