Help! Repeated Measures ANOVA SPSS

[kristall]

Context of study:

32 participants completed a self-esteem questionnaire (scores can go from 0-40). Then they are scanned in a fMRI so we can see their brain activation. On one day, they drink alcohol. On another day they drink placebo.

In my SPSS sheet, I have 32 subjects. Each has a self-esteem score. Then I have variables for 5 brain regions. For each region, I have a fMRI activation score on placebo day and an activation score for the alcohol day.

We think that the level of self-esteem somehow impacts the level of brain activation.

What I did so far in SPSS:
Analyze -> General Linear Model -> Repeated Measures.
Then I correctly input in all 5 of my brain regions in the "Within-Subjects Variables".

Questions:

I do not know where to place the self-esteem scores. Do they go in the "between-subjects factor box" or the "covariates box"?

When I click the "options" button, what do you recommend I add to the analysis? And what do those options mean?

I'm a newbie to statistics. Your help is appreciated!

 

[victorxstc]

From your research goal, it seems to me your analysis is not completely correct. Please explain your research better, with more details and more organized. So you have 30 patients with known self-esteem scores. You are sure that self-esteem does not change under the influence of alcohol, so you just measure their self-esteem once. Right? It appears to me that self-esteem can alter by alcohol consumption and you have probably measured their self-esteem once when sober and once when drunk, right? ...

Ok each participant is scanned twice (once with placebo and once with alcohol), right? Then you want to assess if some brain regions get more active under the influence of alcohol or not, right? But you ALSO think that "the level of self-esteem somehow impacts the level of brain activation."... So you have assessed self-esteem only once I think...

My understanding is that you should run a repeated-measures MANCOVA with all the 5 brain regions as dependent variables, alcohol as repeated measure and self-esteem as covariate. Or you can run a repeated-measures ANCOVA for each brain region as dependent variable and the other variables as said. MANCOVA is better though.

 

[kristall]

Hello! Thank you for your response.

To answer you:

Yes, I took self-esteems scores only once, which was a day prior to the scanning. Yes, each participant is scanned twice (once with placebo and once with alcohol).

I will try to clarify this study better:

The subjects came in a day before the testing and filled in a bunch of questionnaires, one being a self-esteem scale. I am testing the subject's "typical" level of self-esteem. (Not the type of self-esteem with the influence of alcohol or a placebo. This study assumes that under the influence of alcohol/placebo, a different type of self-esteem is present and I don't care about that.)

I want to treat "typical" self-esteem as a variable that could explain a part of the variation for neuronal activation in 5 different brain regions (both the placebo and the alcohol trials). I also want to compare the activation between placebo and alcohol trials.

Is anything else unclear?
Do you still suggest the MANCOVA?