10 different events --> moderator --> 9 symptoms

#1
Hey folks!

I have a rather complicated analysis to compute, and am not 100% sure how best to do it.

Design & Variables
(1) Longitudinal study, large sample, 5 measurement points every three months. At first measurement point big test battery, and consecutive measurement points measurement of some important variables.
(2) Main variables:
* Current symptoms for one mental disorders (9 different items on the questionnaire, selfreport). I am not interested in the composite score, I am interested in symptom clusters.
* Life events experienced during the last 3 months. Multiple choice questionnaire, 10 different life events (e.g. "financial problems"), more than 1 answer possible.
* Genotype: three different categories: low, medium or high.

Core Questions
(1) Do different life events lead to different symptoms? (general question, one analysis, yes or no)
(2) Do different life events lead to different symptoms, controlling for important variables (e.g. gender)
(3) Which life events lead to which symptom clusters?
This seems hard, how can I do this? In the end I want to have "life event x1 leads to symptoms y1, y3, y7", "life event x2 leads to symptoms y1, y2, y4" etc.
(4) Is this moderated by genotype?

Help is appreciated for these four question.
Thank you guys ;)
Torvon
 
#3
For question (1), you most probably would do a MANOVA/omnibus-interaction-analysis. That would tell you: yes or no.

For the other three questions, I have no clue what to do.

Thanks
T-