Hi all,
I'm suffering a lit bit with some data I have to analyse (my first time...) and I have many doubts.

We manipulated 3 IVs that were shown randomly many times in combination for each subj (within subject design):
- social status: high, low
- emotion: 3 conditions
- congruency: 2 condtions


subjects performed a task,so we also have errors and correct responses ^^. In total 2x3x2x2x2 would be 24 conditions.

We have some measures from 2 questionnaires related to the social hierarchy perception.


I already run an ANOVA, but now I want to show that the IV social status really had an influence on my DV. This is important because this manipulation was subtle (unconscious) and I want to demonstrate that there is an influence of the social hierarchy perception questionnaire somehow.

I think the most suitable analysis for this would be a moderation analysis, but I am using SPSS and I think there is only a macro, MEMORE, available for this type of analyses, and it doesn't allow you to include so many variables.

So I'm thinking about alternative analyses:

1) I was thinking about a multiple regression with the manipulated Social status, and the questionnaires as IV... How do I include that social status manipulation on my data for this? I coded it in long format so I have a variable that is social status and it is 1s and 0s, and the DVs for each condition... It seems so odd to include as a predictor the IVs as 1 and 0 and to expect that it considers that the DV are ordered according to the conditions... Is this approach completely wrong?
How can I include in a regression analysis the conditions manipulated?


2) If I can include the conditions manipulated somehow as predictors... Does it matter if they are within subject?


3) then, should I also enter the interaction between these categorical vv and the questionnaires? How would I do that?

3) Can I ignore that in the DV I have correct and incorrect responses and that some of the stimuli were congruent and other incongruent...?

4) If I wanted to ignore the different variables we manipulate and just do a regression with the questionnaires, would it be valid that we have all those different types of DV? How would it the best approach in this case?

Sorry for my ignorance, the really really long text and if the questions are stupid.

can't find the answer anywhere and your help will be very appreciated (Ill love u )