I have a question that I'm hoping somebody here can help me with. I am a student working on my psychology thesis and am having some difficulty analyzing my data.
I gave 300+ participants a questionnaire. They all answered a personality measure (continuous scale) and all gave their gender. I also gave them another set of questions that involved a categorical (yes/no) response to hypothetical situations. The hypothetical situations varied so that they had a 2x2x2 within-subjects design. The conditions involved hypothetical favors for a close friend or stranger, male or female, and emotional or nonemotional problem. Each participant either got "easier" versions of the situations or "harder" ones, so they each got 8 situations and difficulty was a between-subjects factor. Participants responded to a Yes/No question about the situations (whether they would be willing to help with the hypothetical favor or not).
To recap in list form:
Between Subjects Factors:
-Hypothetical Situation Difficulty Group (easy v. difficult; categorical)
-Participant gender (categorical)
-Personality questionnaire (continuous)
Conditions with yes or no answers (hypothetical situations each with 3 manipulations with 2 levels each) coded as:
Hypothetical Situation 1: Closefriend, non-emotional problem, male (1,1,1)
Hypothetical Situation 2: Close friend, non-emotional problem, female (1,1,2)
Sit. 3: Close friend, emotional problem, male (1,2,1)
Sit. 4: Close friend, emotional problem, female (1,2,2)
Sit 5: Stranger, non-emotional problem, male (2,1,1)
Sit 6: Stranger, non-emotional problem, female (2,1,2)
Sit 7: Stranger, emotional problem, male (2,2,1)
Sit 8: Stranger, emotional problem, female (2,2,2)
I am interested in looking at how the personality measure and participant gender were related to the responses to the situations and also at how the different conditions within the situations relate to the responses. My primary interests include the interaction between participant gender and protagonist gender as well as the interaction between emotionality of the favor, time and gender. Ideally I would do this all in one analysis so that I can look at the interactions as easily as possible.
Initially, I analyzed this in SPSS using a repeated measures GLM. However, I have had a couple people suggest that I should do it in a logistic regression instead. One said that the GLM violates an assumption.
So, my questions are: Does the repeated measures GLM actually violate any assumptions? I see other researchers using repeated measures tests for this kind of design. If there isn't a problem with doing this as a repeated measures GLM, I would rather do it that way since I know how.
Any help/input would be greatly appreciated!