I'm starting to get a bit overwhelmed with my project.
My mooting concerns design and sample size.
I have access to between 15-30 people with a clinical disorder and about 100 controls for my PhD.
I want to manipulate two factors, one with 2 levels, the other with 3 levels, on all participants - 2x2x3 factorial design.
I will be recording changes, i.e. the dependent variable, using a list of multi-choice questions.
As it stands, I have planned to do individual sessions with every participant on all six conditions. Presuming I am using the people with clinical disorders only (the clinical sample), they would each act as their own control - so using G-Power analysis of ANOVA: Repeated Measures, between factors at an effect size of 0.35, I would need a sample of 20 only?
To make it easier to understand: I want to test people's ability to remember stories of different lengths (factor 1 - 3 levels), while wearing pink clothing versus blue clothing.
My supervisor is unsure of getting approval from the service to do 20 individual sessions, and has suggested doing it in groups instead, I suppose 4 groups of 5 people, with each group undergoing all conditions in that session. I don't know how I could calculate this using G-Power? I suppose in essence nothing has changed in terms of statistical design, more so in the interpretation?
Ideally I'd like to compare people's ability to remember the stories on both factors, when in a group and when not in a group (adding an additional factor - 2x2x2x3 I think).
Any thoughts really welcome?! Is this too simple for a PhD project? Should I be trying to make more of the normal sample?