Thank you for reading this and for helping.
I have a run a semantic priming experiment. (A brief overview, if you are unfamiliar: Show a prime word before a target word/phrase. If the prime and target are part of the same semantic category, e.g., furniture, animals, football positions, etc., a subject's "reaction time" to identifying the target will be shorter than if the prime was something unrelated.) I am not sure how to get SPSS to take into account the levels of data.
* Layer 1: Subjects either have a Condition X, or they are controls.
* Layer 2: Among the former, there are two variants of Condition X. Let's call them Type Alpha and Type Beta.
* Layer 3: Among subjects with Condition X, they were tested in two states -- State 1 and State 2 -- counterbalanced randomly so that some were tested in State 1 first and the rest were tested in State 2.
(I don't know if it is feasible to clearest to run the analysis with all 5 groups together in the same analysis, or to run it with each group -- Controls, Alpha1, Alpha2, Beta1, and Beta2 -- separately. It's the larger pattern of "did the priming work in group G or not" that matters to me.)
Dependent Variables: Reaction Times to priming experiment
* 8 independently shown congruent prime/phrase pairs -- Call these trials AA.
* 8 independently shown incongruent prime/phrase pairs -- Call these trials BA (different primes, same target category).
Covariates: There are some related to typical demographic measures and ones measuring severity of Condition X.
What I know: I know how to do a repeated measures ANOVA if I only had a single number for AA and BA. I could easily find the means for each subject, but using just the mean seems statistically dishonest as I would lose the variance within AA and BA for each subject.
What I know that I don't know: I know how to do a repeated measures ANOVA that takes into account the variance of AA and BA, too.
Whatever assistance you can provide, including just the best search terms to find the right help in previous threads here or via a search engine, would be greatly appreciated.