I have recorded data from N subjects (for the moment 9, but it will increase). After each experiment I use different analysis and the success rate. The analysis is divided into three parts which can be perform in several ways:

Start : S1, S2, S3, S4

Mid: M1, M2, M3, M4, M5

End: E1, E2, E3, E4

S4 can be combined only with M4 and M5, while S1..S3 can be combined with M1...M3. E1...E4 can be combined with all the previous combinations. In total there are 44 combinations, for example S1M1E1 or S3M2E4.

The first analysis that I performed was a one-way ANOVA  to find which was the "best" combination. The next that I want to test is which is the best Start, Mid and End.

I think that neither n-way ANOVA or one-way MANOVA is the test that I should run, should I group by for example "Start", take the mean value for each subject and perform a one-way ANOVA?  Should I count each of the different results as "Repeated measure"? In both cases I am losing the information about the interaction between "Start", "Mid" and "End". Which is the appropriate statistical test?

More Jacobo Fernandez-Vargas's questions See All
Similar questions and discussions