I am completely stumped on which tests to use for a study that I am working on. Basically, I have two groups, 1 & 2. The two groups are of different size and the samples are small (9 & 12). Participants in both groups have performed the same three tests, A, B & C in the one session. The tests are similar formats but differ in difficulty level. There are two things that I want to find out:

1) When 1 & 2 are pooled into one group (1+2) are the mean scores for A, B and C significicanlty different from each other. I believe that a one way anova with post hoc bonferoni tests is the best option here. Howwever, but I am not really testing three independent groups... just three sets of scores form the same group!

2. Are the mean scores for A, B & C significanlty different from each other within each group (1 & 2) and between the two test groups? Remember the groups are uneven in size. I am stumped on which non-parametric test to use here. Maybe a Friedman two way ANOVA by ranks?

Apologies if my suggestions above seem a bit misguided. Stats are not my strong suit.

Any suggestions welcome, as are any good links to how the data should be set up in SPSS.

Thanks, BOAZ