I have a time course experiment with 3 samples per patient (paired design). I have 10 patients, and have 3 samples from each individual at time point 1,2 and 3. The 30 samples are indexed with #patient (1-10) and #time point (1-3). The experimental design calls for a repeated measures ANOVA. When I analyze with RM, I have to choose the index columns for 'group' and 'treatment'. I tried to make sense of the software help, but I think it's rather sloppy on this point (maybe check that yourself, the columns in your example are not easy to identify). Is it correct to give 'Patient' as the index column for 'group' and 'time point' as the index column for 'treatment'?
There is also no further analysis possible like the post-hoc tests for 1-way ANOVA. What would be the correct analysis to check which group is different to which when low p-values are found in the repeated measures ANOVA analysis? Should I do paired t-tests, would it be good to convert the data to logarithmic scale before analysis to ensure normal distribution? Is it a better idea to do a non-parametric Wilcoxon Test?
Thanks in advance