Dear Donald,
Thank you very much once again for your helpful explanations about the orthogonality matter.
Now I fully understand that there is any problem with the non-orthogonality in the way that FSL runs an ANOVA (1 -1 0; 0 1 -1).
Taken into account that we are asking for an ANOVA from a non-orthogonal pair of contrasts (1 -1 0: 0 1 -1) could you please detail briefly how FSL does in order to achieve and/or correct the orthogonal principle?
thank you again in advance
Daniel
|