Hi Andrew,

thanks for the very quick reply. 

Yes, sorry distraction mistake. To be faster I've copied and pasted the interaction column from the placebo group to the drugs group without paying attention to the switch of sign in the previous columns.

Regarding to model the repeated measure,the matrix then would be (I'm modelling a situation with 4 subjects 2 for group (P = placebo; D = drugs), and 2 condition A and B)








as you suggested: first column would represent the conditions, second column represent interaction, and then 1 column for each subject. However I have some doubt:

1- regarding the column for the subjects, for the treament group I've used -1 is it correct? Otherwise I was wondering how can I differenciated between the 2 treatment group?

2- Considering using "t-test" what would be now the contrast?

 
Is not clear to me what would be the contrast to test if there's a main effect of the treatment.

3- regarding the first desing (2nd column treatment, 3rd condition, 4th interaction)







 
I suppose the results are going to be different compare with the other model. Which solution would you suggest?

I'm sorry for the naive questions but my experience with the GLM is zero, I've just started to approach it.

Thanks again for the attention.
Davide