excellent reply thankyou yanqioung. That is exactly what I want. After posting the message I realized I consulted with a colleague and confirmed that the main effect of factor B matrix I had constructed would not work.
Here is what we came up with instead. I believe it is equivelent to what you provided but please let me know If I am incorrect.
For the main effect of B I used the following set of contrasts:
1 -1 0 0 0 1 -1 0 0 0
0 1 -1 0 0 0 1 -1 0 0
0 0 1 -1 0 0 0 1 -1 0
0 0 0 1 -1 0 0 0 1 -1
and for the interaction I performed the same transformation on the matrix as you have done:
1 -1 0 0 0 -1 1 0 0 0
0 1 -1 0 0 0 -1 1 0 0
0 0 1 -1 0 0 0 -1 1 0
0 0 0 1 -1 0 0 0 -1 1
This now gives me 4 beta weights for the main effect term and 4 beta weights for the interaction term. In order to perform the group level statistics on this data I plan to take the beta weights from each subject (12) and submit them to a one way anova with 4 levels, each level a beta weight. I am assuming that if the differences revealed by any one of the contrasts is large enough across a sificient number of subjects a main effect should be seen. The resulting statistic would be the final map for that comparison (for both the main effect contrasts and the interaction contrasts). Does this seem reasonable?
KJ
