Hi Giovanni
> I've an event-related design with 3 kinds of events (A1, A2, B), each occurring the same number of times.
>
> Given that the contrast vectors [1 0 -1] and [0 1 -1] give A1>B and A2>B, I wonder if the correct vector for testing the hypothesis (A1+A2)>B is [1 1 -1] or [1 1 -2] (which sums to zero).
If you want to know whether the average of A1 and A2 is greater than
B, which is what I'm guessing, then you want [1 1 -2] (or [.5 .5 -1]).
> I also wonder if I should use the t contrast [1 1 0] or the F contrast [1 0 0;0 1 0;0 0 0] to show the (positive) combined contribution of the two regressors A1 and A2.
A [1 1 0] (or [.5 .5 0]) t contrast will show you the averaged
positive effect of A1 and A2. The F contrast you list would show you
regions that show a (positive or negative) effect of A1 *or* A2.
Hope this helps,
Jonathan
|