Hi, thanks for your responses.
The background for this is that I'm picking up an SPM analysis from some people who has left the university. I have some further questions that arise from your responses.
1) I have the two t-tests that Ella mentioned as working for the interactions. I assume these can be considered interactions because they are manipulations of the interaction formulas (A1B1 - A1B2) - (A2B1 - A2B2) [Effect of B at level A1 - Effect of B at level A2) or its opposite. I also assume that I run these interactions separately.
I get results for the 1, -1, -1, 1, 0 contrast, which is the one that springs from the manipulation of the above interaction formula. Am I correct in interpreting this as showing areas where the effect of factor B is larger at Level A1 than at Level A2? Sorry I didn't fully understand Ella's description of "areas with decreased activation from A1B1 to A1B2 and increased activation from A2B1 to A2B2"
2) Flexible factorial design. Is this where subject is modelled as a separate factor? The method notes I have say that the group level analysis involved modellling participant and condition as predictors using flexible factorial design, so I assume I am OK with this. However 'subject' doesn't appear in the design matrix in the constrast manager - is that correct?
Thanks
Rob
|