Hi,
I'm a bit confused by what you are saying here.
If you are looking at the corrected p-values outputs (corrp) then you should never get values larger than 1.0 or less than 0.0. Can you confirm what images you are looking at which contain values greater than 1?
Also, I'm not sure how you've set up the interaction. If you have continuous covariates (and I assume you do) then the interaction needs to be modelled with a separate EV, not just a contrast. From what you describe here it seems that you have a contrast that is just looking at the difference between correlations (or rather the strength of the linear relationship) of FA with A and B. This is not an interaction, if I've interpreted your email correctly.
Anyway, if you can give us some more information then we can try and help some more.
All the best,
Mark
> On 12 Oct 2017, at 11:47, Arman Bordbar <[log in to unmask]> wrote:
>
> Hi Dear all,
>
> I have run an correlational analysis between FA values and two covariates (creativity(A) and logical thinking(B)). In the tfce p-corrected results for individual variable I don't see any signicant cluster but when I examine the interaction between A and B (e.g. contrast x, 1 for A and -1 for B in Glm gui) I see many significant clusters and someones with values more than 1. Does my design make sense? Can I interpret it as "The FA values have positive correlation with A when B decrease"? What do these values more than 1 in the tfce image mean?
>
> Bests
>
> Arman
|