Hi Jane, If the model is Y = intercept*b0 + X*b1 + Z*b2 + X.*Y*b3, then the significant interaction means that the relationship between Y and X varies according to Z, and likewise, the relationship between Y and Z varies according to X. From this test we can't say anything about the relationship between X and Y. All the best, Anderson On 8 May 2017 at 08:41, jane <[log in to unmask]> wrote: > Dear FSL users, > I'm new and I want to test an interaction between two variables on > Fractional anisotropy (FA). I saw that this model has been proposed > > EV1: intercept > EV2: first continuous variable (X) > EV3: second continuous variable (Z) > EV4: EV2*EV3 (interaction) > > > The contrasts are: > C1: [0 0 0 1] > C2: [0 0 0 -1] > > I have a doubt in the interpretation: if C1 is significant it means that > the relationship between X and Y varies according to Z, or that > relationship between Z and Y varies according to X. > This contrast gives me also the direction of the effect? I mean higher > values of both variables, higher FA? > > Thank you in advance. > > Best, > Jane >