All,
Interaction contrast construction has been thoroughly discussed on
this forum - but some advice seems conflicting and I would like to
be explicitly clear. The two things Iwould like to be clear
on are the inclusion of the variables individually and the
construction of the ContXCont variables.
As an example, I am wanting to look at the effects of a categorical (binary) variables A, and two continuous variables B and C on FC in my experiment. I also have age and sex as variables. I have constructed the additional variables (AxB and BxC). Example values below on how that is constructed.
- Age = [ 18 20 22 21 23 19]
- Sex = [ 0 1 0 1 1 0 ]
- A = [1 1 1 0 0 0]
- B = [0.5 0.8 0.5 0.2 0.2 0.4]
- C = [5 3 2 9 1 2 ]
- AxB = [0.5 0.8 0.5 0 0 0 ]
- BxC = [2.5 2.4 1 1.8 0.2 0.8]
For any single variable A or B: I would select All, Age, Sex, and the variable, then use the contrast [0 0 0 1]
For a contrast for A and B: I would select All, Age, Sex, A , B, and AxB, then use the contrast [0 0 0 0 0 1] - keeping the indiviudal variables in the model but adjusted out.
This would tell me if there is a significant interaction effect
of A and B on FC, while adjusting for age and sex,
correct?
Thank you for any clarification and/or confirmation.
Threaded View
Title | Author | Date |
---|---|---|
jtanne98 | Feb 10, 2025 | |
Alfonso Nieto-Castanon | Feb 11, 2025 | |
jtanne98 | Feb 11, 2025 | |
Alfonso Nieto-Castanon | Feb 12, 2025 | |