Hi,
I'm trying to design a reward processing study where subjects can win small, medium, and large rewards. I plan to compare two groups of subjects on this task and I want to be able to test the hypothesis that Group A will show a linear increase in activity to these rewards whereas Group B will not. In particular, I want to be sure I'm looking at linear, not quadratic effects. Is there a way I can setup the glt such that the beta's I take into the group analyses specifically reflect the linear effect alone? I thought about simultaneously running a quadratic effect glt and then partialling out the quadratic effect but that would eat up much of the linear effect variance.
Thanks for your help.
Jatin