Hello experts,
I have a task with a 2 x 2 x 4 design that I am attempting to examine at second-level using a full factorial, along with some custom contrasts. As the third factor relates to time, and I want to examine if activity is decreasing over time linearly, I've gone about setting up a custom (simplified F) contrast based on multiplying the following contrasts: factor 1 is coded [0.5, -0.5] X factor 2 coded as [0.5, -0.5] x and linearly decreasing factor 3 is coded [3 1 -1 -3]. This yields the following contrast: [0.75, 0.25, -0.25, -0.75, -0.75, -0.25, 0.25, 0.75, -0.75, -0.25, 0.25, 0.75, 0.75, 0.25, -0.25, -0.75].
What I wanted to know is if this is a reasonable approach and if I've gone about setting up my initial contrasts correctly. My third level contrast should, ideally, represent a relatively linear decrease, though I'm most interested in what happens in the last session.
Best,
Akos Szekely
|