I am constructing a 2-sample t test on SPM8 comparing females (n = 19) to males (n = 18). In my analysis I would like to add birth control, follicular phase, and luteal phase as three covariates to the female group to reduce error. I have entered three separate covariates in the experimental design and have dummy coded for each.
I understand that if it were just a female vs. male t test, my contrasts would be "1 -1" in which the 1 represents the female group (the female group was entered first and the male group was the second) and the -1 is the male group. When I get to the contrast manager however, I am unable to enter my contrast coding, I get the message "no valid contrast defined."
Which numbers should I use as my contrasts? It has been suggested that I use "1 -1 1/3 1/3 1/3" my contrast for the Female>Male contrast and "-1 1 1/3 1/3 1/3" for my Male>Female contrast. SPM however rejects these numbers. Would anyone please be able to help me with which contrasts I should input for this analysis?