Hi all. I have the following situation: a group of 10 subjects has a variable (called "ag") measured twice (variable "time" with levels 1 and 2) and no between-subject factor. I would like to use a mixed model instead of a paired t-test so that I can take into consideration another covariate (called "ultra") as well as the individual variability.
Is the following procedure correct?
class subject time;
model ag = time ultra;
or instead should I use subject(time) in the random statement? I am asking it because I get substantially different results if I use one procedure or the other.
Or am I even completely wrong since this procedure can only be applied in crossover situations?