Thanks very much.
I tried taking the 2x2 example from the manual (for me, 1 factor is diagnosis and the second factor drug treatment ā everybody gets drug once and placebo once) and replacing the group mean with a mean for every subject ā I still got the rank deficiency error until I removed the EV for diagnosis as I saw in a previous post that I can create a contrast to establish diagnosis: s1+s2+s3-s4-s5-s6 as the first three subjects are patients and the next three controls). Thus, this worked, where EV1 is drug treatment and EV2 is diagnosis by drug interaction:
Gp EV1 EV2 S1 S2 S3 S4 S5 S6
1 1 1 1 0 0 0 0 0
1 1 1 0 1 0 0 0 0
1 1 1 0 0 1 0 0 0
1 -1 -1 1 0 0 0 0 0
1 -1 -1 0 1 0 0 0 0
1 -1 -1 0 0 1 0 0 0
1 1 -1 0 0 0 1 0 0
1 1 -1 0 0 0 0 1 0
1 1 -1 0 0 0 0 0 1
1 -1 1 0 0 0 1 0 0
1 -1 1 0 0 0 0 1 0
1 -1 1 0 0 0 0 0 1
But in fact in my experiment I have 3 diagnostic groups (I thought I need to understand the 2 group situation before going to 3 groups), so now I need an ANOVA with 1 factor at 3 levels (diagnosis), and one repeated measure factor at 2 levels (drug or placebo). I need to modify the above DM by changing the interaction EV, but Iām stuck at this point. I suspect I need more EVS but if I add EVs to indicate groups then I get rank deficiency errors again. Any thoughts on ANOVA with one factor at 3 levels (between subjects) and one factor repeated measures (2 levels) would be very much appreciated.
Thanks
Graham
|