I am really puzzled as to how to set up a repeated measures multiple factor
analysis of covariates and could use some feedback. I have 6 subjects, each
with functional data across 3 sessions with 2 body testing sites at each
session: session 1 is baseline, session 2 is 24 hr post-drug, and session 3 is 2
wks post-drug. My covariates include age and 2 systemic measures of drug
(the drug and its metabolite), in both the csf and in the blood. These last
two measures are basically redundant and I would expect them to be collinear.
So my design so far is:
EV1 EV2 EV3 EV4 EV5 EV6 EV7 EV8 EV9 EV10 EV11 EV12 …
S1_sn1_a 1 0 0 0 0 0 1 0 -1 -1 0 24
S1_sn1_b 1 0 0 0 0 0 0 1 -1 -1 0 24
S1_sn2_a 1 0 0 0 0 0 1 0 1 0 -1 24
S1_sn2_b 1 0 0 0 0 0 0 1 1 0 -1 24
S1_sn3_a 1 0 0 0 0 0 1 0 0 1 1 24
S1_sn3_b 1 0 0 0 0 0 0 1 0 1 1 24
S2_sn1_a 0 1 0 0 0 0 1 0 -1 -1 0 29
S2_sn1_b 0 1 0 0 0 0 0 1 -1 -1 0 29
S2_sn2_a 0 1 0 0 0 0 1 0 1 0 -1 29
S2_sn2_b 0 1 0 0 0 0 0 1 1 0 -1 29
S2_sn3_a 0 1 0 0 0 0 1 0 0 1 1 29
S2_sn3_b 0 1 0 0 0 0 0 1 0 1 1 29
....
S6_sn3_b 0 0 0 0 0 0 1 0 0 1 1 43
Where Evs 1-6 account for repeated measures within each of 6 subjects, Ev 7
= test site a, EV 8 = test site b, EV9 = Scan 2 > 1, EV10 = Scan 3 > 1, EV11
= Scan 3 > 2, and EV12 = age, for example. I really have no idea how to
model the multiple levels of drug and in both the blood and csf—other than to
treat them each like 2 levels of a single factor (factor 1: drug and drug
metabolite, and factor 2: blood and csf).
Can anyone offer some feedback on what I am doing right/wrong? Thanks so
much…
Melissa
|