V [🇷 for BE/BA]

posted by ElMaestro  – Denmark, 2011-01-28 12:14 (4220 d 21:43 ago) – Posting: # 6507
Views: 63,499

Dear d_labes,

» where are the intra-individual (error or residual) variances or sd for Test and Reference?
» You have only one residual.
» The trt1 and trt2 variances (and corr) in your output are the between subject covariance parameter.

Interesting. Let us discuss for a moment the covariance matrix V (not Z and not G, and at this point also not discussing which syntax SAS applies for the given problem (leaves disussions about CSH, FA(0) out here)).
V holds the key.
Further, let us assume for simplicity we only look at the first subject which is the first 4x4 of V where the data are ordered as TRTR (i.e. per1-per2-per3-per4 for the first subject; in the original dataset above they were ordered differently but the order in this context does not change the point):

Data in per1 and per3 correlate within the subject as (s)he gets T (within variance VT).
Data in per2 and per4 correlate within the subject as (s)he gets R (within variance VR).
To this we add a common error variance on the diagonal (V0).

Having no particular mathematical understanding, the first 4x4 of V I think will be:
V0+VT     0         VT       0
0         V0+VR     0        VR
VT        0         V0+VT    0
0         VR        0        V0+VR


Both of the models discussed above correspond to a fit with this V, as far as I can see.

Pass or fail!
ElMaestro

Complete thread:

UA Flag
Activity
 Admin contact
22,305 posts in 4,668 threads, 1,587 registered users;
online 6 (0 registered, 6 guests [including 4 identified bots]).
Forum time: Friday 10:57 CEST (Europe/Vienna)

Doctors are men who prescribe medicines of which they know little,
to cure diseases of which they know less,
in human beings of whom they know nothing.    Voltaire

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5