Hi Andrew and NBS users',
I would like to confirm my design matrix with repeated mesures and adding symptom covariate.

Information about the design:
9 patients
Longitudinal study (time A and time B): repeated level.
2 clinical variables

My design matrix would look like: (to simplify, with 2 patients)
1 0 1 4.5 5
0 1 1 4.8 4
1 0 -1 2.1 7
0 1 -1 2.7 6

Exchange Blocks
1 2 1 2

The 1st two columns: within subject mean, the 3rd: between-group , the 4th: First symptom covariate and the 5th: second symptom covariate

Contrast: with F-Test
Gr1>Gr2 [0 0 1 0 0]

Is-it good?

Kind regards,
Frederic

Hi Fredric,

Your design looks fine. I suggest to start simple and omit the symptom covariates in the first instance. You can then add them in later and see what effect they have on the result.


You might also want to start with a t-test and use [ 0 0 1 ...] or [ 0 0 -1 ], assuming you have a hypothesis about the direction of the effect.


Andrew

Originally posted by Frederic Briend:

Thanks for your help Andrew,
Best,
Frederic