again, more precise: Fisher s exact test for bigger tables than 2x2

[Max Jasper]

Hi all,

I run a multi GLM on several drug doses as covar, and some other factors, and use several blood test results as multiple dependant variables.

QUESTION:
Is there a way to tell GLM that depVar3 is NOT dependant on indepVar5? If this can be done then effects of some drugs on blood test results might be more realistic.

Thanks for all comments,
Max.

[Max Jasper]

Hi Gene,

Thanks for your response.

Yes, as you said it. In your model, blood1,2,3 are blood test results like
GLUF, K, Na, LDL, HDL, TRIG, CHOL, and dose1,2,3 can be doses of several
drugs like amlodipine, ramipril, spironolactone, etc.

I know for example that amlodipine does not affect blood test values of LDL,
HDL, TRIG. I want to specify in my GLM model that GLM considers zero effect
fom amlodipine on LDL, HDL, TRIG while GLM would consider amlodipine's
effect on other dependant variables like K, Na, GLUF.

I don't know if this can be done in SPSS via LMATRIX, KMATRIX, MMATRIX,
SPECIAL matrix, or a ZERO STRUCTURE variable.

Appreciate any clue on this matter.

Max.

|
|Max,
|
|Are you analyzing this
|
|Glm blood1 blood2 blood3 by factor1 factor2 with dose1 dose2.
|
|So that dose1 and dose2 are covaried out of blood1, blood2,
|and blood3.
|
|If this is not correct, what is correct? If this is correct,
|what do you want? I'm sorry but your description just was not
|clear. Please reply back to the list so all can see.
|
|Thanks, Gene Maguin
|
|
|