covariate for only clinical group

#1
HI
I have tried to find an answer on the web but was not able, and will appreciate your help!
I have a question regarding the use of a covariate in my analysis.
I have 2 groups- 1 of patients with clinical variables and a healthy control group. I would like to assess differences between the 2 groups in a continuous dependent variable but to control for 3 vars, using GLM in SPSS.
The problem is that 2 variables which I would like to control for are clinical so ONLY my patient group has these variables (and not the control group-sysmiss). The third variable I wish to control for is a variable that all subjects in both groups have a value.
My question: Can I enter a zero value for all my control subjects in the GLM analysis? I this statistically correct? or am I supposed to run a different analysis to examine differences between the 2 groups while controlling for 3 covariates?
Thank you in advance
 

rogojel

TS Contributor
#2
hi,
I think using the clinical variables would contradict the intent of "controlling" for a variable. The idea with that is to take into account differences between the two groups that are coming from differences in the controlled variable ans are not due to the factor you are checkung. E.g. you want to check whether the time spent playing basketball affects the performance - so your main variable is time. But you might want to control for height, because higher players might be better (and they may also spend more time playing). Now imagine you had the height data for successfull players only. There would be no way to dontrol fir height, right?

regards
 

hlsmith

Omega Contributor
#4
In addition, I am guessing the control group could have a value for the covariate, but they don't given your study design (e.g., BMI, blood pressure, white blood cell count). So their value may not really be "0".