Hello,
I am solving a practical problem I believe it has been solved already many times. I have two groups of parients. One has the disease the other one does not. Patients are described by 4 parameters (height, age,...) 5th parameter comes from the blood analysis. The question we are asking is: Is the value of the 5th parameter signifficantly increased in patients with the disease?
When analyzing the whole data sets the answer was yes. But at the same time we have noticed both groups vary significantly in also in all other parameters. thus we cannot say that this increase in the 5th parameter was caused by the disease only.
So we want to cut the control group so it would have the same values of 4 parameters as the diseased group and do this test again. Therefore my question is how to cut this group to avoid criticism we did that on purposse to prove what we want to prove.
Thanks for any suggestions.
What is to stop you simply running mutivariable logistic regression? You would be looking for differences in parameter 5 when you adjust for the effects of the first four.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size