08-21-2015 07:49 AM
I was wondering if it was possible to add confounding variables to my discriminant analysis in SAS enterprise guide.
Let me first describe the situation.
I have one single dependent variable, which is At Risk 1 or not at risk 0.
Then there are 3 continuos independent variables: Physical activity, sedentary and physical fitness.
The question I have to solve is: In which way can Physical activity, sedentary and physical fitness distinguish the two groups (At risk/not at risk)? Are the people with low physical activity, high sedentary and a low physical fitness more at risk? Age, Sex (0/1) and smoking (0/1) are to be used as confounding variables in this analysis.
So I was thinking about discriminant analysis to adress this question.
Only I couldn't find a way to add this confounding variables. Can anyone help me with this? Or is another test better for this question?