I am doing a logistic regression analysis (my dependent variable is a dichotomous dummy variable). I did Factor Analysis on some of the independent variables and I got reduced dimensions representing that three variables fall into one dimension (as I expected). Since two of these three variables are dummies (hospitalization and physical limitation) and the another one is ordinal with 5 categories (perceived health status), how I can mix these three and form one variable. I also cross-tabbed these variable and the result was same.

So, my questions are:

1) How I can mix all these three variables and put them in my model?

2)If I put all of these three variables individually (without forming a unique variable) multicollinearity will be a problem? 

Similar questions and discussions