I want to use VIF to check the multicollinearity between some ordinal variables and continuous variables. When I put one variable as dependent and the other as independent, the regression gives one VIF value, and when I exchange these two, then the VIF is different. And once the VIF value is higher than 3, and the other time it is lesser than 3.
Then, how I do make a decision to keep the variable or not, and which one should I keep? Ultimately, I am going to use these variables in a logistic regression. How important it is to see multicollinearity in logistic regression?