My professor said that control variables are generally thought to be exogenous in an OLS context, i.e. uncorrelated with the main independent variable in my regression. However, I think that they are endogenous and not only determine the dependent variable but also correlate with the independent variable of my interest. So, my question: Should we think of control variables as being exogenous to the independent variable of our interest in our OLS regression? If so, how do we call an endogenous "control variable"? Maybe it's just terminology that's confusing me.

  • 1
    $\begingroup$ "Exogenous" does not mean "uncorrelated with the main independent variable", it means, roughly, "independent of the error term." This may help clear up your confusion. $\endgroup$
    – jbowman
    Jun 11, 2022 at 20:12


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.