1
$\begingroup$

I am estimating a model for an assignment, and found out that there is a 0.9 correlation between two of the independent variables. So If I am not wrong I should omit one variable and redo the regression. However, it is asked to me, to estimate the log-log model and then the correlation between variables drop from 0.9 to 0.5. so in the new model, should I still omit one of the variables? Also, which should be the cut-off in order to decide where I should omit a variable? Is 0.6 correlation high enough to give problems to the model?

Improve this question
$\endgroup$

0

Reset to default

Your Answer

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