3
$\begingroup$

I used ridge regression in order to dealing with multicollinearity but there is something that i do not understand. I used Stata command:

- ridgereg y x1 x2 x3 x4 x5 x6 x7 x8 x9, model(orr|grr1|grr2|grr3) diag lmcol-

in addition in orr model i used many different values for $\lambda$. The problem is that: when I run the above model and checked VIF it had the same value for all variables for all different models and for all different values of $\lambda$. Is it possible? Does it make any sense? Does it mean that my model suffers very seriously of multicollinearity?

$\endgroup$
  • $\begingroup$ @Jeremy Just noticed your edit. Why did you format the command as a bullet list entry instead of the usual code sample block? Also, the title of this Q is horribly misleading, why not writing an informative title if you decide to edit this old question anyway. $\endgroup$ – amoeba Jan 4 '17 at 9:20
  • $\begingroup$ Oops. Bullet was an accident. I will change title (I should have, but capitalization of Stata annoyed me so I jumped in and changed that). $\endgroup$ – Jeremy Miles Jan 4 '17 at 16:09
2
$\begingroup$

Yes, I guess so. If you look at the procedure in Wikipedia of VIF and think about its implementation, it probabely only takes into considerations specified regressors independent of the model you choose.

https://en.wikipedia.org/wiki/Variance_inflation_factor#Calculation_and_Analysis

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.