0
$\begingroup$

I just ran a hierarchical multiple regression with 6 control variables, 3 main effects and 2 interaction effects. To determine the critical value for the Mahalanobis distance (for identifying outliers) I need the to know the number of predictor variables to determine the critical value. This is probably a pretty straightforward question! How many degrees of freedom are there in my case?

Here the guide to the Mahalanobis distance: https://en.wikiversity.org/wiki/Mahalanobis%27_distance

Thank you so much for your help!

$\endgroup$
  • $\begingroup$ Can you edit your question to clarify why you are computing Mahalanobis distance here? $\endgroup$ – mdewey Jan 22 '18 at 16:52
  • $\begingroup$ Sure I just did! $\endgroup$ – TimH Jan 22 '18 at 16:54
  • $\begingroup$ I am not sure whether the control variables and the interaction variables also contribute to k. $\endgroup$ – TimH Jan 22 '18 at 16:57
  • $\begingroup$ The distance is just referred to the data matrix so your linear regression is irrelevant. If you want to identify outlying points affecting your regression try looking on this site for Cook's distance. $\endgroup$ – mdewey Jan 22 '18 at 17:13
  • $\begingroup$ Alright thanks! What would be the amount of predictor variables though? 11? Does one count the interaction variable and the control variables as well? $\endgroup$ – TimH Jan 22 '18 at 17:18

Your Answer

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

Browse other questions tagged or ask your own question.