2

I am trying to test whether a model I am using has multicollinearity. The only two methods I've learned are condition numbers and variance inflation factor (VIF) to determine whether multicollinearity is present. However, all the examples I have used involved numerical variables.

Can VIF and condition numbers be used to evaluate a model with both numerical and categorical variables? If not, what is the appropriate way to assess multicollinearity in such circumstances?

Based on this question,

Collinearity between categorical variables

I would suspect the answer is no these two tools are not suitable, but I am uncertain since the answerer focuses on collinearity not multicollinearity.

kjetil b halvorsen
  • 77,844
Stan Shunpike
  • 4,333
  • Asked before ? https://stats.stackexchange.com/questions/274445/does-the-vif-make-sense-for-a-model-with-categorical-variables, https://stats.stackexchange.com/questions/430412/vif-for-categorical-variable-with-more-than-2-categories, https://stats.stackexchange.com/questions/285722/vif-doesnt-show-up-values-for-categorical-variables, https://stats.stackexchange.com/questions/583434/vif-but-with-a-categorical-response-variable, https://stats.stackexchange.com/questions/457962/how-is-vif-calculated-for-dummy-variables, – kjetil b halvorsen Dec 09 '23 at 01:30
  • ... https://stats.stackexchange.com/questions/409529/how-is-gvif-calculated-for-categorical-variablesalso-is-there-any-other-way-to – kjetil b halvorsen Dec 09 '23 at 01:30

0 Answers0