I'm a beginner and have this question: In a negative binomial regression analysis, is it possible to check if there is multicollinearity? I'm trying to introduce some moderating effects (e.g. interaction terms) in my model with count data as a dependent variable, and I'd like to make sure there's no multicollinearity (I've standardized the variables of interest previously).

Any help will be greatly appreciated!

  • $\begingroup$ Ways to check for multicollinearity between your predictors don't change depending on how you model the response. $\endgroup$ – Scortchi - Reinstate Monica Jul 21 '13 at 0:31
  • $\begingroup$ Look here for example. $\endgroup$ – COOLSerdash Jul 21 '13 at 8:25

As @Scortchi says in the comments, you can use all the usual techniques for examining multicollinearity. Multicollinearity is a quality of the linear predictor - a linear combination of the explanatory variables. The fact that, in your Generalized Linear Model process you are subsequently going to transform that linear predictor and that the result of that is modelled as having a particular distribution, doesn't change the basic conceptual issues around the linear predictor.

| cite | improve this answer | |

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.