I have a data set with a categorical dependent (14 categories - party choices), and three independent variables - age (continuous), education (8 categories) and gender (male or female).

Online I understood that I need to run a multiple regression analysis. However, advice online is conflicting, with some proposing the use of dummies while others do not.

Could you please help me with some guidance? The software I work in is RStudio.

Thanks in advance.

Regards, Vlad

  • $\begingroup$ Was my answer helpful? $\endgroup$ Dec 13 '17 at 23:22

I'm worried that this question is a duplicate, but I'll answer it anyway. What you are describing can be called "supervised classification". There are a wide range of tools that you can use to solve such a problem, and each tool has its own strengths and weaknesses. Here is a brief summary of your options, borrowed shamelessly

  • Multinomial logistic regression: the crux of what is discussed here
  • Multinomial probit regression: similar to multinomial logistic regression with independent normal error terms. This is a nice R package for that.
  • Multiple-group discriminant analysis: My personal favorite, but it does have some limitations (assuming multivariate normality of the errors for one). See this post on how to perform it in R.

The article lists a few other options for you to explore. Beyond their options, Random Forests via Classification Trees is also very powerful, but can be too "black box" for some people.


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.