I have a dataset where I want to find the relationship between a factor(10 levels) and several independent variables(all numeric). I don't think it will be a good idea to convert the dependent factor to a dummy variable since I have ten levels. Anova seems like the best way to do this but my dependent variable is not numeric. Any advice on how to approach this?

  • $\begingroup$ How would you do this problem if the dependent variable were numeric? Let’s start there. $\endgroup$
    – Dave
    Mar 11, 2021 at 0:49

1 Answer 1


If your dependent variable is ordered (Bad / Okay / Good), you could use a proportional odds logistic regression. If there is no ordering (Apple / Banana / Orange), you can use a multinomial logistic regression.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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