I have two outcome variables: A) 5-point Likert scale B) the number of drinks a participant had (ranges from 0-5)

My independent variables are sex, age, income, location, and ethnicity. I'm planning on conducting two regression models - one for each outcome. For the first outcome, since a Likert scale is ordinal, I assumed I should do an ordinal logistic regression. For the second outcome, I'm unsure whether to use a Poisson regression since it is the # of drinks or whether I should consider this an ordinal variable as well and do an ordinal logistic regression.

Thanks for any help in advance!

  • $\begingroup$ There's not a perfect answer here; it comes down to which model serves your goal best. If you do choose to use a Poisson distribution, be sure to renormalize the likelihood based on the constraint that $x \leq 5$. $\endgroup$ – Arya McCarthy Apr 8 at 19:32

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.