It might sound like an easy question. What makes me confused is that the independent variable takes any values between 0 and 5 including the numbers such as 1.987. So, I am not sure if I should consider it as an ordinal variable.

  • $\begingroup$ So the outcome is on an interval? Like 0 to 10 and can be any number in that interval? $\endgroup$ Sep 9 at 22:57
  • $\begingroup$ It takes any values between 0 to 5 such as 0, 3.562, 1.97 etc. @DemetriPananos $\endgroup$
    – Reyhan
    Sep 9 at 23:20
  • 2
    $\begingroup$ Rescale the outcome to be between 0 and 1 and you can use beta regression. $\endgroup$ Sep 9 at 23:23
  • $\begingroup$ Thank you very much. $\endgroup$
    – Reyhan
    Sep 9 at 23:37
  • $\begingroup$ What is your dependent variable? $\endgroup$ Sep 10 at 8:51

follow the lead of Demetri and rescale the variable to be between 0 and 1. Then you can use beta regression but you also have the fractional logit model in your arsenal, see here.

I hope this helps.

  • $\begingroup$ See also from 22 years earlier R. W. M. Wedderburn. 1974. Quasi-Likelihood Functions, Generalized Linear Models, and the Gauss-Newton Method. Biometrika 61: 439-447 jstor.org/stable/2334725 $\endgroup$
    – Nick Cox
    Sep 10 at 7:50
  • $\begingroup$ Thank you very much to you both. $\endgroup$
    – Reyhan
    Sep 10 at 16:03

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.