0
$\begingroup$

I've looked into other questions that dealt with what regression models to use when the dependent variable is continuous between 0 and 1 (some people have mentioned fractional logit, GLM, and even OLS), but not something that is the same question as I have.

I have a dependent variable that can only take four possible values between 0 and 1, namely 0, 1/3, 2/3, and 1. Which regression models would work well with this kind of dependent variable?

Thank you in advance.

$\endgroup$

1 Answer 1

3
$\begingroup$

That looks like a discrete, ordered variable, which makes ordinal regression (e.g., ordinal logistic regression) appropriate for it. If you don't want to make the proportional odds assumption, you can also treat the variable as a nominal variable and run multinomial regression (e.g., multinomial logistic regression) on it. Ordinal regression would be the most interpretable model, though.

$\endgroup$

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.