Hi,
I'm trying to figure out whether the use of the Tobit Regression model is appropriate even when there is an ordinal outcome as dependent variable (instead of continuous one)
If not, do you know which other censored model would it be more suitable?
Thanks for any help!
Denise
No its not, Tobit assumes the dependent variable is continuous but censored, Tobit models a latent continuous variable where the censoring is usually due to measurement or observation limits. Ordinal is ordered, but where the distance between the levels are not necessarily equal!
Just use an ordinal regression model instead (logistic regression, probit regression or censored ordinal depending on the nature of the DVs)
Many thanks to both of you!
I'll have a look at this paper as well. Thanks!
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators