Page 1 of 1

Problem understanding Tobit

Posted: Mon Jun 25, 2012 12:39 pm
by starik123
Hi all,

I have some problems understanding, when Tobit is appropriate. So far I know, that

- when the dependent variable is censored, i.e. the probability distribution is truncated, I have to use Tobit. An example would the capacity of a football stadium with 20,000 seatings. If there is demand for 50,000 tickets, the variable will be 20,000.

But here I am reading a paper, studying international investment. The dependent is a ratio between 0 and 1 (1-foreign market capitalization/world market capitalization, from the perspective of one specific country). The authors claim, that because it is bounded it has to be estimated via Tobit.
As far as I understand, just the fact of boundedness does not require Tobit, only if the variable is censored???

Thanks for your help!

I am really confused right now.