That depends on quite a few things.
It depends on the assumptions you are willing to make on the distribution of
the error and reasons for such bounds.
There are few chapters on the User Guide that are helpful.
Otherwise you can start read up -tobit- -cnreg- -truncreg-
Consider taking the log of your variable so that at least you get rid of the
problem that it is always positive, you can then interpret the coefficients
as if the variable had not been tranformed by retransforming the dummies
back using exp()-1
Renzo
----------------------------------------------------------------------------
----
*From "Apostolos Ballas" <[email protected]>
To [email protected]
Subject st: Limited Dependent Variable
Date Sat, 21 Aug 2004 09:28:46 +0300
----------------------------------------------------------------------------
----
Dear Statalist members,
I want to estimate a regression model where the dependent variable is a
strictly positive integer number which is also bound from above (ie, there
is a maximum value). Obviously, OLS (regress) can do it. However, what
alternative procedures does STATA offer?
Thanks in advance for your assistance
Apostolos Ballas
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
