Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Tobit interaction effect


From   Maarten Buis <maartenlbuis@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Tobit interaction effect
Date   Wed, 9 May 2012 12:32:17 +0200

On Wed, May 9, 2012 at 12:09 PM, Ebru Ozturk wrote:
> But, it has been discussed that there is no Stata code for estimating interaction effects in Tobit by Norton, Wang & Ai, 2004. So, how can we just add them and interpret the results?

They were wrong, as long as you interpret your results in terms of the
latent outcome. This is what I meant when I said:

"The resulting model is easy to interpret in terms of the expected
latent variable, but gets a lot more complicated for other outcomes
like the probability of being censored, or the expected outcome
conditional on not being censored."

The logic is that a Tobit model is just a linear model terms of the
latent variable, so none of the problems discussed by Norton and
colleagues apply to that interpretation of the Tobit model. It is
non-linear in terms of the probability of being censored or the
expected outcome condition on not being censored, etc. If those are
the outcomes of interest you will have to do more work.

-- Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index