Statalist


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

Re: zero inflated beta [was: st: Information request]


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   stata list <statalist@hsphsun2.harvard.edu>
Subject   Re: zero inflated beta [was: st: Information request]
Date   Mon, 17 Aug 2009 19:35:17 +0000 (GMT)

--- Jay Verkuilen wrote:
> The ZI-Beta model is very hard to identify, too. The problem is that the
> beta includes a J-shaped distribution. It's hard to know if this will
> work.

I am not so sure that that is a problem. The identification comes from the 
fact that you believe that all zeros are modeled by the zero-inflation 
part, while all non-zeros are modeled by the beta part. So there is no 
ambiguity, even in a J-shaped beta distribution. The price you pay for 
this identification of the model is that you have to believe this (strong)
assumption. This is different for a zero-inflated fractional logit, where 
you believe that some zeros are modeled by the fractional logit part and 
some by the zero-inflation part.

--- Austin Nichols wrote:
> There is no intrinsic problem with a lot of zeros in the fractional
> logit model; those people could simply have very small bonuses in
> expectation due to covariate patterns (with an expected bonus of 17
> cents, you will realize a zero quite often).

In principle I agree, but Fabio was analysing data with more than 50% 
zeros, which I think is pushing this model too hard.

-- Maarten

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

http://home.fsw.vu.nl/m.buis/
-----------------------------------------


      

*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index