RE: Subject: st: Stata & LIMDEP comparison for count models

["Kaganova, Yevgeniya" <ykaganova@rti.org>]

Thank you so much for your answers! One of the models that we were interested in running was a hurdle model , but I found that Stat can do it http://www.stata-journal.com/sjpdf.html?articlenum=st0040 , and then found a chapter in Long and Freese book. But I was also asked to get some general information   on packages capabilities. 

Thank you again. Eugenia.

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of David Jacobs
Sent: Monday, May 03, 2010 12:35 PM
To: statalist@hsphsun2.harvard.edu
Subject: Re: Subject: st: Stata & LIMDEP comparison for count models

While it is true that Limdep has a few more count models including panel ZIP models and maybe hurdle models, Limdep is far more difficult to use.

For example, the older versions of Limdep I've used do not handle missing values well.  In addition to this important difficulty, Limdep offers few of the convenience features that make Stata so convenient.  And graph quality is not as strong in Limdep.  Data manipulations are considerably more difficult as well.

My advice:  Purchase both packages if you can afford them.  But if you need to restrict your purchase to just one package, Stata clearly is superior for most tasks.

At 12:20 PM 5/3/2010, you wrote:
>I just saw your query. Apologies for the delay.
>
>LIMDEP is a very complete package for modeling counts, but still lacks 
>in certain areas, which Stata has. Stata, LIMDEP, and recently R offer 
>for more count models than other software. They are in a league by 
>themselves in this respect.
>
>There are a host of count models. I think all of the basic and perhaps 
>not-so basic, count models are offered in both Stata and LIMDEP. LIMDEP 
>has a few more count models for panel data than Stata; eg random 
>coefficient negative binomial, and offers the full range of truncated 
>and censored Poisson and NB models. Official Stata does not  have 
>censored count models, for example, but user authored censored count 
>models can be downloaded from the SSC site. LIMDEP does not have 
>various count models that are available as user created ado files; eg 
>finite mixture models, quantile count models. In many respects it's a 
>trade off.
>
>LIMDEP may have a few more count models, but many are seldom used. 
>Stata is a really very good package for modeling counts, and should be 
>satisfactory for most needs. New models are being created by someone 
>within the Stata community all of the time.
>
>Joseph Hilbe
>
>
>
>
>
>
>Hilbe, Joseph M. (2006). “A Review of LIMDEP 9.0 and NLOGIT 3.0”, 
>The American Statistician, Vol 60, Nu 2 [May 2006].
>Hilbe, Joseph M. (2005), "A Review of Stata 9.0", The American 
>Statistician Vol. 59, Nu 3 [Aug 2005] 14 pages.
>
>From: "Kaganova, Yevgeniya" <ykaganova@rti.org>
>Subject: st: Stata & LIMDEP comparison for count models Dear 
>Statalisters, I  need to compare Stata and LIMDEP capabilities for 
>count models.
>Is there a good source where I can read about it ?
>Thank you. Eugenia
>
>
>*
>*   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/


*
*   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/

*
*   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/