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: "Can Your Results be Replicated?" (Stata error?)
From
Richard Williams <[email protected]>
To
[email protected], [email protected]
Subject
Re: st: "Can Your Results be Replicated?" (Stata error?)
Date
Fri, 20 Sep 2013 08:35:55 -0500
At 02:08 AM 9/20/2013, Maarten Buis wrote:
A solution for -oglm- would be to use the approximation of that model
as a set of independent binary logistic regressions as is used in the
Brant and Wolfe-Gould test. Once you have the data prepared for that,
you can use -_rmcoll- with the -logit- option to detect perfect
predictors. In addition you could then easily estimate those
independent binary logits, which would give very good starting values.
-- Maarten
Thanks Maarten. Are you perhaps thinking of gologit2 though? Or do
you think this would apply to oglm as well? (For anyone else who is
reading, both programs are available from SSC).
On Fri, Sep 20, 2013 at 2:03 AM, Richard Williams
<[email protected]> wrote:
> At 02:56 PM 9/19/2013, Jeff Pitblado, StataCorp LP wrote:
>>
>> As for Richard's experience with perfect predictors, we believe you can
>> expect
>> one of the following outcomes when dealing with perfect predictors:
>>
>> 1. The model will fail to converge because of the lack of a concave
>> Hessian.
>>
>> 2. The model "converged", but there are some extreme valued coefficients
>> with
>> just as extreme standard errors. Sometimes, but not always, the fitted
>> intercept will be extreme too. For -mlogit- with the -rrr- option, these
>> extreme values tend to translate to zero relative-risk ratios with missing
>> standard errors.
>
>
> Along those lines -- if you run the following code
>
> webuse repair, clear
> hetprob foreign b3.repair, het(b3.repair)
> oglm foreign b3.repair, het(b3.repair) link(p) log
>
> hetprob drags on for 16,000 iterations before giving an incorrect answer.
> oglm, on the other hand, gives an incorrect answer almost
immediately. Which
> I guess means that oglm is better. ;-)
>
> Anyway, you may want to add hetprob to the todo list.
>
>
>> --Jeff
>> [email protected]
>> *
>> * For searches and help try:
>> * http://www.stata.com/help.cgi?search
>> * http://www.stata.com/support/faqs/resources/statalist-faq/
>> * http://www.ats.ucla.edu/stat/stata/
>
>
> -------------------------------------------
> Richard Williams, Notre Dame Dept of Sociology
> OFFICE: (574)631-6668, (574)631-6463
> HOME: (574)289-5227
> EMAIL: [email protected]
> WWW: http://www.nd.edu/~rwilliam
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/statalist-faq/
> * http://www.ats.ucla.edu/stat/stata/
--
---------------------------------
Maarten L. Buis
WZB
Reichpietschufer 50
10785 Berlin
Germany
http://www.maartenbuis.nl
---------------------------------
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME: (574)289-5227
EMAIL: [email protected]
WWW: http://www.nd.edu/~rwilliam
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/