Statalist


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

st: Re: Backed up message stata list


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   stata list <statalist@hsphsun2.harvard.edu>
Subject   st: Re: Backed up message stata list
Date   Wed, 28 Oct 2009 18:48:06 +0000 (GMT)

--- On Wed, 28/10/09, Michaela Kesina wrote privately:
> one year ago you comment a question on the stata list. It
> dealt with the backed up message after a ml regression. 
> I´m estimating a fractional logit model with glm and have
> the same problem.
> I´ve tried to contact the person, who posted the question,
> but the email adress doesn´t exist any more. I´m very
> sorry for disturbing you, but you were the only person who
> has put a comment on this question.
> I´ve changed the algorithm with the option technique but
> this was seldom successful.

The way to ask this type of questions is to sent it to 
the Statalist, and not to persons privately, so I am 
forwarding this to the Statalist. This policy is discussed
on the Statalist FAQ:
http://www.stata.com/support/faqs/res/statalist.html#private

The fractional logit model is usually pretty stable, so
this indicates that there are some extreme patterns in 
your data. One such extreme pattern that might break the 
fractional logit model is when there is a large proportion 
of exact zeros or ones and there is some (combination of) 
categorical variable(s) that (almost) perfectly predicts that. 

I also remember a case where this type of model became 
unstable when a linear link function was specified (which 
means that you wouldn't be estimating a fractional logit,
but mistakes like that happen).

In summary, seeing this message for this model means that 
either your data is not suitable for this model, or there is
some error in your data, or there is something wrong with 
the way you specified the model (and as you didn't say how 
you specified your model, I can't check that).  

-- Maarten

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

http://www.maartenbuis.nl
--------------------------




Send instant messages to your online friends http://uk.messenger.yahoo.com 

*
*   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