Re: st: logistic in version 11

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Tue, 16/3/10, Peter Goff wrote:
> The code below is from a do-file that
> worked well under version 10 but no longer functions under
> version 11 (log likelihood is perpetually backed up).
> 
> . logistic reporting_whi n_tested_whi
> 
> "n_tested_whi" is the sum of all white students tested.
> Oddly enough, the logistic command runs if I substitute in
> any of the grade-levels that contributed to the
> "n_tested_whi" variable. For example:
> 
> . logistic reporting_whi g03_whi
> 
> works under both version 11 and version 10. I'd like this
> to run under both versions - any suggestions as to why one
> variable would work with version 10 but not version 11 would
> be appreciated!

Sounds to me like your model is barely identified, and small
changes in the algorithm from version 10 to 11 pushed your 
model "over the edge". One thing you might want to look at
is if almost all the zeros in reporting_whi are below a 
certain value of n_tested_whi, and almost all the 1 above. 
The effect of n_tested_whi is not identified in the extreme 
case that this seperation is perfect, and I could imagine 
that Stata gets into trouble when the separation becomes 
very near to perfect.

Hope this helps,
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/