Statalist


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

RE: st: nonconvergence of switching probit when data weighted


From   "Paula Albuquerque" <pcma@iseg.utl.pt>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: nonconvergence of switching probit when data weighted
Date   Thu, 12 Mar 2009 13:40:57 -0000

Thank you for your suggestion. I tried it. This time, even the estimation that had found a rho=1 (but had produced results) declares that " flat log likelihood encountered, cannot find uphill direction". It seems there must be something else...

Paula

--- On Thu, 12/3/09, Paula Albuquerque <pcma@iseg.utl.pt> wrote:
> I find no negative weights and I don't think there are
> any exceptionally large ones.

I usually get a bit suspicious with weights larger than 3, especially since you have quite a small sample sze for a rather complicated model like -movestay-. To see if this is the problem you could try something like this:

gen junkweight = min(multipli, 3)
movestay ... [pw=junkweight], ...

If this does converge, than you know at least what the problem is.

-- 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/


*
*   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