On 1/10/07, [email protected] <[email protected]> wrote:
Hello.
I am working on a case of using GLLAMM for structural equation modeling with a
binary outcome - it is a path model, with one latent variable, and I received
the message (Stata 9.2):
"
numerical derivatives are approximate
flat or discontinuous region encountered
"
If your estimation process strandled into an area with the linear
predictor around 60, then indeed you won't see much change in the
likelihood around that area even if you make huge steps of 10. And
also you might have identification problems as your random effect
variance seems to depend on the same variable as your linear
predictor, so what you have as a result is the equation that looks
like (a + bx)/(c + dx) which may not be identifiable unless you fix
some parts of it.
--
Stas Kolenikov
http://stas.kolenikov.name
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
