Dear Edlira,
I had also some convergence problems with Gllamm and mlogit link, in particular when I tried to define the random effect as you have done. Have you looked at the Gllamm Manual? In the 9 chapter there are two examples of gllamm with mlogit and how to expand the data to be in appropriate form. When I followed the Manual and used random intercept model my estimation converged fine, although it took long time.
I do not know if this helps you, but one advice a got from Professor Rabe-Hesketh was to start by reduced the number of integration points, then use matrix a=b(e) command and use from(a) option in the gllamm command.
Best,
Laura
-----Alkuperäinen viesti-----
Lähettäjä: Edlira Narazani [mailto:[email protected]]
Lähetetty: ma 20.9.2004 14:22
Vastaanottaja: [email protected]
Kopio:
Aihe: st: Still gllamm
Dear all,
1. I'm using gllamm as follows and at the end of iterations I dont get t-values
or standard errors. Therefore I cant use it without estimation statisctics.
gllamm depvar indepvar, robust i(nfirm) link(mlogit) fam(binom) nocons adapt
trace
Is it ok? Do you know how to resolve it?
2. when I want to include some random effect (33 firms in 23 years follow some
choices which take value 0, 1,2)I have done in this way. At the end of
iterations I get the note: Hessian matrix is not symmetric.maximization did not
converge.
gen const = 1
eq intercept: const
eq slope: year
gllamm depvar inddepvar, robust i(nfirm) link(mlogit) family(binom) nrf(2)
eqs(intercept slope) adapt nocons trace
Is it right?
I have written even previously to you regarding these questions but no reply so
far. So I would really appreciate your answer.
Edlira
----------------------------------------------------------------
This message was sent using IMP - Facolta' di Economia - Torino.
*
* 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/
<<winmail.dat>>
