Re: st: Trouble with ml init and convergence

[Richard Williams <Richard.A.Williams.5@ND.edu>]

There are at least two other programs that will do ordered probit: 
oglm and gologit2, both available from SSC.  I have no reason to 
expect they will succeed where others have failed, but you can 
try.  Sometimes adding the -difficult- option to them will make the 
results converge.  My guess is that the model is just too complicated 
or the data are being spread too thin, but I haven't looked at your 
other messages yet so I can't say.  If you happen to have SPSS or 
PASW or whatever it is called these days, you might see if the PLUM 
routine can handle it.

At 08:11 PM 3/10/2010, Dana Chandler wrote:
> Hi fellow statalisters -
>
> I am working with an ordered probit that is having difficulty
> converging. I don't have that much experience with maximum likelihood
> estimators in Stata and think I'm probably missing something obvious.
> I would greatly appreciate any help or insights that you guys have.
>
> I wrote once before when the canned oprobit command wasn't converging
> (http://www.stata.com/statalist/archive/2010-01/msg00631.html). Since
> then, I have figured out how to program an ordered probit using ml lf
> thanks to comments from helpful statalisters and Gould's Stata MLE
> book. However, I still can't figure out how to get my function to
> converge.
>
> I'm estimating a model where I have a cross-section that includes the
> population and the number of religious organizations within several
> hundred towns. Each town has 0 to 5 organizations (top-coded) and I'd
> like to estimate an ordered probit where the outcome is the number of
> religious organizations and the predictors are population or some
> transformation of the population variable.
>
> In the simplest model, I estimate "oprobit outcome pop". This works
> fine. However, for another specification, I want to allow population
> to have a differing effect on the outcome variable depending on how
> many organizations are already in the town. Hence, I create dummies
> pop_w0-pop_w5 where I calculate DUMMY(outcome>=N)*pop. Next, I
> estimate "oprobit outcome pop pop_w1 pop_w2 pop_w3 pop_w4 pop_w5".
> This DOESN'T work which is why I start to use the oprobit that I
> programmed with ml lf where hopefully I can set initial values to help
> it converge. Unfortunately, I'm still having trouble with that.
>
> I have attached a link to a stata dataset and example do file that
> explain the methodology
> (http://www.4shared.com/dir/33394700/3eb89ce2/statalist.html). I
> haven't had any luck setting initial values and would greatly
> appreciate anyone's suggestions.
>
> Thanks everyone,
> Dana
> *
> *   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/

-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu
WWW:    http://www.nd.edu/~rwilliam

*
*   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/