Jeff,
Thank you for your very informative answer.
I guess the bottom line is that if one cannot/does not want to use
analytical gradient for a model that includes covariates, then one is
better off sticking with -ml-?
Antoine
Jeff Pitblado, StataCorp LP wrote:
> Antoine's observations to the contrary are due to the choice of evaluator
> types in the comparison. In fact having predictors in the probit model
> accentuates the timing difference between the -v0- and -lf- evaluator types.
*
* 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/
