Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: bsample and convergence


From   "Oliver Rivero-Arias" <oliver.rivero@dphpc.ox.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: bsample and convergence
Date   Tue, 16 Sep 2008 12:40:15 +0100

Dear all,

I have prepared a looping exercise using the command "while" where in each loop bootstrap samples from an original dataset (using bsample) and a multinomial logit (mlogit) in each bootstrap sample are conducted. I have programmed 5000 loops and each loop saves the probabilities of being in a particular category of my categorical dependent variable.

However, in some bootstraps samples the multinomial logit does not converge (error r(430)) and therefore stops all my simulations. Is there a code I could use to tell STATA that in case a particular multinomial logit from a bootstrap sample does not converge, just discard this loop and draw another bootstrap sample?

Any help is greatly appreciated.

Oliver Rivero-Arias


*
*   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