Dear Stata Users,
I'm running a self-written program with several svy: probit models of the form: retired= age + female (coded as 1/0), with different weights and the `difficult' option and get the error `convergence not achieved - an error occurred when svy executed probit'.
When I use svy: logit instead, everything works fine (however, for comparison with a heckprobit model I want to estimate the coefficients via probit models).
Is there any general difference in the convergence of probit and logit models or does someone of you have an idea for my situation?
Thank you for your help,
Melissa
I'm running a self-written program with several svy: probit models of the form: retired= age + female (coded as 1/0), with different weights and the `difficult' option and get the error `convergence not achieved - an error occurred when svy executed probit'.
When I use svy: logit instead, everything works fine (however, for comparison with a heckprobit model I want to estimate the coefficients via probit models).
Is there any general difference in the convergence of probit and logit models or does someone of you have an idea for my situation?
Thank you for your help,
Melissa
Comment