May I have to find an estimator that works better than Poisson when there are no zeros?

[Hm Saleh]

Hello,

I run a Poisson estimation but like the one with a log-likelihood of -7181.8, which was a failure because almost none of the coefficient estimates are significant.

I may have to find an estimator that works better than Poisson when there are no zeros (female without children).

I run NBREG and I got no Alpha value I think this regression is a failure too.

Can you guys give me any suggestions for regressions or ideas to solve this problem?

Thanks

[Clyde Schechter]

Take a look at -help tpoisson-.

[Joao Santos Silva]

Dear Hm Saleh,

We may be able to be more helpful if you tell us more about what you want to do, but one thing you can try is to model the count minus 1. Anyway, this looks more like a problem with the model than a problem with the estimator.

Best wishes,

Joao