I have a zero-inflated negative binomial model (ZINB) for highly skewed, high zero-count data, $n=6800$. The null model does not reject the ZINB model, and Stata count fit also indicated ZINB over other count models.

Yet, when I run the model with more than the minimum of covariates, the inflation intercept has a coefficient but no error, no $t$, no $p$-value. Does this mean the model is incorrect? Or just the zero-inflated assumptions? Any suggestions?


Are you using Stata's zinb command to run your regressions?

When there are no standard error estimates, the things that comes to mind as possible explanations are that the model may be too collinear, or that there may be zero-cells for some variables.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.