Hi,
I’m wondering if anyone can tell me the correct way to use an offset
variable in a zero inflated poisson regression model. I’ve used the
following command which seems to test the offset variable as a component of
the zero inflation part of the outcome variable, but not the count part of
the outcome:
zip numtxdays cond, inflate(cond, offset(offset)) vce(cluster rctr) vuong
My question is whether or not there’s a way to account for the offset
variable while testing the count part of the model in Stata, and if so, is
it appropriate to do so? The reason I ask is that, if I use SAS PROC NLMIXED
or Mplus, I can include the offset variable in the test of the count part of
the outcome variable, as well. If I do, I get radically different results
than if it is only accounted for in the inflation part of the model.
Thanks for your help - any advice is greatly appreciated,
Jen Rose
*
* 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/
