Dear Statalists,
Is there a difference between using zinb with the option exposure and using zinb with the option constraint on Stata 17?
When comparing the zinb models in the example below, estimated coefficients slightly differ in the last decimal space.
Is there a difference in the two procedures?
Is there a difference between using zinb with the option exposure and using zinb with the option constraint on Stata 17?
When comparing the zinb models in the example below, estimated coefficients slightly differ in the last decimal space.
Code:
webuse fish gen lnPersons=ln(persons) constraint define 1 lnPersons=1 zinb count livebait, inflate(child camper) exposure(persons) zinb count livebait lnPersons, inflate(child camper) constraint(1)
