marginal effect after xtlogit, re nocons for a hazard model (person-period data)

[Ayesha Bukar]

Hi Statalist,

Apologies, I am not yet familiar with using dataex. I am running a hazard model using xtlogit using the following command.

xtlogit ibn.period age agesq sex married . . . , re vce(robust) nolog nocons

Then I run the post estimation command : margins dydx(*), predict(pu0)

One of the periods gets dropped. How do I impose a condition to ensure that I report the marginal effect for the period that gets omitted after the margins command, to match the results I have reported for the coefficients with noconstant? I am reporting the output in a nested table so I wish to have a consistent output to display.

Thanks

[Ayesha Bukar]

I have attached the output here. One thing I have also noticed is that while the coefficient is not significant in some periods, the marginal effects is, and vice versa. Please can anyone help?