difficulty with Running Heckman model to address sample selection bias

[Saleh sharmah]

Dear members
I am trying to run the Heckman model to solve any sample selection bias in my model. My regression is y=b b1x1, firms controls, industry FE, year FE, vce, (robust, firm_id)
my question is how to run the Heckman model with fixed effects and how to introduce the variable with dummy 0,1 to be implemented in the model?

thanks for your help.

[George Ford]

are you using the heckman command?

[Saleh sharmah]

Yes I’m . I am inexperienced with using Heckman regression. So how is it implemented in the regression and where does the selection bias dummy go exactly?

thanks for your help

[George Ford]

heckman command takes care of all that. basically, a probit is run as a first step, then an inverse mills ratio is computed from the predictions and inserted into the second stage regression. Stata will take care that the covariance matrix is correct (which it isn't if you force it).

heckman depvar [indepvars], select(depvar_s = varlist_s) [twostep]