Why using the probit estimation in first stage as opposed to OLS? in Heckman sample selection model?

[Sarah Adem]

I am using the Heckman sample selection model to study the impact of having a home care provided by parent or a paid helper on the wage of high and low skill females.

I don't have any problem applying the estimation. I just want to under stand why in Heckman sample selection model we use probit in the first stage ?


thank you

[William Lisowski]

Because in the first stage you are modeling the initial outcome that the dependent variable was selected to be observed, so an appropriate model is needed for a 0/1 outcome (not selected/selected). In the second stage you model the dependent variable's observed value, given that it has been observed.

Greater detail in the documentation for the heckman command in the Stata Base Reference Manual PDF included in your Stata installation and accessible through Stata's Help menu. Or run

help heckman

and click on the link at the top to the fuill PDF documentation. The Remarks and examples section has a general discussion of the methodology.

[Sarah Adem]

thank you William that is really helpful .