I am posting with regards to Heckit Models (for self-selection bias) and the inclusion of an interaction term. It has already been extensively documented in the literature the impacts of interaction terms in non-linear model and the corresponding interpretations- I have adjusted accordingly.
I need to test my model for self-selection into the public/private sector which is a dummy that makes up one half of my interaction term in my original probit model.
In a footnote, the following article states that "No canned program is available for calculating the interaction effects results from the Two-Part Model, whereas the most recent version of Stata can calculate interaction terms for the HECKMAN model"
Frondel, Manuel, and Colin Vance. "On Interaction Effects: The Case of Heckit and Two-Part Models." Jahrbücher für Nationalökonomie & Statistik 233.1 (2013).
Is anyone familiar with the code for running a Heckman model with the inclusion of just one interaction term? Is it just a straight forward Heckman command?
Any information is greatly appreciated.
Anna
I need to test my model for self-selection into the public/private sector which is a dummy that makes up one half of my interaction term in my original probit model.
In a footnote, the following article states that "No canned program is available for calculating the interaction effects results from the Two-Part Model, whereas the most recent version of Stata can calculate interaction terms for the HECKMAN model"
Frondel, Manuel, and Colin Vance. "On Interaction Effects: The Case of Heckit and Two-Part Models." Jahrbücher für Nationalökonomie & Statistik 233.1 (2013).
Is anyone familiar with the code for running a Heckman model with the inclusion of just one interaction term? Is it just a straight forward Heckman command?
Any information is greatly appreciated.
Anna
