I am trying to read on how to interpret interaction effects in Heckman's selection model and come across with this article ￼On Marginal and Interaction Effects: The Case of Heckit and Two-Part Models. The authors mention that they use a stata command to calculate the effects yet to specify what they used.
Let assume the regression of interest: $$ y = \alpha_0 + \alpha_1 x_1 + \alpha_2 x_2 + \alpha_i x_1*x_2 + \beta X + \epsilon $$ where $y$ is the limited dependent variable and $X$ is a set of controlling variable.
I am interested in interpreting $\alpha_i$, that is effect of the interaction term between $x_1$ and $x_2$.
Could any one please point out how to calculate, and interpret, the interaction effects in Heckit/Tobit model using Stata?