Is the interaction term in my probit-model significant?

Hey guys

I have a model with two independent dummy variables (white/black and male/female). My dependent variable is a dummy on getting a job (yes/no).

By using simple math I can calculate the probabilities of getting a job for each group. White men: 36%; White women: 30 %; black men: 19 %; and black women 25 %.

This can be fitted in an interaction model: Getting a job = 0,36 - 0,17 (Black) - 0,06 (women) + 0,012 (black*woman).

Now I want to do this in stata.

My questions are: 1) How do I know if the interaction term is significant, 2) would it make sense to calculate a marginal effect of the interaction term?


I do as follows:
*I run a probit model:

Probit Getjob i.race i.gender i.race#i.gender


This output reports that my interaction term is signifcant. Is that what I need to answer question 1)? Or should I do something else?


*Then I use the margins command:
margins race#gender

Here I receive same results as I wrote above: White men: 0,36; White women: 0,30; black men: 0,19; and black women 025.

But I dont know how to get stata to calculate the interaction term without me computing the interaction term myself by generation race*gender. Is that ok?


cheers