Oaxaca Decomposition

[Siham Beshir]

I'm trying to run an oaxaca decomposition for differences in log of wages: White men vs. Black men. I have dummy variables for Black (black=1 and all other races=0), White (white=1 and all other races=0) and Female (women=1, men=0). However, I'm having trouble integrating both race and gender in the "by" part of the oaxaca. I also have indicators for industry and state which I'm not sure I can use with the oaxaca. Other control variables are age, education and migration status.

[William Lisowski]

generate group = .
replace group==1 if white==1 & female==0
replace group==2 if black==1 & female==0
label define group 1 "White men" 2 "Black men"

may be what you need.

[Siham Beshir]

Thank you William! So after I do the regression, how do I incorporate this code^ into the oaxaca. Would it be "oaxaca dependent variable independent variables ,by(group)"? Then when I'm interpreting the results group 1 is White men aka the reference group?

[William Lisowski]

Yes, "by(group)", and yes, "group 1 is White men".

[Siham Beshir]

Thank you so much!!