Regression based on the returns of males, while excluding the returns of females and missing values

[Victoria Rogers]

Cannot delete the topic but it seems that I found the solution after all.

[Maarten Buis]

The problem with your Male variables is that it is missing when the person is female. This is not what you want. Instead you can do:

gen byte Male = 0 if !missing(gender)
replace Male = 1 if gender == 0
regress return risk_premium if Male == 1

You seem to want to do this separately for different (groups of) dates, but you neither mention this in your question nor does your code really do this. You could look at help statsby or help rolling.

[Carlo Lazzaro]

Victoria:
I would follow this route:

gen alpha_male3=.
regress return risk_premium if Male==0
replace alpha_male3=_b[_cons] if Male==0

Kind regards,
Carlo

[Victoria Rogers]

Thank you for the help. Hopefully you can also help me with a very difficult problem.
http://www.statalist.org/forums/foru...f-observations

[Carlo Lazzaro]

Victoria:
please see my reply at http://www.statalist.org/forums/foru...rvations/page2 #20

Kind regards,
Carlo