How to assign a random value for missing months of birth?

[Matthew Williams]

Dear all,

My data has about 700 observations with missing months of birth and I am trying to fill in those missing values. I am thinking of two options, among others. The first option is to choose 30 June or 1 July for the missing values because those dates are the middle of year. The second one is to assign a random number from 1-12 to the missing values. Any suggestion on how to do the 2nd option in Stata is highly appreciated. Thank you.

Data

* Example generated by -dataex-. For more info, type help dataex
clear
input double id byte mob int yob
42063 .    .
38067 . 1982
46079 . 1973
38024 . 1984
31047 . 1977
33118 . 1981
42059 . 1977
42027 . 1975
38010 . 1984
37061 . 1965
38017 . 1979
46038 . 1970
12003 . 1981
33046 . 1988
 8098 . 1983
37092 . 1974
33071 . 1986
42038 . 1988
31033 . 1967
42067 . 1975
end

[Clyde Schechter]

replace mob = runiformint(1, 12) if missing(mob)

Also, to make this reproducible, before running this code you should set the random number generator seed if you have not already done so. -set seed pick_a_number_any_integer_will_do-. That way if you re-run this code in the future, you will get the same results.

[Matthew Williams]

Thank you Prof. Clyde for the code as well as suggestion on the use of setting seed.