[Andrew Musau]

Pat, since parentfigure = 1 if the parent is a biological mother and 0 if the parent is a biological mother, then the command

list if parentfigure

will list the observations corresponding to biological mothers. On the other hand,

list if !parentfigure & !missing(parentfigure)

will list the observations of biological fathers. The negation character in Stata is the exclamation mark "!". Stata allows you to run a regression for a subsample using the if qualifier. Here is a reproducible example.

sysuse auto, clear
*FULL SAMPLE
regress price mpg weight

*SAMPLE OF FOREIGN CARS
regress price mpg weight if foreign

*SAMPLE OF DOMESTIC CARS
regress price mpg weight if !foreign

So, as you see above, you can always do analysis on a subsample while the data is in long layout.

[Pat Burns]

Andrew: Thank you so much for your response. I read the "reshape long" information as you had suggested in a prior post, so no longer have the following syntax:
expand 2, gen(x) gen X1HPAR= cond(x==0, X1HPAR1, X1HPAR2) gen parentfigure = cond(X1HPAR == 1, 1, cond(X1HPAR == 2, 0, .)) label define parentfigure 1 "BIOLOGICAL MOTHER" 0 "BIOLOGICAL FATHER" label values parentfigure parentfigure tab X1HPAR
Instead I used this:
reshape long X1HPAR, i(CHILDID) j(parentfigure) (note: j = 1 2)
now parent figure =1 or =2 I then generated Mother's Education (and a number of other variables in a similar way) as follows:

gen MOMED = X12PAR1ED_I if X1HPAR == 1

(16,247 missing values generated)

When I try to regress the variables that I created from X1HPAR as above (DADED),I get a message that there are no observations?

I hope I have not confused you, I certainly have confused myself!

[Pat Burns]

Andrew: Also, I am not limiting my regression analysis to a subsample. I have a number of other variables that I created that are unrelated to those generated by using the X1HPAR variable.
For example here is my regression command:
reg ZX1MTHETK4 i.MOMDEG##i.DADDEG MOMAGE DADAGE MOMOSCR DADOSCR MOMWORK ///
> DADWORK FEMALE CHILDAGE CHILDRACE i.SIBLINGS ///
> KDGREPEAT NONENG SINGPAR PRIVATE i.PKCARE i.SCHLOCATE i.INCOME
no observations
r(2000);