st: Using expand

[Grethe Søndergaard <dkstatstata@gmail.com>]

Hi,

I have a dataset that includes full siblings from different families
(two full siblings from each family).

Variables:
id_child
id_mother
id_father
birthdate (of the child)

I only want to use full siblings with an age different less than 5
years. In order to do this, I am using expand:

by id_mother id_father, sort: gen gfreq = _N
expand gfreq, generate(dub)
sort id_mother id_father id_child
by id_mother id_father id_child: gen num_child = _n

by id_mother id_father: gen id_child2  = id_child[gfreq * num_child]
by id_mother id_father: gen birthdate2 = birthdate[gfreq*num_child]

gen age_difference = (birthdate2 - birthdate)/365.25
replace age_difference = age_difference*-1 if age_difference<0
keep if age_difference<5

After doing this, I want to revert to the original observations:
keep if dub==0

- but is that the right way to do it?
The reason why I am asking is, that the number of observations before
I use dub==0 isn't exactly twice the number of observations after I
use dub==0. (N=1.426.648 before and N=713.466) - And shouldn't it be -
or am do I misunderstand something?
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/