Connecting twins in a dataset using twin status and family ID

[Stephen Cranney]

I thought this would be rather simple, but I can't figure it out.

I have a dataset that includes twin pairs. I want to make a twin variable (e.g. variable X is person i's value for Y, and variable X2 is the value of person i's twin for Y) so that I can do MZ and DZ twin correlations. I have family ID and twin status. I tried to do it by doing various versions of by X: replace twinvariable[_N]=_N-1, etc., but it isn't quite doing it. Is there some way to tell Stata, "look in this variable Y-defined group. If there are two cases in this group that have the same value for variable X, insert a value in variable X2 that consists of the value of the other pair case"?

Many thanks

[Jorge Eduardo Perez Perez]

What about this?

clear
* Create some data
input family x
1 2
1 3
2 1
2 4
end
* Generate an identifier for twin within family
bys family: gen twin=_n
* Create twin variable
gen x_twin=.
* Replace by value of other twin within each family
bys family: replace x_twin=x[2] if twin==1
bys family: replace x_twin=x[1] if twin==2

[Nick Cox]

This may help too:

SJ-8-4 dm0043 . Tip 71: The problem of split identity, or how to group dyads
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. J. Cox
Q4/08 SJ 8(4):588--591 (no commands)
tip on how to handle dyadic identifiers

For example, within groups of 2, 3 - _n is always the index of the other observation. (If _n == 1, then 3 - _n == 2 and vice versa.)

For that and other tiny tricks, the pdf is free to all:

http://www.stata-journal.com/sjpdf.h...iclenum=dm0043

[Stephen Cranney]

Thanks! The code above worked once I dropped all the non-twin observations (then I re-merged the twin variables back in with the main dataset).