Creating sibling-pairs using unique id

[Devon Smith]

Hi:

I have a dataset that contains the following variables: unique household id (hhold), unique household member id (id) within each household, a variable that defines relationship with head of the household (relationship). I want to create a dataset where each row represents a unique sibling pair within the household. Specifically, individuals with the relationship values 1 and 7 would form one type of sibling pair, 2 and 10 would form another, individuals with values 3 would be siblings as well, so would individuals with values 5 and 5, 8 and 8. How would I go about doing this?

[Nick Cox]

A data example would help (FAQ Advice #12).

This may help. Code not tested at all.

gen work = inlist(relationship, 1. 7) 
bysort hhold (work) : gen pair1_7 = work == work[_N] & work == 1 & work[_N-1] == 1 
by hhold : gen other1_7 = cond(_n == _N & pair1_7, id[_N-1], cond(_n == _N - 1 & pair1_7, id[_N], .)) 

replace work = inlist(relationship, 2, 10)
bysort hhold (work) : gen pair2_10 = work == work[_N] & work == 1 & work[_N-1] == 1 
* compare previous case 

replace work = relationship == 3 
* and so on

What should happen if e.g. 3 people satisfy any such criterion?

See also https://journals.sagepub.com/doi/pdf...867X0800800414

[Nick Cox]

https://www.statalist.org/forums/for...ng-identifiers appears related.

Please don't run such closely related threads at the same time -- or at least give cross-references.

[Devon Smith]

Thanks, Nick! Sorry for the double-post.