# Identifying households participating in Guinea83 AND Guinea96?

**URL:** <https://forum.ipums.org/t/identifying-households-participating-in-guinea83-and-guinea96/695>\
**Category:** INTERNATIONAL\
**Created:** [October 29, 2014, 12:42pm UTC](https://forum.ipums.org/t/identifying-households-participating-in-guinea83-and-guinea96/695 "2014-10-29T12:42:00Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![MaxMueller](https://avatars.discourse-cdn.com/v4/letter/m/779978/32.png) [@MaxMueller](https://forum.ipums.org/u/MaxMueller)\
**Post date:** [October 29, 2014, 12:42pm UTC](https://forum.ipums.org/t/identifying-households-participating-in-guinea83-and-guinea96/695/1 "2014-10-29T12:42:00Z")

</div>

Is there a way to identify households/ persons across the samples in Guinea in 1983 and 1996, i.e. to exploit some panel data potential? As far as I understand the variable “serial” is used as a unique identifier in a given sample, but would be given to another household in the subsequent sample?!

---

<div class="post-metadata">

**Author:** ![Tim\_Moreland](https://avatars.discourse-cdn.com/v4/letter/t/f1d935/32.png) [@Tim\_Moreland](https://forum.ipums.org/u/Tim_Moreland)\
**Post date:** [October 29, 2014, 9:06pm UTC](https://forum.ipums.org/t/identifying-households-participating-in-guinea83-and-guinea96/695/2 "2014-10-29T21:06:00Z")

</div>

Unfortunately, this is not possible. In each year, the [Guinea census](https://international.ipums.org/international/sample_designs/sample_designs_gn.shtml) contains a random sampling of 10% of the households in the country. As a result, one would expect a minimal amount of overlap to exist, even if households/persons could be linked across samples. You are correct that “serial” is unique within a sample, but not across samples.