# It it okay to combine the six 1% samples of the 1970 US Census into a 6% sample?

**URL:** <https://forum.ipums.org/t/it-it-okay-to-combine-the-six-1-samples-of-the-1970-us-census-into-a-6-sample/276>\
**Category:** General\
**Created:** [September 13, 2013, 10:59pm UTC](https://forum.ipums.org/t/it-it-okay-to-combine-the-six-1-samples-of-the-1970-us-census-into-a-6-sample/276 "2013-09-13T22:59:00Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![mgdataprj](https://avatars.discourse-cdn.com/v4/letter/m/9dc877/32.png) [@mgdataprj](https://forum.ipums.org/u/mgdataprj)\
**Post date:** [September 13, 2013, 10:59pm UTC](https://forum.ipums.org/t/it-it-okay-to-combine-the-six-1-samples-of-the-1970-us-census-into-a-6-sample/276/1 "2013-09-13T22:59:00Z")

</div>

The MPC provides six samples from the 1970 US Census (1% metro fm1, 1% metro fm2, 1% state fm1, 1% state fm2, 1% neigh fm1, and 1% neigh fm2).

Can I combine the six sample to obtain a 6% sample? - Assuming that I divide the sample weight by 6 (e.g., new\_weight = perwt/6).

---

<div class="post-metadata">

**Author:** ![Brandon\_Trampe](https://avatars.discourse-cdn.com/v4/letter/b/35a633/32.png) [@Brandon\_Trampe](https://forum.ipums.org/u/Brandon_Trampe)\
**Post date:** [September 16, 2013, 3:50pm UTC](https://forum.ipums.org/t/it-it-okay-to-combine-the-six-1-samples-of-the-1970-us-census-into-a-6-sample/276/2 "2013-09-16T15:50:00Z")

</div>

Within each form, the three samples were selected by an independent random draw with replacement. As a result, some cases could appear on both the Metro Form 1 and the State Form 1, for example. Some researchers choose to disregard this overlap due to the small number of cases and treat a combined 6% sample as if there are no duplicates.