# How to interpret NIU for sampweight questions

**URL:** <https://forum.ipums.org/t/how-to-interpret-niu-for-sampweight-questions/853>\
**Category:** HEALTH SURVEYS\
**Created:** [April 6, 2015, 8:08pm UTC](https://forum.ipums.org/t/how-to-interpret-niu-for-sampweight-questions/853 "2015-04-06T20:08:00Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![albertsun](https://avatars.discourse-cdn.com/v4/letter/a/7bcc69/32.png) [@albertsun](https://forum.ipums.org/u/albertsun)\
**Post date:** [April 6, 2015, 8:08pm UTC](https://forum.ipums.org/t/how-to-interpret-niu-for-sampweight-questions/853/1 "2015-04-06T20:08:00Z")

</div>

I’m using the online analyzer tool and trying to fill in the blank in this sentence. “In 2013, TK% of adults in the US had their cholesterol checked by a medical professional.”

So I put cholchek1yr as the row, year as the column, and sampweight as the weight. That gave me NIU 23.6%, No 27.5%, Yes 46.8% and some small percentages as unknown. Now this is a question that was asked to all sample adults right?

Does that mean I should exclude the NIU responses and scale the rest up so that Yes would be .468/(.275+.468) ?

Or is the right answer that 46.8% of adults had their cholesterol checked?

A bit confused about how these studies work.

Thanks so much,

Albert

---

<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:** [April 7, 2015, 7:58pm UTC](https://forum.ipums.org/t/how-to-interpret-niu-for-sampweight-questions/853/2 "2015-04-07T19:58:00Z")

</div>

You should exclude the NIU responses. Since you are using [SAMPWEIGHT](https://www.ihis.us/ihis-action/variables/SAMPWEIGHT#description_section), and non-sample adults have SAMPWEIGHT=0, the NIU observations above consist entirely of respondents under the age of 18.

Notice that if you add a Selection Filter of “age(18-\*)” to your table, it will drop all of the NIU observations and nothing else. Equivalently, you could instead add a Selection Filter of “cholchek1yr(1-\*)” to explicitly exclude NIU observations. The ‘\*’ in my filters can be used in [SDA](https://sda.ihis.us/cgi-bin/sdaweb/hsda?harcsda+ih2013) to signify the highest (or lowest) valid value.

Hope this helps.