Can I delete all the NIU variables in my data set?

I downloaded the ATUS-X for 2003-2014. I dropped all the observations where age is less than 40. I looked at the variable sex (tabulation) and 70% of the observations were NIU. Why does it show up like that? Can I just delete them without compromising the data?

This same issue comes up with marital status, highest level of education completed, retired, and other variables. They don’t all show up as 70% of the observations, but can I just drop all the NIU observations?


When creating a data extract, ATUS-X users are given a choice of four possible groups of people to be included in their data extract:

ATUS respondents
ATUS respondents and members of their households
ATUS respondents and ATUS non-respondents
ATUS respondents, ATUS non-respondents, and members of their households

On the codes page for SEX, you can look at how each of these options may change data output by selecting “case-count view” and clicking between each of the four sample groups. This will show many NIU cases when ATUS non-respondents are included in a sample, and your current extract may contain such individuals. The Defining Sample Criteria page can help you choose whether including non-respondents is necessary for your research goals.

If your current selection is appropriate for your needs, ATUS respondents can be identified as those with a RELATE value of 010: Self. Applying a filter to observe only these respondents will rid your samples of the NIU cases. However, if you are only interested in the ATUS respondents and want to change your sample criteria, you can revise your extract, change the Sample Members option back to the default: ATUS respondents, and resubmit.

I hope this helps.