The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » General Data Questions » Not in Universe - Woman Surveys (Data)
Not in Universe - Woman Surveys [message #15840] Tue, 25 September 2018 14:42 Go to next message
micpender is currently offline  micpender
Messages: 1
Registered: September 2018

I'm a little confused by the "Not in universe" labeling - in particular how it might differ from missing values. I used the IPUMS-DHS website to pull several West African countries' women surveys over several years and I have several variables with a "Not in universe" label along with missing values.

I'm kind of assuming that "not in universe" simply means that the question didn't apply to that woman? For instance, the variable DELDOC_01 being equal to "Not in Universe" simply means that that woman didn't give birth in the last three to five years, because if she did she could have answered in either the positive or negative to having some kind of medical professional deliver the previous baby. However, assuming that my assumption is right, how do these women differ from the missing values? Or am I just completely off base and have no idea?

Thank you in advance for your reply.
Re: Not in Universe - Woman Surveys [message #15973 is a reply to message #15840] Fri, 12 October 2018 14:34 Go to previous message
boyle014 is currently offline  boyle014
Messages: 73
Registered: December 2015
Location: Minneapolis
Senior Member
Good question. Glad you asked.

You are correct that "not in universe" means "the question does not apply to this person." Respondents are assigned "not in universe" due to skip patterns in the questionnaire. For example, a woman who says she does not work will not be asked whether she is paid in cash or in kind. She will show up as "not in universe" for the latter question.

You will find more exact information if you consult the "Universe" tab of a variable description. For example, DELDOC_01 for Bangladesh 2011 has the Universe of: Ever-married women age 12-49 who gave birth in the 5 years before the survey. You get to this and all other tabs by clicking on a variables name.

Cases coded as "Missing" should theoretically have been covered by the question but have no response in the data. A woman who refused to answer a question would be coded as "Missing" for that variable; if an interviewer accidentally failed to ask a question to a woman theoretically covered by that question, then the woman would also be given a "Missing" response code.

Hope this helps.

Professor Elizabeth Boyle
Sociology & Law, University of Minnesota, USA
Principal Investigator, IPUMS-DHS
Previous Topic: Missing Data Question
Next Topic: Disability
Goto Forum:

Current Time: Mon Nov 30 06:08:35 Coordinated Universal Time 2020