The DHS Program User Forum: Domestic Violence

Home » Topics » Domestic Violence » Missing records?

Show: Today's Messages :: Show Polls :: Message Navigator

Missing records? [message #23851]

Thu, 16 December 2021 10:07

Mason92
Messages: 33
Registered: December 2021

Member

I need help in understanding as to why some records/respondents have missing values:

For DHS Liberia 2019-20
The total number of women who completed the 'Domestic Violence' was 3120. But for 789 women no information is available on the intimate partner violence perpetrated by husband/partner. I need help in understanding as to why for 789 women this information is coded as "." i.e. blank/missing please?

For DHS Rwanda 2019-20
The total number of women who completed the 'Domestic Violence' was 2788. But for 841 women no information is available on the intimate partner violence perpetrated by husband/partner. I need help in understanding as to why for 841 women this information is coded as "." i.e. blank/missing please?

Thanks!

Report message to a moderator

Re: Missing records? [message #23859 is a reply to message #23851]

Mon, 20 December 2021 19:29

Bridgette-DHS
Messages: 3199
Registered: February 2013

Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

Those are NA or "not applicable" cases. They are not "missing". If you go to the beginning of the DV module in the questionnaire for the Liberia 2019-20 survey, for example (Q1502 on page 511 of the final report) you will see a filter. There is a group of questions that are not asked of women who were never married and never lived with a man. For those women, those questions are coded with a dot (in Stata) for NA.

Report message to a moderator

Re: Missing records? [message #23903 is a reply to message #23859]

Thu, 06 January 2022 05:55

Mason92
Messages: 33
Registered: December 2021

Member

Thanks for the clarification, it was very helpful.

I have a follow-up question on Liberia DHS 2019-20: the variable d105a (ever been pushed etc. by husband/partner) has 2331 responses. So, is it correct that there were 2331 women who were ever in a relationship with a man? Secondly, for the variable v743a (person who decides on respondent's health care) has following categories with counts:

Respondent alone = 925
Respondent and husband/partner = 2,782
Husband/partner alone = 931
Someone else = 1
Other = 1
Total = 4654

However, for the 2331 women who were selected for the DV module and had ever lived with a man there are 343 'missing' records for v743a; am I correct that these are in fact missing records?

Thank you

Report message to a moderator

Re: Missing records? [message #23905 is a reply to message #23903]

Thu, 06 January 2022 07:57

Bridgette-DHS
Messages: 3199
Registered: February 2013

Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

In Stata, a dot (".") means Not Applicable. That is, a respondent will have a dot for v743a if, and only if, the question about who decides on her health care was not asked of her. You have to go to the questionnaire, find the question, and look at the preceding filters and skips to see who was asked this question and who was not. You would have to do the same thing for the DV questions. It can be tedious, but you will probably find that there is a good reason why you don't have more overlap of v743a and the subsample that was asked the DV questions.

Report message to a moderator

Re: Missing records? [message #23914 is a reply to message #23905]

Mon, 10 January 2022 04:19

Mason92
Messages: 33
Registered: December 2021

Member

Thank you so much!

Report message to a moderator

Previous Topic:	How to use weighting when working with both Domestic violence and women's data
Next Topic:	Afghanistan DHS 2015 - Zabul province data

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Nov 23 22:01:06 Coordinated Universal Time 2024