The DHS Program User Forum: General Data Questions

Home » Data » General Data Questions » Data inconsistencies

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

Data inconsistencies [message #23063]

Wed, 07 July 2021 10:28

TD
Messages: 3
Registered: July 2021

Member

I'm currently carrying out data analyses on the most 2018 Nigeria and 2019 Sierra Leone datasets.

In the individual recode documents, there is a section titled "inconsistencies in the data file" for both countries. Please do you know if the DHS provides guidance on how to deal with these inconsistencies when completing the analyses?

Report message to a moderator

Re: Data inconsistencies [message #23064 is a reply to message #23063]

Wed, 07 July 2021 11:00

Bridgette-DHS
Messages: 3230
Registered: February 2013

Senior Member

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

On the basis of my own experience, I'd say it is very unlikely that you will actually encounter any of these inconsistencies in your analysis. For all practical purposes you can ignore the list. But if you have a table or model that includes a variable with known inconsistencies, you can define and use an alternative variable. For example, add the letter "r" for "recoded" and recode it as a dot (which normally means Not Applicable) for the inconsistent values. It is important not to save "corrected" variables under the original name. I would not use a statistical method to adjust for inconsistencies or to impute "corrected" values. If you change values in any way, it's helpful to state that you have done so. Otherwise, other users cannot match your results.

Report message to a moderator

Re: Data inconsistencies [message #23071 is a reply to message #23064]

Fri, 09 July 2021 06:20

TD
Messages: 3
Registered: July 2021

Member

Thank you

Report message to a moderator

Previous Topic:	Cluster sampling
Next Topic:	Where to get the country specific variables/questions that begin with 's'

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Dec 13 20:57:10 Coordinated Universal Time 2025