The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » IPUMS Demographic and Health Surveys (IPUMS-DHS)  » What becomes of DHS weights after region harmonization by IPUMS-DHS
What becomes of DHS weights after region harmonization by IPUMS-DHS [message #18059] Tue, 03 September 2019 07:27 Go to next message
camilleatpik is currently offline  camilleatpik
Messages: 3
Registered: July 2019
IPUMS-DHS did a very useful job of harmonization of subnational regions (with each region keeping the same geographic footprint across samples).
From what I understand, IPUMS does this by merging or splitting subnational regions based on historical redefinition of administrative sub-divisions. As a result, a given country has the same administrative sub-division over all its waves. For example, IPUMS-DHS have merged Bengladesh'region which have been slit over time: Bengladesh has only 5 regions across all the waves in the IPUMS-DHS data while the in the latest subdivision one can find 7 regions.

Now I am wondering how does the IPUMS-DHS deals with DHS weights when it merges or split regions. DHS weights are put in order to adjust for the disproportionate sampling: small regions will be over-sampled and big regions will be under-sampled. This means that the DHS weighting is specific to the regional subdivision.
Does the IPUMS-DHS do some re-weighting to account for this ? If so, could you explain how it works, or where to get useful resource?
I would really appreciate any help !
Thank you
Re: What becomes of DHS weights after region harmonization by IPUMS-DHS [message #18662 is a reply to message #18059] Wed, 22 January 2020 14:32 Go to previous message
boyle014 is currently offline  boyle014
Messages: 73
Registered: December 2015
Location: Minneapolis
Senior Member
Thank you for the question. The weights are not affected by the regional harmonization. This because IPUMS DHS never splits a region to accomplish harmonization; we only merge regions. Within any merged region, the weights assigned to individuals are still valid.

Professor Elizabeth Boyle
Sociology & Law, University of Minnesota, USA
Principal Investigator, IPUMS-DHS
Previous Topic: IDHSSTRATA & IDHSPSU: Inconsistent Lengths
Next Topic: Merging DHS with IPUMS-DHS to get harmonized districts
Goto Forum:

Current Time: Sat Jul 4 06:13:50 Eastern Daylight Time 2020