The DHS Program User Forum

Today's Messages (off) | Unanswered Messages (on)

Topic: Combining ADM1 regions in a single survey (KR, Children's recode)

Combining ADM1 regions in a single survey (KR, Children's recode) [message #30946]

Wed, 30 April 2025 12:18

deepitbhatia
Messages: 1
Registered: April 2025

Member

I'm looking to create generate standard errors for the proportion vaccinated against measles from the children's recode (KR) files for a number of countries across DHS 5, 6, and 7.

So far, I've used the instructions in the forum post linked here:
https://userforum.dhsprogram.com/index.php?t=msg&goto=28 772&&srch=%22standard+error%22#msg_28772
to get region-level (i.e., ADM1-level) estimates of the standard error by running intercept-only logistic models for each region and extracting the standard error of the intercept.

My question is: if I want to pool certain regions within a single survey (so that I can compare regions later on between surveys for a given country, in case of changes in the ADM1 boundaries), do I need to do anything different with the weights or sampling strategy other than:
1) subset my survey design for all regions I'm interested in combining (instead of subsetting it by just one region)
2) subset my data frame for all regions I'm interested in combining (instead of subsetting it by just one region)

Thanks,
Deepit

Report message to a moderator

Current Time: Mon May 5 16:41:52 Coordinated Universal Time 2025