| Unequal observations per gps cluster point [message #27856] |
Thu, 12 October 2023 17:57  |
pcmitchell
Messages: 1
Registered: October 2023
|
Member |
|
|
Hello,
This is my first foray using DHS data. I am exploring nutritional data in Malawi and am specifically interested in childhood stunting, wasting, and underweight by cluster location. However, I have found that there are a different number of observations per cluster--for example, some clusters have as many as 17 survey entries and other clusters just 1. How should i go about excluding clusters and aggregating data by cluster in order to show some sort of concentration of stunting, etc. by cluster or district (by district would require aggregating clusters to a district polygon)?
Thank you,
Penelope
|
|
|
|
| Re: Unequal observations per gps cluster point [message #27879 is a reply to message #27856] |
Mon, 16 October 2023 16:25  |
Janet-DHS
Messages: 891
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
In the typical DHS sample design, about 30 households are selected from each cluster (or census enumeration area). However, there can be a lot of variation in the number of specific types of cases, such as children under 5. The variation from 1 to 17 children that you describe is random. Your analysis should include the number of cases per cluster, as well as the proportion stunted (etc.) and the estimation procedure should give less weight to a proportion with a small n than to a proportion with a large n. I don't think you need to do any exclusion or aggregation. Let us know if this is not clear, and we will ask the GIS team.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
