The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Dataset access » Records missing in child survey for specific clusters
Records missing in child survey for specific clusters [message #8711] Fri, 04 December 2015 10:29 Go to next message
karoun11 is currently offline  karoun11
Messages: 4
Registered: December 2015
Location: Gainesville, FL
I am working with the Pakistan 2006-2007 DHS child survey. I have noticed this before- but some clusters have data reported in the household survey (indicating that a cluster has children in it) but do not have data in the child survey. For example, cluster 129 is missing from the PKKR52FL file, but is present in the pkhr52dt file. I was joining the data to the cluster GPS points, and am concerned that I am making an error or misunderstanding something. I have noticed this in other countries datasets as well.


Karoun H. Bagamian
Postdoctoral Associate, HEDGE lab
Emerging Pathogens Institute/Department of Environmental and Global Health
University of Florida
Re: Records missing in child survey for specific clusters [message #8740 is a reply to message #8711] Fri, 11 December 2015 07:35 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 2280
Registered: February 2013
Senior Member
Following is a response from Senior DHS Stata Specialist, Tom Pullum:

I looked at cluster 129 in that survey. It includes 319 people in the PR file. Only 11 are under 5 years of age. One of them has hv103=0, i.e., is not a de facto resident. Then I looked at hv101, relation to the household head. 2 are not related to the household head, 2 are a child of the household head, and 7 are a grandchild.

A child under 5 will get into the KR file if the child's mother is a de facto resident of the household. It's my guess that none of these children has a mother living in the household. Unfortunately, in this survey, the variables on parental survival and co-residence (hv111-hv114) are not included. If they were, we would probably see that hv112 is . or 0 for all of these children. (. means not living and 0 means living but not in the household.)

This is obviously a pretty unusual cluster. However, this survey had 972 clusters, which is a lot of clusters.

In general if there are children under 5 in the PR file for a cluster, but no children under 5 in the KR file for that cluster, I would expect it to be because hv112 is . or 0 for all the children (under 5) in the PR file. I don't think there's a way to confirm that explanation for surveys such as this one that did not include hv112, but it's probably true.

There also can be children in the KR file who are not in the PR file--the reverse kind of discrepancy. These will be the children of those eligible women who had children in the past five years but those children have since died or they live in another household.
Previous Topic: Restructuring Healthcare data
Next Topic: download multiple surveys
Goto Forum:

Current Time: Tue May 18 10:18:13 Coordinated Universal Time 2021