The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Nutrition and Anthropometry » Stunting indicators in KR vs. PR file
Stunting indicators in KR vs. PR file [message #13405] Tue, 31 October 2017 23:22 Go to next message
dflood011 is currently offline  dflood011
Messages: 7
Registered: October 2017
Member
Hi DHS team,

I am using the Guatemala 2015 DHS dataset to analyze child stunting risk factors.

Using the Household (PR) file, I am able to replicate the stunting prevalence rates from the DHS report. However, I am wondering why I am getting such dramatically different numbers when I use the Children's file (KR)

My understanding is that the children in the KR file are a subset of children in the PR file. So it would make sense to have slight differences in the numbers between the two files.

Unweighted PR analysis: 6,607 stunted children out of 12,258 total children (53.9%)
Unweighted KR analysis: 8,170 stunted children out of 11,444 total children (71.4%)

But why would the percentages be so vastly different? Any why would there be MORE stunted children in the KR file if this is a subset of PR?

I have a feeling I am missing something obvious. Any help or guidance would be appreciated.
Re: Stunting indicators in KR vs. PR file [message #13498 is a reply to message #13405] Fri, 10 November 2017 10:42 Go to previous messageGo to next message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
Dear User,
Here is a description of the two types od data files you reference:
Quote:

Household Listing Data - Household Member Recode (PR)
This dataset has one record for every household member. It includes variables like sex, age, education, orphanhood,
height and weight measurement, hemoglobin, etc. It also includes the characteristics of the households where the individual
lives or was visiting. The unit of analysis (case) in this file is the household member.


Quote:

Children's Data - Children's Recode (KR)
This dataset has one record for every child of interviewed women, born in the five years preceding the survey. It
contains the information related to the child's pregnancy and postnatal care and immunization and health. The data for the
mother of each of these children is included. This file is used to look at child health indicators such as immunization
coverage, vitamin A supplementation, and recent occurrences of diarrhea, fever, and cough for young children and treatment
of childhood diseases. The unit of analysis (case) in this file is the children of women born in the last 5 years (0-59 months).

Re: Stunting indicators in KR vs. PR file [message #13502 is a reply to message #13498] Fri, 10 November 2017 11:16 Go to previous messageGo to next message
dflood011 is currently offline  dflood011
Messages: 7
Registered: October 2017
Member
Hi Liz, thanks for your message. I understand the differences between the two types of data files.

My basic questions:

(1) If children in the KR file are subset of children in the PR file, why would there be more stunted children in the KR file for Guatemala DHS 2014?

(2) Even accounting for the subtle differences in data sets, why would stunting rates be so much higher (~20%) in the KR file?

My assumption about the subset comes from these 2 threads which, were previously answered by Trevor and Tom:
https://userforum.dhsprogram.com/index.php?t=msg&goto=88 52&S=Google
https://userforum.dhsprogram.com/index.php?t=msg&goto=26 1&
Re: Stunting indicators in KR vs. PR file [message #13504 is a reply to message #13502] Fri, 10 November 2017 11:54 Go to previous messageGo to next message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
Dear User,
Thank you for posting again. I have forwarded your query to one of our experts. As soon as we have a response, we will post.
Re: Stunting indicators in KR vs. PR file [message #13517 is a reply to message #13504] Mon, 13 November 2017 12:50 Go to previous messageGo to next message
dflood011 is currently offline  dflood011
Messages: 7
Registered: October 2017
Member
Thank you very much! I know I'm missing something fundamental about the datasets, so any help would be greatly appreciated.
Re: Stunting indicators in KR vs. PR file [message #13534 is a reply to message #13517] Wed, 15 November 2017 12:06 Go to previous messageGo to next message
dflood011 is currently offline  dflood011
Messages: 7
Registered: October 2017
Member
Hi Liz, I realized the mistake I was making. I had mixed up maternal and child anthropometrics. Sorry for bringing this question to your attention.
Re: Stunting indicators in KR vs. PR file [message #24096 is a reply to message #13534] Wed, 23 February 2022 17:18 Go to previous messageGo to next message
AkhilK28 is currently offline  AkhilK28
Messages: 8
Registered: February 2022
Member
Hello,

I have the same question but for child stunting. In the KR file, there are 86,312 stunted children in India, but in the PR file, there are 80,020 in India in NFHS-4 (2015-16), according to DHS.

Why are there more stunted children in the KR file if the KR file is supposed to be a SUBET of the PR file!

Thanks,
Re: Stunting indicators in KR vs. PR file [message #24101 is a reply to message #24096] Thu, 24 February 2022 07:25 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 2588
Registered: February 2013
Senior Member

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

Measurements of height and weight are taken during the household interview, for all children under 5 in the household and (usually) all women 15-49. That information goes into the PR file. Some households include children whose mother is not alive, or is alive but not living in the same household as the child. Then the eligible women are interviewed, and the data for the KR file is obtained. Height and weight are copied (during file construction) from the PR data to the KR data. The copying is limited to children who are living with their mothers.

Thus the calculation of stunting based on the PR file is representative of children in the household population, regardless of whether they are living with their mother. When based on the KR file, it is representative of children in the household population whose mother is alive and living in the same household as the child.
Re: Stunting indicators in KR vs. PR file [message #25016 is a reply to message #24101] Fri, 19 August 2022 19:40 Go to previous messageGo to next message
AkhilK28 is currently offline  AkhilK28
Messages: 8
Registered: February 2022
Member
Then shouldn't it mean that the KR stunted children population should be lower than the PR stunted population? If it was copied from the PR --> KR, then why does KR have a higher stunted N of children? PR includes all children living with mother or otherwise whereas KR only contains living with mother so therefore KR should be less than PR but in actuality KR has a higher N of stunted children in NFHS-4!

[Updated on: Fri, 19 August 2022 19:41]

Report message to a moderator

Re: Stunting indicators in KR vs. PR file [message #25036 is a reply to message #25016] Mon, 22 August 2022 09:27 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 2588
Registered: February 2013
Senior Member
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

Stunting is based on the HAZ, the height-for-age Z score. I checked how many children have a HAZ score in the KR and PR files.

/index.php?t=getfile&id=1916&private=0


That is, there are 211,089 measured children in the KR file and 213,860 in the PR file. The N is smaller in the KR file than in the PR file, as you would expect. I don't see how you could have gotten the reverse.

Note that the KR file includes children who died, and children who are not living with the mother and for that reason there are children in the KR file who are not in the PR file. The Z scores--and stunting, etc.--are only obtained for children who are alive and in the household and whose height and weight were measured. There are more children with Z scores in the PR file than in the KR file because the height and weight measurements are part of the household survey and there are a few children in the household survey whose mother had died or is living elsewhere.
  • Attachment: hw70.jpg
    (Size: 46.18KB, Downloaded 68 times)
Previous Topic: height-for-age difference (HAD) for children 0-5 years
Next Topic: Staple food by type in DHS nutrition surveys
Goto Forum:
  


Current Time: Fri Sep 30 09:49:48 Coordinated Universal Time 2022