The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » General Data Questions » Data Query
Data Query [message #10125] Wed, 29 June 2016 11:18 Go to next message
DHS user is currently offline  DHS user
Messages: 111
Registered: February 2013
Senior Member
I am using Bangladesh and Nepal DHS surveys for one of my projects. Both the surveys refer to the year 2011. For my analysis, I am using the Children Recode (KR) files for both countries. I will really appreciate your help in clarifying few of my doubts. I am rushing against a deadline to submit this paper to a journal and really appreciate your help.

1. In the KR file for both countries, the weighted sample size for WAZ scores is 7683 children (for Bangladesh) and 2380 children (for Nepal). Whereas the DHS final reports for both countries report that the weighted sample size on WAZ scores is 7861 (Bangladesh) and 2475 (Nepal).
a. Is the difference arising because the KR file only report sample from the interviewed women? And not the de facto children?
b. Further, for the children for whom the DHS have the information on WAZ scores but their mothers were not interviewed, does the DHS data set provide information on their socioeconomic, maternal and demographic characteristics?
2. As the DHS report shows, the Bangladesh DHS resulted from completed interviews of 17,842 eligible women and from 17,141 selected households. Similarly, the Nepal survey resulted from successful interviews of 12,674 eligible women and from 10,826 households. I am trying to create a flow chart for each country to show that how do I arrive at my final sample of WAZ scores. For example for Bangladesh, I want to show how interviews of 17,842 eligible women and from 17,141 selected households results in the sample (weighted) of 7683 children for WAZ scores.
a. This is what I have understood. In Bangladesh, every child in 0-59 months age group in 17,141 selected households was measured for height and weight. For the children who have valid measurements, the sample size is 7861 children. Once we remove the children (de facto) whose mothers were not interviewed the sample size is 7683. The same is applicable for Nepal. Please let me know if my understanding is wrong and DHS has used some other criteria.
Re: Data Query [message #10127 is a reply to message #10125] Wed, 29 June 2016 11:19 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3013
Registered: February 2013
Senior Member
Following is a response from Senior DHS Stata Specialist, Tom Pullum:

You are correct. Height and weight are obtained within the household interview for children under five. The KR file (also the BR file) is limited to the children of women who are de facto members of the household. For some of the children in the household file, the mother has died or is not in the household (survival and co-residence for the mother are given in hv111 and hv112).

As much as possible, DHS uses the data in the household file (hc70-72) for the z scores, but when linked to data for the mother in the child files, we are limited to hw70-72, with somewhat fewer children. It appears to me that everything you said is correct -- that is, it's consistent with the structure of the data.
Previous Topic: Marital Instability
Next Topic: Query about variable
Goto Forum:
  


Current Time: Tue Mar 19 05:48:11 Coordinated Universal Time 2024