| Re: Merging datasets in SPSS [message #14703 is a reply to message #14684] |
Tue, 01 May 2018 07:34   |
Bridgette-DHS
Messages: 3214
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
As has been stated in other postings, the PR file includes all children living in the household. It gives the most representative information about children in the household population. The KR file is limited to children whose mother is living and in the household. It does not include children whose mother has died or is living in a different household than the child. The children who are in the PR file but not the KR file tend to be relatively deprived and are more likely to be stunted, etc. However, if you want to relate stunting (etc.) to other characteristics of the child, or characteristics of the mother, it is necessary to use the KR file.
The variable hw1 (in the KR file) or hc1 (in the PR file) is age in months, calculated as month of interview minus month of birth (v008-b3). The new variable b19 is similar but is calculated down to the level of the day. The day of the interview is given as v008a (in century day codes, or cdc). The day of birth is the new b18. The length of a month is approximated as a constant number of days, 365.25/12=30.4375. Then b19 is the integer part of v008a-b18 when divided by 30.4375. Previously, children would be included in the KR file if hw1 was between 0 and 59, inclusive. Currently, the requirement is that b19 is between 0 and 59, inclusive.
If you have a question about an indicator such as food security, please refer to a specific table in which the indicator appears.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
