Re: Merging BR and PR files - India - NFHS 1998 [message #26045 is a reply to message #26041] |
Wed, 01 February 2023 11:46 |
Bridgette-DHS
Messages: 3199 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS staff member, Tom Pullum:
We at DHS have put a lot of effort into procedures to match children in the KR or BR file with children in the PR file in the surveys that omitted b16. We have concluded that this is not possible without a lot of incorrect matches. You know that the child must be alive and living with the mother, which is captured with b9=0. Beyond that, all you have to work with is the child's sex and age in the PR file (hv104 and hv105 and the child's sex and age in the KR/BR file (b4 and b8). You can probably safely assume that the sex will be correct in both files. However, age is not enough to go on. For a specific child that DOES appear in both places, the two ages may not agree. Also you can have multiple children at the same age (for example, twins). There can be 2+ mothers (with their children) in the same household, but in this survey you do not even have hv112 to distinguish between 2+ possible mothers. Any procedure that does not use b16 will result in some spurious matches. I recommend that you use one of the more recent India surveys that does include b16.
|
|
|