| Re: Creating Couples' file from IR and MR datasets for India [message #25162 is a reply to message #25157] |
Thu, 08 September 2022 07:52  |
Bridgette-DHS
Messages: 3199
Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:
If you need to add variables to the CR file, it is best to just do an IR/CR merge for any new women's variables (sort both files by v024 v001 v002 v003) and an MR/CR merge for any new men's variables (sort both files by mv024 mv001 mv002 mv003). If you try to reconstruct the CR file from scratch it is likely that you will have problems. This is a more complicated merge than it may seem to be. In the standard CR file, women are only allowed to name one man with v034, but men can name multiple women with mv034_*. The same man may appear in multiple couples. Otherwise, there would be women who would not appear at all. How you handle polygamy is usually the reason for the kinds of discrepancies you describe. I think it would be easier just to do IR/CR and MR/CR merges, as I said. The first one would be 1:1 and the second one would be 1:m (but you can do such merges without specifying "1:1" etc. so long as you enter "keep if _merge==3").
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
