| Re: Height and Weight Scores (WHO Child Growth Standards) data [message #19200 is a reply to message #19150] |
Fri, 08 May 2020 14:53  |
Bridgette-DHS
Messages: 3215
Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:
I always use the old version of merge, without "1:1", etc, because it tends to give me problems. After your first merge, do you have a new variable called "_merge"? If so, you need to drop it.
The main problem is that the hc7x variables are only applicable to children under 5 who are in the household survey. You should use the KR file rather than the BR file, because the BR file includes ALL births in the birth histories, not just those in the past five years. Next, both the BR and KR files include children who have died or are not in the same household as the mother. You need to eliminate those cases. You can have multiple children with the same mother who have died or are living elsewhere and that's why you get the "not unique observations" message. Let us know if the following does not work.
I suggest that you revise this
*merge to birth data
use "ETBR51FL.DTA", clear
sort v001 v002 b16
merge 1:1 v001 v002 b16 using "HW_to_merge.dta"
to become this:
*merge to birth data
use "ETKR51FL.DTA", clear
drop if b16==0 | b16==,
sort v001 v002 b16
merge 1:1 v001 v002 b16 using "HW_to_merge.dta"
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
