| Re: number of women with a live birth in the 2 years preceding the survey [message #28376 is a reply to message #28305] |
Fri, 22 December 2023 09:15   |
Janet-DHS
Messages: 888
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
It will be easier if you use the NR file (a new type of file for DHS-8 surveys, with all the events in the pregnancy history for the past 5 years), rather than the IR file. Open NPNR82FL.dta and enter the following Stata line for the indicator in table 9.9:
*Match table 9.9.
tab p32 if p19<=23 & p32<=2 [iweight=v005/1000000]
This will match the total numbers of live births and stillbirths in table 9.9.
Table 9.1 is earlier in the tab plan but is actually a slight subsample of table 9.9. It is restricted to the most recent live birth or stillbirth in the past two years. You need to construct a recode of pidx, limited to those outcomes, and select the youngest:
* Match table 9.1
egen pidxr=rank(pidx) if p19<=23 & p32<=2, by(v001 v002 v003)
tab p32 if p19<=23 & p32<=2 & pidxr==1 [iweight=v005/1000000]
Looking at the report, I do see a slight problem for table 9.1. The table gives 1933 births and 18 stillbirths (weighted), and a total of 1949. There is obviously some rounding here. I'd say there is also some kind of an error, because it is impossible for two numbers that round to 1933 and 18, respectively, to have a total that rounds to 1949. When I run the lines above, I get 1932.37 births and 16.64 stillbirths, before rounding, and a total of 1949.01. These numbers round to 1932 and 17 and 1949, respectively. The total of 1949 is matched exactly. I would not worry about reconciling 1933 and 18 in the report with 1932 and 17, respectively.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
