The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » India » NFHS-4 Median Birth interval mismatch
NFHS-4 Median Birth interval mismatch [message #18002] Tue, 13 August 2019 13:03 Go to next message
dfnbcc is currently offline  dfnbcc
Messages: 3
Registered: June 2019
Member
Hello,

I am trying to calculate the median birth interval using STATA from the NFHS-4. However, I am getting the wrong median compared to the report. The report says the median in India is 32 months, however, I am getting 28 months.

I am subtracting 9 months from the variable as I want birth-to-pregnancy interval, not birth-to-birth. But even if I don't subtract 9, I get different values.

use IABR74FL.dta
gen weight=v005/1000000

rename sdistri shdistri
tostring  shdistri, replace 
gen strata = shdistri
svyset v021 [pweight=weight], strata (strata)

keep shdistri v008 b3 bord b0 b11 v106 hv024 v025 v190 v447a v133 s301 hv001 hv002 hvidx v005 v021 weight strata


*Births in the month of interview are included and births 60 months before the interview are excluded.
	gen monthssncbirth=v008-b3
	replace monthssncbirth=. if monthssncbirth<=60
	drop if monthssncbirth==.
	
*First-order births (and their twins) are excluded from both numerators and denominators.	
	svy: tab bord
	gen birth_order=bord if (b0 > 1) 
	svy: tab birth_order

*subtract by 9 months to find birth-to-pregnancy interval (WHO)
	gen birth_interval=b11-9 
	replace birth_interval=. if birth_interval<0


Members Search Help Register Login Home Home » Topics » Fertility » NFHS-4 Median Birth interval (Median birth interval does not match NFHS-4 report)
Show: Today's Messages :: Show Polls :: Message Navigator
Switch to threaded view of this topic Create a new topic Submit Reply
- NFHS-4 Median Birth interval [message #17992] Sun, 11 August 2019 11:28
dfnbcc is currently offline dfnbcc
Messages: 2
Registered: June 2019
Member
Hello,

I am trying to calculate the median birth interval using STATA from the NFHS-4. However, I am getting the wrong median compared to the report. The report says the median in India is 32 months, however, I am getting 28 months.

I am subtracting 9 months from the variable as I want birth-to-pregnancy interval, not birth-to-birth. But even if I don't subtract 9, I get different values.


Code: [Select all] [Show/ hide]
use IABR74FL.dta
gen weight=v005/1000000

rename sdistri shdistri
tostring shdistri, replace
gen strata = shdistri
svyset v021 [pweight=weight], strata (strata)

keep shdistri v008 b3 bord b0 b11 v106 hv024 v025 v190 v447a v133 s301 hv001 hv002 hvidx v005 v021 weight strata


*Births in the month of interview are included and births 60 months before the interview are excluded.
gen monthssncbirth=v008-b3
replace monthssncbirth=. if monthssncbirth<=60
drop if monthssncbirth==.

*First-order births (and their twins) are excluded from both numerators and denominators.
svy: tab bord
gen birth_order=bord if (b0 > 1)
svy: tab birth_order

*subtract by 9 months to find birth-to-pregnancy interval (WHO)
gen birth_interval=b11-9
replace birth_interval=. if birth_interval<0

Would appreciate some help please.

Thanks in advance.
Re: NFHS-4 Median Birth interval mismatch [message #18041 is a reply to message #18002] Sat, 24 August 2019 06:37 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 1748
Registered: February 2013
Senior Member
DHS staff are very busy right now, and we will try to answer in a couple of weeks?
Re: NFHS-4 Median Birth interval mismatch [message #18152 is a reply to message #18041] Mon, 30 September 2019 06:19 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 1748
Registered: February 2013
Senior Member
Following is a response from DHS Research and Data Analysis Director, Tom Pullum:

Instead of the BR file, use the KR file to get births in the past five years. Look at the distribution of b11. There are some very short intervals and I would edit them out. Then enter the following line in Stata:

summarize b11 [fweight=v005] if b11>9, detail

The 50th percentile is 32 months. That matches the report.

To get a birth-to-pregnancy interval you could subtract 9 months, but that would be the interval to pregnancies that resulted in a live birth, and presumably you would want to include all pregnancies, not just those that resulted in a live birth. For that you would need to use the calendar, and probably include current pregnancies. That's more complicated, partly because of left censoring.
Previous Topic: NFHS 4 and women's work
Next Topic: School attendance for 6-17 years
Goto Forum:
  


Current Time: Sun Feb 23 11:29:03 Eastern Standard Time 2020