The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Other countries » Missing Data in Indonesia DHS 2012
Missing Data in Indonesia DHS 2012 [message #9813] Tue, 24 May 2016 16:01 Go to next message
DHS user is currently offline  DHS user
Messages: 102
Registered: February 2013
Senior Member
I am using the women dataset from Indonesia DHS. I have transformed the data set from wide to long format to see the data in sibling's record. However, I got stuck with the data. I found a missing data that I don't know why it's missing. From the new dataset, there are 195,839 sibling records. There are 94,715 records of female siblings after excluding the male siblings. Of these, there are 11,975 women who died and 82,740 are alive.


When I tried to count the women who died from maternal causes, there are 9,142 missing which gets me confused. Why are they missing? Is it because these 9.142 records are the unknown deaths? And, if I exclude the missing data and deaths not related to pregnancy, to describe maternal mortality number, it means that there are 372 deaths related pregnancy, isn't it?
  • Attachment: Indonesia.jpg
    (Size: 21.63KB, Downloaded 377 times)
Re: Missing Data in Indonesia DHS 2012 [message #9814 is a reply to message #9813] Tue, 24 May 2016 16:04 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 2307
Registered: February 2013
Senior Member
Following is a response from Senior DHS Stata Specialist, Tom Pullum:

In situations like this it is good to look at the questionnaire. In the woman's questionnaire for this survey, given in an appendix (see page 451 of the report), #1109 is a filter based on sex and age at death. The questions about whether the death was pregnancy-related are not asked for men, of course, but they also are not asked about girls who died before age 10. The last two lines below will show you that the relevant number of missing cases is more like 447 than 9142. I would prefer that those 447 cases were not missing, but that's only about 5% of the number you thought. There may be some other reason why they are missing, such as a reluctance by the interviewer to ask whether a sister who died in her early teens had been pregnant. I agree that 372 deaths were identified to be pregnancy-related--in terms of their timing, at least.

set more off
use e:\DHS\DHS_data\IR_files\IDIR63FL.dta, clear

keep v001 v002 v003 mm*
describe mm*_1
drop mm5* mm6* mm10* mm11* mm12* mm13* mmc* mmidx*
rename *_0* *_*
describe mm*_1

reshape long mm1_ mm2_ mm3_ mm4_ mm7_ mm8_ mm9_ mm14_ mm15_,i(v001 v002 v003) j(mmidx)
rename *_ *
drop if mm1==.

drop if mm1==1
tab mm9 mm2,m
tab mm7 if mm2==0 & mm9==., m
tab mm7 if mm2==0 & mm9==. & mm7>=10 & mm7<=49, m
Previous Topic: The Meaning of Camp Rur (4) in v134 of HTIR12
Next Topic: Zambia 2013-14 DHS
Goto Forum:

Current Time: Sat Jun 19 04:48:31 Coordinated Universal Time 2021