Following is a response from DHS staff member, Tom Pullum:
I only use Stata and will respond with that, but you should be able to translate to SPSS. I think the problem is that when you reshape the data you are keeping the full allowance of up to 20 siblings. You need to delete the lines of data that go beyond the number of siblings that a woman actually had. Also you need to restrict to women who died at age 15-49.
I entered the following lines:
use "...NGIR7BFL.DTA", clear
* remove leading zeroes in the subscripts
rename *_0* *_*
* For this example just keep mm1 and mm7 and mm9
keep v001 v002 v003 mm1_* mm9_*
reshape long mm1_ mm9_, i(v001 v002 v003) j(index)
rename *_ *
tab mm1, m
* Drop the lines that go beyond the woman's actual number of siblings
drop if mm1==.
tab mm9 if mm1==2 & mm7>=15 & mm7<=49,m
I see 616,859 blank lines in the data, before dropping the lines with mm1=1:
. tab mm1, m
mm1 | Freq. Percent Cum.
------------+-----------------------------------
male | 112,971 13.51 13.51
female | 106,590 12.74 26.25
. | 616,859 73.75 100.00
------------+-----------------------------------
Total | 836,420 100.00
The last line in the program produces this:
. tab mm9 if mm1==2 & mm7>=15 & mm7<=49,m
mm9 | Freq. Percent Cum.
------------------------+-----------------------------------
death not related | 2,484 67.52 67.52
died while pregnant | 415 11.28 78.80
died during delivery | 535 14.54 93.34
6 weeks after delivery | 202 5.49 98.83
2 months after delivery | 43 1.17 100.00
------------------------+-----------------------------------
Total | 3,679 100.00
There are no "missing" cases if you restrict to women age 15-49 at death. These are the only siblings to whom mm9 applies.
Note that there is an updated version of the files, with "C" in place of "B".
Please let us know if you are still seeing blanks or NA cases
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
