Home » Topics » Fertility » Table on Fertility Planning Status from NFHS-4 data
Table on Fertility Planning Status from NFHS-4 data [message #14023] |
Sun, 04 February 2018 01:53 |
preejayan
Messages: 1 Registered: February 2018
|
Member |
|
|
Dea DHS Team-
I am using NFHS-4 data to recreate Table 4.18 on Fertility Planning Status of India Report (2015-16). I used the following set of variables in different permutation and combination but couldn't create this table exactly what was published.
I used birth history data (IABR71FL.DTA) and variables
v005 (weight), v208, v213, v225, b10, v013 and bord.
I ended up with 275,648 cases however, India (NFHS-4) report states 281,067 cases.
Secondly, I used both types of weights, i.e., "aweight" and "iweight" while generating this table. No of cases in "aweight" is more than the number of cases if we use "iweight". This is opposite when we use previous rounds of NFHS datasets (like NFHS 1, 2, & 3) where number of cases are more when use "iweight".
Please help me - i) in providing the stata code for Table 4.18 and ii) explain why number of cases are more while using "aweight" in NFHS-4 compared to number of cases in "iweight" calculations, which is exactly opposite with other NFHS previous rounds.
Thanks for your help!
- Jayachandran
|
|
|
|
Re: Table on Fertility Planning Status from NFHS-4 data [message #14153 is a reply to message #14065] |
Mon, 26 February 2018 13:43 |
Bridgette-DHS
Messages: 3203 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
I apologize for the delay in this response to your Feb. 4 posting. The table on "fertility planning status"--table 4.18 in the India 2015-16 report--is a relatively difficult table to reproduce in Stata. I am attaching a Stata program, as a text file, that reproduces table 4.18, and the accompanying log file. To use the program you must change the paths. The structure of the program, with subprograms, may seem too complex, but it is actually much easier to segment the procedure as I have done, because the IA71 files are huge and you waste time just waiting for them to open.
Everything can be done from the IR file. You first break out the births in the past five years, using "reshape long" and save them. Then you return to the IR file an break out the pregnancies. You then append the pregnancies to the births. There are actually two problems with table 4.18 as it appears in the report, both of them with the "number of births" column for "mother's age at birth". The problem is for the cases that are current pregnancies, rather than actual births. The mother's age at birth, for current pregnancies, is calculated by assuming that the cmc of the birth will be the cmc of interview plus 9 minus the number of months pregnant. The number of months pregnant is given by v214. The distribution of v214 has two problems. One is that there are 7 cases in which the woman says she is pregnant but the duration of pregnancy is not given. The numbers given do not actually add up to 281,067 because of those 7 cases. More important, there are 75 women who give the duration of their current pregnancy as 10 months. The calculation that I just gave, in those cases, would mean that the child with whom the woman is pregnant was already born, the month before the interview! It would have been better to recode duration=10 months to 9 months. Of course, there are actually many errors in the reporting of v214, and the effect of those 75 cases is negligible. In the program I indicate where this particular problem could be "corrected".
[Updated on: Mon, 26 February 2018 13:44] Report message to a moderator
|
|
|
Re: Table on Fertility Planning Status from NFHS-4 data [message #14392 is a reply to message #14153] |
Mon, 02 April 2018 10:02 |
Bridgette-DHS
Messages: 3203 Registered: February 2013
|
Senior Member |
|
|
The following is from Senior DHS Stata Specialist, Tom Pullum:
This is an update to my Feb. 26 posting on the calculation of "fertility planning status". The Stata program attached (as a text tile) will calculate the indicator to match the table in the report. The program is set up for Nepal rather than India but if you just change the file name (and the path) you should get a match.
[Updated on: Mon, 02 April 2018 10:03] Report message to a moderator
|
|
|
|
|
Goto Forum:
Current Time: Sun Dec 8 11:08:47 Coordinated Universal Time 2024
|