The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Bangladesh » BDHS2011 - Tab. 5.4 Children ever born and living
BDHS2011 - Tab. 5.4 Children ever born and living Thu, 21 May 2015 10:26
 geoK Messages: 30Registered: May 2014 Member
Hello,
I am trying to replicate percentages and mean number of children ever born as in Tab. 5.4 for ALL WOMEN but I get different numbers. I can get the total number of women 20,797 using the ALL WOMEN FACTOR but all % by age and total come different. I am sure it is something to do with the all women factor which I am struggling to understand how it is used...

Can anyone help, please? Maybe suggesting a script or explaining step by step? Thanks.

Also, on the same topic (ALL WF), I'd like to further disaggregate mean number of children ever born for all women by e.g. Region. I am aware that the specific factor for regional disaggregation is created, but just wondering how to use it in the case of mean number of children ever born.

Thanks again,
Best regards!
Re: BDHS2011 - Tab. 5.4 Children ever born and living [message #4418 is a reply to message #4405] Sat, 23 May 2015 00:45
 Trevor-DHS Messages: 680Registered: January 2013 Senior Member
The issue with using the all women factors is often that you need to inflate the denominator of your indicator to inflate the number of women from the ever married women to all women, but you don't want to inflate the numerator as the assumption is that the never married women have had no children. Below is some code in stata that demonstrates this:
```* Set up svy paramaters
gen wt=v005/1000000
svyset v021 [pw=wt], strata(v023)

* All women factor
gen aw = awfactt/100
* Check the denominator
tab v013 [iw=aw*wt]

* Children ever born
gen ceb = v201
* Use the ratio of children ever born and all women
svy: ratio ceb/aw, over(v013)
svy: ratio ceb/aw```
Re: BDHS2011 - Tab. 5.4 Children ever born and living [message #4429 is a reply to message #4418] Tue, 26 May 2015 07:32
 geoK Messages: 30Registered: May 2014 Member
Dear Trevor, thanks a lot, this is really helpful. Can I please double check with you if I have understood it correctly, when calculating other indicators with different subpopulations:
For example,
If I want to calculate women's literacy (taking variable V155 and creating an indicator variable where LITERATE=1 and NOT-LITERATE=0, following tab. 3.3.1 definition of literate women) by Religion (V130):
1- should I create a new women factor specific for variable Religion (V130)? Given that the dataset lists 5 factors which look sub-pop specific: awfactt; awfactu; awfactr; awfacte; awfactw
2- assuming that the correct factor to use is called awfactt, and following your script, would the following be correct? (IN RED MY CHANGES)

* Set up svy paramaters
gen wt=v005/1000000
svyset v021 [pw=wt], strata(v023)

* All women factor
gen aw = awfactt/100
* Check the denominator
tab v130 [iw=aw*wt]

* Children ever born
gen Literacyw = (derived from V155 where 1=literate; 0=illiterate)
* Use the ratio of children ever born and all women
svy: ratio Literacyw/aw, over(v155)
svy: ratio Literacyw/aw

#################################################

Thanks a lot,
regards

• Attachment: RESULTS.pdf
Re: BDHS2011 - Tab. 5.4 Children ever born and living [message #5482 is a reply to message #4429] Fri, 29 May 2015 05:14
 geoK Messages: 30Registered: May 2014 Member
Hi again. Any idea/suggestion about my last question?

Sorry, In my program there is a mistake, it should be

svy: ratio Literacyw/aw, over(v130)
svy: ratio Literacyw/aw

It is however correct in the pdf attached called results.
thanks a lot
regards

[Updated on: Sun, 07 June 2015 13:55]

Report message to a moderator

Re: BDHS2011 - Tab. 5.4 Children ever born and living [message #5579 is a reply to message #4429] Wed, 10 June 2015 10:58
 Trevor-DHS Messages: 680Registered: January 2013 Senior Member
1) Ideally you would create all women factors for religion, but given that two of the religious groups are tiny it is impractical to do so, so I would just use the total all women factor AWFACTT.
2) Your corrected code looks fine.

In your results, though, you should ignore the results for Christianity and Buddhism as you have far too few cases and the results are not representative at all.
Re: BDHS2011 - Tab. 5.4 Children ever born and living [message #5609 is a reply to message #5579] Tue, 16 June 2015 06:19
 geoK Messages: 30Registered: May 2014 Member
Thanks a lot!
Re: BDHS2011 - Tab. 5.4 Children ever born and living [message #5625 is a reply to message #5609] Wed, 17 June 2015 15:49
 Trevor-DHS Messages: 680Registered: January 2013 Senior Member
I just realized that my last response was misleading. My comments 1) and 2) are not wrong, but the whole idea of applying all women factors for literacy is problematic. We use all women factors with with variables for fertility, marital status and sexual activity because we can assume (in the countries that use all women factors) that all never married women have had no pregnancies, are never married, and have never had sex. We can't make an assumption about their literacy, and thus we can't use this approach to estimate literacy of all women. My apologies that I didn't spot this before.
Re: BDHS2011 - Tab. 5.4 Children ever born and living [message #5631 is a reply to message #5625] Thu, 18 June 2015 05:03
 geoK Messages: 30Registered: May 2014 Member
Thanks a lot!!!!!!
 Previous Topic: All men factor - Literacy rate Next Topic: No of pregnant mother and U5 child
Goto Forum:

Current Time: Sat Mar 28 19:02:01 Eastern Daylight Time 2020