The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Fertility » Ideal fertility sons and daughters v613 v627 v628
Ideal fertility sons and daughters v613 v627 v628 Tue, 10 April 2018 16:27
 rkbroussard@prc.utexas.ed Messages: 4Registered: April 2018 Location: Austin, Texas Member
Hello, I'm currently using a compilation of individual Peru data from 2003-2011. There seems to be a problem with the raw variables v613, v627, v628, and v629. According to the DHS recode manual, v613 (ideal number of children) is supposed to be the sum of v627(ideal number of boys)+ v628(ideal number of girls) +v629 (ideal number of any sex), but the numbers are way off.

For example, 53,526 women said they their ideal number of boys would be 0 in variable v627, but 3891 respondents said their ideal number of children would be 0 in v613.

Any help would be much appreciated.

Kathleen Broussard
Re: Ideal fertility sons and daughters v613 v627 v628 [message #14664 is a reply to message #14472] Thu, 26 April 2018 12:39
 Liz-DHS Messages: 1516Registered: February 2013 Senior Member
Dear User, A response from Dr. Shea Rutstein:
Quote:

As I interpret this question, the researcher has combined all the data between 2003 and 2011. There was no survey in 2003. The continuous survey started in the last 2 weeks of 2003 but the data are identified as 2004 as is the report.

Secondly, If everyone (say a survey with 20,000 women) gives a number of ideal number of children of either sex then 0 boys and 0 girls would apply to everyone but 20,000 women could have a number other than 0. If all wanted 2 children of either sex, then 20,000 would have a value of 2 for v629 and v613 but 0 for v627 and v628. Therefore, no inconsistency.

Re: Ideal fertility sons and daughters v613 v627 v628 [message #14714 is a reply to message #14664] Tue, 01 May 2018 15:40
 rkbroussard@prc.utexas.ed Messages: 4Registered: April 2018 Location: Austin, Texas Member
Thanks so much for your response! Just to clarify, then, that would mean that a value of 0 on variable v629 (Ideal number of either sex), does NOT necessarily mean that the respondent would like 0 children. Instead, it means that the respondent simply did not provide a response for v629, but rather only provided a response for v627 and v628?

If this is the case, then why are these non-responses labeled as 0 rather than just missing? And is there any way to find the respondents who actually said they would ideally have 0 children?

Kathleen Broussard
Re: Ideal fertility sons and daughters v613 v627 v628 [message #14747 is a reply to message #14714] Thu, 03 May 2018 13:40
 Bridgette-DHS Messages: 2549Registered: February 2013 Senior Member
Following is a response from Senior DHS Stata Specialist, Tom Pullum:

I recommend that you do these calculations separately for different years, rather than pooling across the range 2003-11.

These variables can have codes 96 or 98, and cases with those codes must be excluded from the calculations.

Think of the data structure as just a set of rows representing respondents and columns representing responses to questions, or variables, just like an Excel spreadsheet. In a given row, v613 will be the sum of v627 and v628 and v629 (ignoring non-numeric responses). That is, if v617=1 and v628=1 and v629=1, then v613 will be 1+1+1=3. Of course, there are many other combinations of v627, v628, and v629 for which the sum will be 3. For example, 0+0+3=3.

Below are Stata lines showing that v613 is indeed the sum of v627, v628, and v629 in the file for 2010 (PEIR61FL.dta).

```set more off
set maxvar 10000

use "C:\Users\26216\ICF\Analysis - Shared Resources\Data\DHSdata\PEIR61FL.DTA", clear

tab v007

describe v613 v627 v628 v629

gen v613r=v613
replace v613r=. if v613>=96

gen v627r=v627
replace v627r=. if v627>=96

gen v628r=v628
replace v628r=. if v628>=96

gen v629r=v629
replace v629r=. if v629>=96

regress v613r v627r v628r v629r
```

• Attachment: table.GIF