The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Bangladesh » Number of observations and age at time of birth
Number of observations and age at time of birth [message #26678] Sun, 16 April 2023 18:55 Go to next message
jangels1 is currently offline  jangels1
Messages: 4
Registered: February 2023
Member
First, a very basic question: I'm working with the 2017/2018 birth recode dataset and I'm just trying to figure out how many women were interviewed. I notice in the narrative report for 2017/2018 in section 1.9 that "Among the 20,376 ever-married women age 15-49 eligible for interviews, 20,127 were interviewed, yielding a response rate of 99%." When I do a simple .codebook caseid for the birth recode file, the number of observations is 18,134. And it appears that there are 47,828 observations in the file in total. What does that number of observations refer to, and why is there a difference between the number of case IDs and the number of women reported interviewed?

Second, I'm trying to figure out how the age of the woman at the time of each birth was determined. In the narrative report on variables associated with place of delivery (table 9.6), it notes "Mother's Age at Birth." I see a number of variables in the .dta file that record the mother's age, but can't figure out which would be the relevant ones for age at a particular birth and how to tab that with place of delivery (or cesarean section, etc).

Thanks in advance for your help!
Re: Number of observations and age at time of birth [message #26697 is a reply to message #26678] Wed, 19 April 2023 12:10 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3087
Registered: February 2013
Senior Member

Following is a response from DHS staff member, Tom Pullum:

I will assume you are talking about the Bangladesh 2017-18 survey. Users are asked to specify which survey they are asking about. And which file? The PR file includes everyone in the household, regardless of age or marital status. In that file, hv017 is 1 for women who were eligible for the IR file. BD 2017-18 was an ever-married survey, so marital status was relevant for the construction of hv017. The response rate for women is the number of cases in the IR file divided by the number of cases with hv017=1 in the PR file.

There are weighted totals and unweighted totals. The weighted total and unweighted total should agree for the entire PR file or IR file, but not for subpopulations. The hv005 and v005 weights are a little different. Not sure which ones you are talking about.

The birth recode file has one record for each child in the birth histories. The weights come from the mother. The weighted totals in the BR file do NOT match the unweighted totals.

Mother's age at the birth of a child should be int((b3-v011)/12). Here, b3 is the cmc of the child's birth and v011 is the cmc of the mother's birth.

Re: Number of observations and age at time of birth [message #26717 is a reply to message #26697] Fri, 21 April 2023 14:08 Go to previous message
jangels1 is currently offline  jangels1
Messages: 4
Registered: February 2023
Member
This is very helpful, thank you!
Previous Topic: Calculation of standard error of stunting at small subpopulation such as district
Next Topic: Determining child's birth order
Goto Forum:
  


Current Time: Wed Jun 19 19:04:24 Coordinated Universal Time 2024