The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Malaria » Computing the number of U5 children by age classes in Burundi
Computing the number of U5 children by age classes in Burundi [message #9297] Sun, 06 March 2016 08:24 Go to next message
Kwizera136 is currently offline  Kwizera136
Messages: 1
Registered: March 2016
Location: Burundi
Member
Dear DHS users,

I'm using 2010 Burundi DHS data and I'm interested in malaria.
In the DHS database (children recode), there are missing values for the child's age (HW1, in the households) in months. After weighting cases, I found 4282 children with missing values for age in months. The DHS report (page 153) shows, however, that all the 7418 children have their age in months.
I used also the variable HC1 (for interviewed women) and I found that only 3873 children have their age in months.

How can I get the correct numbers of children in age classes as found in the DHS report? What is the variable age used?

Thank you.
Re: Computing the number of U5 children by age classes in Burundi [message #9409 is a reply to message #9297] Thu, 24 March 2016 20:32 Go to previous message
Liz-DHS
Messages: 1350
Registered: February 2013
Senior Member
A response from one of our technical experts, Cameron Taylor:
Quote:

Dear user,
Great question! The age variables are always tricky with DHS data. The variable Hw1 in the kids recode (KR file) is the variable that responds to the ages of the kids that were measured for height and weight as part of the household questionnaire. The abbreviation hw is short hand for height and weight. So that variable is corresponding to the ages of children who were weighed and measured, which is not every child in the survey.

If you are in the KR file and wanting to produce the indicator "Pourcentage ayant eu de la fièvre au cours des deux semaines ayant précédé l'enquête" you don't have to specify the age of the children since data in the children's recode is only children under five years of age of interviewed mothers. However, it is always good to include an age restricting variable just in case. In this example the variable agemonth is calculated by subtracting the child's age from the date of the interview.

You can use the code below to replicate the indicator on page 153
*Open Dataset
use "BUKR61FL.DTA", clear 

*Generate the weight variable
gen wt= v005/1000000

*Mother reported fever in child during 2 weeks preceding the survey
g fever=(h22==1)

*Create age in months variable
g agemonth=.
replace agemonth=v008-b3

*Percentage who had fever in the two weeks preceding the survey
tab fever if b5==1 & agemonth<60 [iweight=wt]

*Percentage who had fever in the two weeks preceding the survey
tab fever if b5==1 [iweight=wt]


Previous Topic: ITN Ownership
Next Topic: Identifying Fevered Children Who Received an Antimalarial
Goto Forum:
  


Current Time: Fri Oct 19 01:43:08 Eastern Daylight Time 2018