The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Weighting data » Sampling weight calculation for Under five children of Bangladesh
Sampling weight calculation for Under five children of Bangladesh [message #15426] Tue, 17 July 2018 22:32 Go to next message
sabbiroeh@gmail.com is currently offline  sabbiroeh@gmail.com
Messages: 2
Registered: February 2018
Member
Dear,



I , Dr. Sabbir Ahmed , am currently doing public health research under MPH in University of Malaya, Malaysia.



I am using the secondary data set ( DHS 2014) of Demographic health Survey (DHS) regarding a topic


I am facing trouble to convert the sample estimate to population estimate. Specially the V005 variabe of BDKR72FL file.

It has only weightage for women . so to divide the data of V005 by 1000000 to get the women population.

But how can I calculate the total number of Under five children of Bangladesh from this data set ?

Please give your valuable input.

Regards,
Sabbir
Re: Sampling weight calculation for Under five children of Bangladesh [message #15941 is a reply to message #15426] Mon, 08 October 2018 18:07 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 1504
Registered: February 2013
Senior Member
Following is a response from Senior DHS Stata Specialist, Tom Pullum:

The best source for an estimate of, say, the number of children under 5 in Bangladesh at the time of the survey would be from the website for the UN Population Division (https://population.un.org/wpp/). Once you have such a number, you can use DHS data to estimate, say, the number of children under five who do not have a full set of immunizations, by applying the proportion in the DHS sample to the population age 0-4 estimated by the UN.

The DHS weights are calculated in such a way that the sum of v005/1,000,000 is equal to the total sample size (of women). You might think that if you inflate those weights by a factor 1/F, where F is the sampling fraction, you will get population estimates. That's correct, but the estimates would not be as good as if you use another source such as the UN numbers I mentioned above.

Previous Topic: Using weights in regression analysis
Next Topic: Weights to be used for matching the sex ratio from the NFHS report
Goto Forum:
  


Current Time: Tue Dec 11 02:40:14 Eastern Standard Time 2018