The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » General » WEIGHTS AFTER DROP OUTS ( Do I Need to Generate New Weights After Dropping ICT, AJK, and GB Regions in PDHS 2017-2018?)
WEIGHTS AFTER DROP OUTS [message #29975] Tue, 03 September 2024 16:27 Go to next message
Noorulain is currently offline  Noorulain
Messages: 3
Registered: August 2024
Member
Hello,

I am working with the PDHS 2017-2018 Individual Recode (IR) file, and I have excluded the regions of Islamabad Capital Territory (ICT), Azad Jammu and Kashmir (AJK), and Gilgit-Baltistan (GB) from my analysis using the variable V139 (Region).

My question is:

Do I need to generate new sample weights after excluding these regions (ICT, AJK, GB) from my dataset?
If new weights are required, could you please guide me on how to properly generate these weights to ensure my analysis remains representative?

2) My other Question is,
I need to calculate the proportion or total of all recent live births within the last 5 years.

I'm unsure whether I should extract this information from the Birth Recode (BR) file or the Individual Recode (IR) file. Could you please advise on which file is more appropriate for obtaining accurate estimates of recent live births? What is the code for this in DHS Recode manual?
Re: WEIGHTS AFTER DROP OUTS [message #30001 is a reply to message #29975] Mon, 09 September 2024 10:18 Go to previous message
Janet-DHS is currently offline  Janet-DHS
Messages: 899
Registered: April 2022
Senior Member
Following is a response from DHS staff member, Tom Pullum:

First, you do not need to re-calculate the weights.

You can calculate what I think you want from either the IR or the BR file. You just have to remember that in the IR file, the cases are women, and in the BR file, the cases are births. In the IR file, the number of births in the past 5 years is v208 and the number of births ever is v201.

Open Stata and enter the following lines
use "...PKIR71FL.DTA" , clear

keep v208 v005

gen birth_5yrs=1

expand v208

total birth_5yrs [iweight=v005/1000000]

 

use "...PKIR71FL.DTA" , clear

keep v201 v005

gen birth_ever=1

expand v201

total birth_ever [iweight=v005/1000000]

You will see that the weighted number of births in the past 5 years is 10,494 and the weighted number ever is 40,831.

You can also get these numbers from "tab v201 v208 [iweight=v005/1000000]" but it's more complicated because the cases are women, not births.

You will get these numbers from the BR file with these lines:
use "...PKBR71FL.DTA" , clear

gen last_5yrs=0

replace last_5yrs=1 if b19<60

tab last_5yrs [iweight=v005/1000000]
Previous Topic: Weithted analysis in R
Next Topic: 2011 wealth class data on access to electricity in Benin Burundi Cote d'Ivoire Ethiopia and Cambodia
Goto Forum:
  


Current Time: Mon Dec 9 05:05:41 Coordinated Universal Time 2024