The DHS Program User Forum: Weighting data

Home » Data » Weighting data » weighting data

Show: Today's Messages :: Show Polls :: Message Navigator

weighting data [message #30345]

Sun, 10 November 2024 04:01

Hejie Wang
Messages: 16
Registered: July 2024

Member

I have a question about new weights for merged surveys, my idea is to multiply the proportion of each survey to the merged file by the weight of that observation (wt), is this the right thing to do?

This R code is here:
data$wt<-data$v005/100000
survey_counts <- table(data$v000)
total_samples <- nrow(data)
survey_proportions <- survey_counts / total_samples
data$survey_proportion <- survey_proportions[data$v000]

Can you give me some advice about this weighting?

Report message to a moderator

Re: weighting data [message #30351 is a reply to message #30345]

Mon, 11 November 2024 14:09

Bridgette-DHS
Messages: 3230
Registered: February 2013

Senior Member

All of your questions have been asked and answered earlier on the forum.

Report message to a moderator

Re: weighting data [message #31151 is a reply to message #30345]

Thu, 10 July 2025 05:37

MichaelMWilson
Messages: 1
Registered: July 2025

Member

Make sure your calculations reflect the correct proportions, as this can significantly affect the results of your analysis!
Edit by tap road a day ago

[Updated on: Thu, 10 July 2025 08:17]

Report message to a moderator

Previous Topic:	Using weights for small geographical locations
Next Topic:	Sampling

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Tue Jul 22 15:33:06 Coordinated Universal Time 2025