The DHS Program User Forum: Weighting data » analysing change over time and weights normalisation

Home » Data » Weighting data » analysing change over time and weights normalisation

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

analysing change over time and weights normalisation [message #25752]

Mon, 05 December 2022 10:20

geoK
Messages: 32
Registered: May 2014

Member

Hi all,
I am trying to replicate a similar type of analysis to this work published on BMJ (https://gh.bmj.com/content/7/4/e007798) where i want to study changes over time between NFHS-4 and NFHS-5 for a set of outcome variables and compare round 1 with round 2. I was wondering whether there is need to denormalise the weights at all as i will be working with 2 rounds of surveys. Many thanks.

Report message to a moderator

Re: analysing change over time and weights normalisation [message #25761 is a reply to message #25752]

Tue, 06 December 2022 08:39

Bridgette-DHS
Messages: 2721
Registered: February 2013

Senior Member

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

What you propose is a comparison of two statistically independent samples. We recommend no changes to the weights.

Report message to a moderator

Re: analysing change over time and weights normalisation [message #25764 is a reply to message #25761]

Tue, 06 December 2022 11:48

geoK
Messages: 32
Registered: May 2014

Member

Many thanks.

Report message to a moderator

Previous Topic:	NFHS-5, when to calculate new weights?
Next Topic:	Resampling data to obviate the need for weighting

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Mon Mar 20 06:30:27 Coordinated Universal Time 2023