Project

General

Profile

Support #1182

Pooling data for analysis by calendar year

Added by Rebecca Benson over 3 years ago. Updated 4 months ago.

Status:
Resolved
Priority:
Normal
Category:
Weights
Start date:
04/07/2019
% Done:

100%


Description

I am trying to get to the mean of certain variables for each calendar year and I am uncertain about rescaling weights. I have read the note on this (https://iserswww.essex.ac.uk/support/attachments/download/3/Pooling%20data%20from%20different%20waves.pdf). In the note, the example code is for a year where only two waves are contributing data, with the weights for data in the second wave being multiplied by (sum of w1 weights/sum of w2 weights).

For years 2013-2016, data from three waves are available. Would the appropriate modification to the weights then be:
w1 weight= w1 weight *(sum of w2 & w3 weights/sum of w1 weights)
w2 weight= w2 weight *(sum of w1 & w3 weights/sum of w2 weights)
w3 weight= w3 weight

Thanks for any guidance!

Also available in: Atom PDF