Project

General

Profile

Support #1892

Question regarding longitudinal weights

Added by Johanna Pauliks about 1 year ago. Updated 7 months ago.

Status:
Resolved
Priority:
Normal
Category:
Weights
Start date:
04/19/2023
% Done:

100%


Description

Dear support,

I've got another question regarding longitudinal weights in UKHLS. I'm using data from adult respondents (individuals) from wave 2 up to wave 10 and using the fixed-effects model (so I'm doing longitudinal analysis). According to the user guide, longitudinal weights are appropriate for this. But if I use the longitudinal weight from wave 10 I lose all cases of individuals who drop out of the sample in previous waves, even so they participated in all waves prior to this (let's say they participated from wave 2-8). This is a problem, as some of my subgroups are very small to begin with. Would it be possible to use the longitudinal weight from wave 10 for everyone who participated in all waves up to wave 10, the longitudinal weight from wave 9 for every respondent who participated until wave 9 and so on, or would this not be appropriate, and I need to create my own tailored weights?

Best regards
Johanna

Also available in: Atom PDF