I would like to check my understanding of how to use the weights from the covid survey. If I'm combining with earlier full waves (e.g. wave I) would I just use the most recent _xw weight from the covid surveys? Do I need to talk about any caveats from using this weight when looking back on earlier full waves or is it completely fine to use as I would normally? Will it still be representative in the same way or do I need to make sure to point out that only those who answer the covid surveys are in the sample and that this may not be as representative of the UK as usual?
Updated by Alita Nandi over 3 years ago
- Status changed from New to Feedback
- Assignee changed from Olena Kaminska to Laura M
- % Done changed from 0 to 80
- Private changed from Yes to No
As it says in the user guide "or the purposes of weighting, the COVID-19 study is treated as an instrument of the wave nine annual interview." These weights have been constructed such that it is non-zero for the sample who responded to both covid19 survey and Wave 9 (not either or).
If you use these weights with covid19 survey variables or with covid19 survey + W9 variables, your estimates will be representative of the UK population.
If you use these wights with covid19 survey data + earlier than W9 survey data, your effective sample will be that subsample of those who responded to both covid19 survey and Wave 9 (not either or) and those who also responded to the earlier waves you are including. And so the weights will not have adjusted for the earlier wave non-response.