Survey participant drop out across waves
I'm using data from wave 1, 4 and 10. I've merged them and keep only participants who provided response in all waves.
The number of observations are higher when I merged only wave 1 and 4 than when I merged wave 1,4 and 10.
I assume it is because some of participants drop out in wave 10 and the number is quite significant.
Could you please suggest the information/evidence of number of participants across waves in a time series?
Updated by Understanding Society User Support Team 2 months ago
- Category set to Data management
- Status changed from New to Feedback
- % Done changed from 0 to 80
- Private changed from Yes to No
1) Yes, the drop results from survey attrition which could have happened in wave 10, but also in any of the earlier waves a given respondent participated in after wave 4.
2) We do not offer such statistics, but users can calculate that quite easily on their own. I'd recommend creating a file containing data from all 11 waves and checking the number of respondents, you could also do more detailed analysis of the patterns of attrition using such file. You can use the syntaxes available on our website "Merging individual files across waves into wide format" and "Merging individual files across waves into long format" (https://www.understandingsociety.ac.uk/documentation/mainstage/syntax)
Understanding Society User Support Team