Measurement error in savings data
I am currently working with the savings data in USoc (save, saved, svamt etc.) and the savdocs variable shows that very few (less than 10%) actually have shown to the interviewer any documents on their savings. I was wondering if there is any systematic information on how the distributions given by the variables is compared to real data on savings or more focused wealth-assets surveys (such as WAS). And, secondly if anything has been done to deal with it.
#1 Updated by Stephanie Auty 11 months ago
- Private changed from Yes to No
- % Done changed from 0 to 70
- Assignee set to Mihai Codreanu
- Status changed from New to Feedback
- Category set to Data analysis
As this question was only asked of those who have savings accounts, in Wave 8 13% of respondents providing a valid response checked relevant documents. As the variables you mention are not derived variables, they have not been edited and are as reported by the respondent.
For previous research using our data, you could search the website for relevant key words, or for discussion of analysis you could join our jiscmail list by emailing UKHLS-REQUEST@JISCMAIL.AC.UK