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.