32 – Analysis of Safety Data and Rare Events
Stay Out of the Emergency Room: A History of Coping and Improving in the SCF
Arthur Kennickell
Federal Reserve Board
Viewed in a larger context, sometimes the worst problems become the greatest opportunities for learning. The Survey of Consumer Finances (SCF) could be seen as a methodological nightmare. Among other things, the survey must describe the highly skewed distribution of wealth, the subject of the survey is generally considered sensitive, the questions are necessarily at least somewhat technical, financial literacy varies across households, understanding of and attention to language more generally may also vary, the interview is long, and interviewers may vary in the extent to which they understand and follow the intended protocols; the joint consequence is that both unit and item nonresponse and other sources of nonsampling error have been important problems. There is an obvious immediate need to cope with such "emergency room" problems in order to have data for analysis, but for the long term it is more important to frame resolution in terms of steps toward prevention and other "public health" measures. Such reframing can lead to continuous improvement.