Numerical Impact of Topcoding on CE Microdata Utility

• Activity Number: 636
• Type: Topic Contributed
• Date/Time: Thursday, August 8, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Daniel Yang*+ and Daniell Toth
• Companies: Bureau of Labor Statistics and Bureau of Labor Statistics
• Keywords: Consumer Expenditure (CE) ; microdata ; confidentiality ; disclosure limitation ; data utility ; topcoding

Abstract

The Consumer Expenditure (CE) Survey implements a statistical disclosure limitation (SDL) process in its microdata release to conceal sensitive and identifiable information in order to protect the household's confidentiality. This process is also known as "topcoding," e.g. the high (low) end income household's annual income will be replaced by the average of high-income (low-income) household's annual income in the microdata for public users. Topcoding will have a numerical impact on microdata's utility and data quality. Several data utility measures have been applied to assess synthetic microdata, such as confidence interval overlap, propensity score comparisons, cluster analysis and difference in empirical CDFs (cumulative density functions). In this study, we evaluate the numerical impact of topcoding on CE microdata utility focusing on the analysis of expenditure items that are: highly correlated with income and highly topcoded, highly correlated but not highly topcoded, not highly correlated but highly topcoded.