Abstract:
|
Policymakers, businesses, researchers, and many others rely on the data collected and released by the federal statistical system. When producing these data products, statistical agencies must ensure that respondent privacy has been properly protected. As the risk of re-identification grows (driven by improvements in computing power and the increasing availability of data), statistical agencies must take new and better steps to protect privacy, and those steps have serious implications for the availability of data for research and evaluation, and the usability/quality of the data being made available. This panel will examine these trade-offs, and discuss how to find the appropriate balance (i.e., the right "privacy budget") between protecting individuals' privacy, and producing high-quality, high-utility data products.
|