Online Program

Public use business establishment microdata: Protecting confidentiality and providing utility with synthetic data

Keywords: confidentiality, disclosure, imputation, microdata, synthetic

Business microdata are very difficult to disseminate as unrestricted public use files, because the risks and consequences of disclosure are high. Indeed, even the fact that a business appears in a national database may be protected by law. Because of the high risks, agencies may need to alter data significantly before sharing them. One approach to doing so is via synthetic data, in which agencies release data with values simulated from statistical models. I review the synthesis of the U.S. Longitudinal Business Database, including discussions of disclosure protection and evaluations of data utility.