|
Activity Number:
|
468
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 7, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #302192 |
|
Title:
|
Controlling Disclosure Risk in Synthetic Public Use Microdata: The Longitudinal Business Database
|
|
Author(s):
|
Arnold Reznek*+ and John Abowd and Saki Kinney and Jerome Reiter and Javier Miranda
|
|
Companies:
|
U.S. Census Bureau and Cornell Univeresity and Duke University and Duke University and U.S. Census Bureau
|
|
Address:
|
4600 Silver Hill Road, Suitland, MD, 20233,
|
|
Keywords:
|
Confidentiality ; Disclosure risk ; Synthetic microdata
|
|
Abstract:
|
The U.S. Census Bureau's Center for Economic Studies (CES) and academic researchers are developing synthetic microdata public use files from the Longitudinal Business Database (LBD). The LBD longitudinally links variables for business establishments from the Census Bureau's Business Register annual files. Synthetic microdata sets are designed to mimic the distributions of the underlying "real" microdata data and allow researchers to carry out analytically valid studies, without disclosing confidential information about respondents. The "synthetic LBD" files contain synthesized data on establishments' employment and payroll and on their birth and death years, as well as industry and geography. After briefly describing how we are creating the synthetic LBD and evaluating its analytic validity, we describe the confidentiality protection measures we are applying.
|
