Federal statistical agencies are the national's largest gatherer and consumer of data. They are supposed to disseminate information and also protect the privacy of individuals and establishments that are described by the data. It is hard to find ways to get enough data to the people who need to make important decisions, such as budgeting for government programs, and also keep the person or business' identity private.
Outcomes & Results
NISS developed a prototype World Wide Web-based system to allow adequate access to confidential data from Federal agencies while maintaining low risk of disclosure. NISS also developed a data swapping took kit (DSTK), which was a set of software programs and tools for performing and analyzing data swapping, which is basically switching subsets of attributes between randomly selected pairs of records. This way, an intruder will not be sure if the record is real or not.
NISS was hired to help develop and build systems that expanded to Federal data but that preserved the confidentiality of the data and privacy of subjects. The systems would respond to queries from networked users of Federal data bases by performing and reporting statistical analyses that extract knowledge from the data, but preserve confidentiality. Their distinguishing characteristic was history-dependence. The response to each query will depend on the history of previous queries and responses.
System Development included: Susan Karimi, Karen Littwin, Bonnie Parrish and Syam Sundar, all from MCNC
Federal Agency Partners included: Bureau of Labor Statistics, U.S. Census Bureau, National Agricultural Statistics Service, National Center for Education Statistics, National Center for Health Statistics
NISS Technology Day, Washington DC, May 2002
Technical Report 106: Visual Scalability
Technical Report 116: Bounds for Cell Entries in Contingency Tables Induced by Fixed Marginal Totals
Technical Report 121: Disclosure Risk vs. Data Utility: The R-U Confidentiality Map
Develop and build systems for federal statistical agencies that preserved the confidentiality of the data and privacy of subjects
Funding Sponsor(s): National Science Foundation
Principal Investigator(s): Alan Karr, NISS
Senior Investigator(s): Stephen Fienberg, Carnegie Mellon; George Duncan, Carnegie Mellon; Sallie Keller-McNulty, Los Alamos National Laboratory; M. Franklin, Maryland; Alan Saalfeld, Ohio State; Andrew Moore, Carnegie Mellon; Stephen Roehrig, Carnegie Mellon, Latanya Sweeney, Carnegie Mellon
Post Doctoral Fellow(s): Ashish Sanil; James Hilden-Minton;
Graduate Student Assistants: Karen Brady; Adrian Dobra; Christopher Hollowman, Duke