The Census Bureau and its state partners are committed to protecting the confidentiality and integrity of the underlying workforce and employment data files. Statistical techniques are employed by the Census Bureau to ensure that actual statistics are not shown if the numbers in a cell are small. Rather, the mapping displays and tabular reports show synthetic data that are statistically analogous to actual worker counts and locations but not exact.
Technically, the method used to protect the residential address information for each workplace address is called "synthetic data," which involves modeling the underlying data in a manner that protects confidentiality, then making random draws from that model. The method used to protect the workplace information is to combine the confidential data with noise in a manner that ensures that the published data, while not exact, become increasingly more accurate as the number of businesses in a workplace area gets large. No business's actual data are used for any workplace reports. The employment data for private employers are controlled to state-level totals provided by the Bureau of Labor Statistics.
Only Census Bureau employees or individuals who have Special Sworn Status are permitted to work with the underlying confidential data. Everyone who has access to U.S. Title 13 data must have an official security clearance based on a background check, including fingerprinting and face-to-face interviews of selected references, conducted by the Office of Personnel Management. The Census Bureau and state data custodians closely review all projects before release to avoid disclosure of confidential information.
The underlying confidential data used to create OnTheMap Versions 3-6 were protected using synthetic data methods that offer formal privacy guarantees. The peer-reviewed paper describing these methods is available from the International Conference on Data Engineering as:
Machanavajjhala, Ashwin, Daniel Kifer, John Abowd, Johannes Gehrke, and Lars Vilhuber, "Privacy: Theory Meets Practice On the Map," International Conference on Data Engineering (ICDE) 2008, pp. 277-286.
Preprint version of the Machanavajjhala paper
(204 KB)
The final version of the paper can be accessed at IEEE Xplore
.
Other documents helpful for understanding the confidentiality protection system used in OnTheMap: