Producing Transportation Data Products from the American Community Survey that Comply with Disclosure Rules

For the last 40 years, the American Association of State Highway and Transportation Officials (AASHTO) has worked with the U.S. Census Bureau to produce special tabulations from the decennial census long form. These tabulations serve a variety of transportation planning needs. In 2006, AASHTO approved a new Census Transportation Planning Products (CTPP) program to continue to provide home, work place, and journey to work data. The CTPP will now use data from the U.S. Census Bureau's new American Community Survey (ACS) to produce 3-year and 5-year data tabulations to support a host of state and local transportation planning efforts, including air quality and environmental analyses, transit studies, policy and investment scenarios, and travel demand modeling. Over the years, transportation planning mandates and requirements have increasingly called for census data at finer levels of granularity for smaller areas of geography. At the same time, the U.S. Census Bureau has been increasingly concerned about the potential disclosure risk posed by such finer granularity. The U.S. Census Bureau currently uses data suppression and use of thresholds to protect confidentiality. These methods severely limit the availability and utility of the resulting data. For example, in travel demand modeling, data are typically required for smaller geographic units defined as Transportation Analysis Zones (TAZs). However, because TAZs tend to be small, data for many geographic areas will be suppressed under U.S. Census Bureau disclosure rules aimed at protecting an individual's confidentiality. The transportation community and U.S. Census Bureau staff acknowledge the need to develop and implement improved disclosure avoidance techniques and tools that will preserve the geographic detail and content of the data. Some methods that are either currently used by the U.S. Census Bureau and other federal statistical agencies or have been proposed for use in the statistical literature for purposes of disclosure limitation or disclosure avoidance include: aggregation (coarsening of categorical data, often geographical); top-coding of reported values (another form of aggregation); tabular data suppression based on thresholds; data swapping and rounding (possibly using linear programming); noise addition or inoculation to prevent compromise by introducing an error; microaggregation (clustering data in groups of a small number of records replacing each member's value with the average value); imputation of sensitive values using multiple imputation to assess variability; and partial or whole synthetic creation of a micro data set with closely related distributional properties to that for the collected data (using various parametric statistical models fit to the collected data, especially log-linear models for contingency tables). Although disclosure avoidance can be achieved using different techniques, these methods must be acceptable to the transportation community and meet with the U.S. Census Bureau's rules, requirements, and processes. Preliminary research done under NCHRP Project 8-36 (71), "Disclosure Avoidance Techniques to Improve ACS Data Availability," ( suggests that credible techniques can be found to produce disclosure-proofed representative data for small areas. The objective of this research is to develop, evaluate, and test credible techniques to produce specially tabulated data products using 5-year ACS data. The resulting data products must satisfy U.S. Census Bureau disclosure rules and support transportation planning at small area geography (TAZ).


  • English


  • Status: Active
  • Contract Numbers:

    Project 8-79

  • Sponsor Organizations:

    Federal Highway Administration

    1200 New Jersey Avenue, SE
    Washington, DC  USA  20590

    American Association of State Highway & Transportation Officials

    444 North Capitol Street, NW, Suite 225
    Washington, DC    20001

    National Cooperative Highway Research Program

    Transportation Research Board
    500 Fifth Street, NW
    Washington, DC  USA  20001
  • Project Managers:

    Srinivasan, Nanda

  • Performing Organizations:

    Westat, Incorporated

    1650 Research Boulevard
    Rockville, MD  USA  20850
  • Principal Investigators:

    Krenzke, Thomas

  • Start Date: 20100107
  • Actual Completion Date: 20110706
  • Source Data: RiP Project 22349

Subject/Index Terms

Filing Info

  • Accession Number: 01464203
  • Record Type: Research project
  • Source Agency: National Cooperative Highway Research Program
  • Contract Numbers: Project 8-79
  • Files: TRB, RiP, USDOT
  • Created Date: Jan 3 2013 2:38PM