NCHRP 08-36/Task 071 [Final]
Disclosure Avoidance Techniques to Improve ACS Data Availability
[ NCHRP 08-36 (Research for the AASHTO Standing Committee on Planning) ]
| Project Data
||Cambridge Systematics |
||The project has been completed and the final report submitted to AASHTO|
Data dissemination rules initiated by the DRB caused significant loss of data for CTPP 2000, and are expected to cause an even more serious loss when applied to products emanating from the American Community Survey. This report summarizes the development and testing of synthetic data techniques that the Census Bureau could employ to provide alternative data, particularly in regards to journey-to-work flow data at small geography from the Decennial Census (or from the 5-year accumulation of the American Community Survey). Multi-year ACS data tables are inherently protected from potential harmful disclosure of participants, and therefore do not require further disclosure restrictions. However, if disclosure-proofing is nevertheless deemed necessary, using a form of data synthesis is probably the best approach from the viewpoint of the transportation planning data user. This research examines alternative data synthesis approaches that could be employed for multi-year ACS journey-to-work tables, including iterative proportional fitting, combined Bayesian/iterative proportional fitting, and the Generalized Shuttle Algorithm.
The report, Disclosure Avoidance Techniues to Improve ACS Data Availability for Transportation Planners, may be found here.