Variance Estimation for Census Transportation Planning Products with Perturbed American Community Survey Data

• Activity Number: 625
• Type: Contributed
• Date/Time: Thursday, August 4, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): Jianzhu Li*+ and Tom Krenzke and J. Michael Brick and David Judkins and Michael D. Larsen
• Companies: Westat Inc. and Westat Inc. and Westat Inc. and Westat Inc. and The George Washington University
• Keywords: Perturbation ; disclosure limitation ; successive differences replication ; confidentiality

Abstract

Census Transportation Planning Products (CTPP) is a set of tabulated data products designed for transportation planners. As the underlying data is moving from the Census Long Form sample to the smaller American Community Survey (ACS) five-year combined sample, disclosure risk becomes an avoidable concern, especially for small geographic areas such as Travel Analysis Zones. Three perturbation approaches were evaluated with respect to utility of the CTPP products, and conforming to Census Bureau's requirements for protecting against disclosure risk. This paper discusses the variance estimation on CTPP using perturbed ACS data from our research. ACS implements the Successive Differences Replication (SDR) method for variance estimation, with the advantage that the variance estimates can be computed regardless of the form of the statistics or the complexity of the design. However, the SDR estimator applied naively to the perturbed data does not account for the variance due to perturbation. As a remedy, an additional term was added to reduce the bias. The proposed estimators are compared with a few alternatives and evaluated through a simulation study.