Travel demand models have advanced from zone-based methods to favor activity-based approaches that require more disaggregate data sources. Household travel surveys gather disaggregate data that may be used to inform advanced travel demand models better and also to improve the understanding of how nonmotorized travel is influenced by a household's surrounding built environment. However, the release of these disaggregate data is often limited by a confidentiality pledge between the household participant and survey administrator. Concerns about the disclosure risk of survey respondents to household travel surveys must be addressed before these household-level data may be released at their disaggregate geography. In an effort to honor this confidentiality pledge and facilitate the dissemination of valuable travel survey data, this research (a) reviews geographical perturbation methods that seek to protect respondent confidentiality; (b) outlines a procedure for implementing one promising practice, referred to as the "doughnut masking technique"; and (c) demonstrates a proof of concept for this technique on 10 respondents to a household activity travel survey in the Portland, Oregon, metropolitan region. To examine the balance between limiting disclosure risk and preserving data utility, four trials were conducted and measures of household anonymity and built environment variation were analyzed for the relocated household in relation to its actual location. Results of this demonstration revealed that increases in the potential displacement distance of a geographically perturbed household generally reduced disclosure risk and also limited data utility.
Available at: http://works.bepress.com/kelly_clifton/26/