Finding a Needle in a Haystack: The Theoretical and Empirical Foundations of Assessing Disclosure Risk for Contextualized Microdata
Kristine Witkowski · Deep Blue (University of Michigan) · 2008
Contextualized microdata are one way to safely release geographic data without identifying the location of survey respondents. This study informs the design of such datafiles with its needle-in-haystack approach to disclosure and its discussion of associated methodological concerns. Drawing a sample of counties, tracts, and blockgroups, I illustrate how the reidentification of individuals is shaped by aggregating geographies into look-alike sets. I detail the complexity of reidentification patterns by assessing the likelihood that young adult white and black males would be pinpointed within reconstituted haystacks given: (1) the size of the total population of aggregated contexts; (2) the amount of error in population counts; and (3) differential search costs stemming from spatially dispersed contexts.