Record-level Measures of Disclosure Risk for Survey Microdata

Download

PDF Download (342Kb)

Description/Abstract

Measures of disclosure risk at the record level have a variety of potential uses in statistical disclosure control for microdata. We propose a new record level measure of disclosure risk which is the probability that a unique match between a microdata record and a population unit is correct. For discrete key variables subject to no measurement error, we study this measure under the assumption of a Poisson model and a Poisson-gamma model. Moreover, we apply the approaches to a sample of microdata from the U.K. General Household Survey. The results indicate that the risk measure may be used to establish whether sample unique records are unique in the population.