A CRONYM : Data without Boundaries D

Disclosure limitation methods for protecting the confidentiality ofrespondents in survey microdata often use perturbative techniques whichintroduce measurement error into the categorical identifying variables. Inaddition, the data itself will often have measurement errors commonly arisingfrom survey processes. There is a need for valid and practical ways to assess theprotection against the risk of identification for survey microdata withmeasurement errors. A common disclosure risk scenario is when an intruderseeks to match the microdata with an external file. We will examineprobabilistic record linkage as a means of assessing disclosure risk and relate itto disclosure risk measures under the probabilistic framework of the Poissonlog-linear models.