A Bayesian Approach to Data Disclosure: Optimal Intruder Behavior for Continuous Data

Stephen E. Fienberg, Udi E. Makov and Ashish P. Sanil


In this paper we develop a Bayesian approach to data disclosure in survey settings by adopting a probabilistic definition of disclosure due to Dalenius and the principle that a data collection agency must consider disclosure from the perspective of an intruder in order to efficiently evaluate data disclosure limitation procedures. Our approach leads to a formal model involving mixture distributions. We then discuss the implementation of a simplified version of the model which is made computationally feasible by the use of Gibbs sampling in conjunction with several approximations. We apply the methods in a small-scale simulation study using data extracted and adapted from an actual survey conducted by the Institute for Social Research at York University. Keywords: Confidentiality; Disclosure limitation; Gibbs sampling; Inferential disclosure; Measurement error.

Here is the full postscript text for this technical report. It is 180 kbytes.