Event
Mark Stinner, MSc Statistics & Mathematics, Methodologist, Statistics Canada
Tuesday, February 20, 2018 15:30to16:30
Purvis Hall
Room 24, 1020 avenue des Pins Ouest, Montreal, QC, H3A 1A2, CA
Disclosure Control and Random Tabular Adjustment.
Education: MSc Statistics (University of Manitoba); MSc Mathematics (McMaster); BSc (Honours) Mathematics (University of Manitoba) Work: Instructor in Mathematics and Statistics (University of Manitoba, University of Winnipeg); Methodologist at Statistics Canada, Generalized Systems Unit, International Cooperation and Corporate Statistical Methods Division, Analytical Studies, Methodology and Statistical Infrastructure Field
Statistical agencies are interested in publishing useful statistical data but doing so may lead to the disclosure of individuals鈥� pri卢vate data. This is a problem, as it leads to a trade-o铿� between the utility of the published data and the risk of disclosure of con铿乨en卢tial data. Disclosure control can be seen as the use of methods to deal with this problem by assessing and controlling the risk of dis卢closing con铿乨ential data while also providing researchers with useful statistical data. This talk describes a disclosure control model based largely on Bayesian decision theory. This model allows for the description of the concepts of disclosure control in terms of familiar statistical con卢cepts such as expectation and variance. A method of disclosure con卢trol, called Random Tabular Adjustment (RTA), is described. This method controls the risk of disclosure by randomly adjusting the data instead of suppressing cells. It 铿乼s naturally into the disclosure control model described.
Statistical agencies are interested in publishing useful statistical data but doing so may lead to the disclosure of individuals鈥� pri卢vate data. This is a problem, as it leads to a trade-o铿� between the utility of the published data and the risk of disclosure of con铿乨en卢tial data. Disclosure control can be seen as the use of methods to deal with this problem by assessing and controlling the risk of dis卢closing con铿乨ential data while also providing researchers with useful statistical data. This talk describes a disclosure control model based largely on Bayesian decision theory. This model allows for the description of the concepts of disclosure control in terms of familiar statistical con卢cepts such as expectation and variance. A method of disclosure con卢trol, called Random Tabular Adjustment (RTA), is described. This method controls the risk of disclosure by randomly adjusting the data instead of suppressing cells. It 铿乼s naturally into the disclosure control model described.