Email)">
10:00 - 11:00
Contributed Paper Session
Room: JENK
Chair:
Aleksandra BUJNOWSKA, Eurostat, Luxembourg, (Email)
Discussant:
Rachel Skentelbery, Office for National Statistics, United Kingdom, (Email)
Perturbative methods for ESS census tables
Peter-Paul de Wolf, (Email)
Statistics Netherlands (CBS), The Hague
The population census has been an important output of official statistics for a long time. In the current global situation, researchers are increasingly interested in combining census information from different countries. In Europe an attempt to facilitate this more easily was taken by Eurostat in developing the Census Hub: a software system where census tables from the participating EU member states can be found. These member states each filled the Census Hub with their own census tables. The census tables are about personal information and are giving information at quite detailed level: crossings of many descriptive variables, often with detailed categories. Hence, even though tabulated data, some statistical disclosure control methods are needed to protect the individual privacy of the people in the census tables of those member states. Up until the 2011 census, all member states defined and used their own 'proper' SDC method(s). This led to the undesired situation that although the tables from the member states were available through a single portal, combining the information over different countries in a useful way was in some cases difficult if not impossible. Consequently, a project in which the European Centre of Excellence on SDC was asked to harmonise the SDC approaches of the different countries was launched. As a result two SDC methods were proposed: targeted record swapping (TRS) and the cell-key method (CKM). These methods were tested by the partners of the Centre of Excellence on SDC that took part in the project, using SAS software obtained from ONS. By suggesting to use one or both of these methods as harmonised approach by the EU member states, it became apparent that a more general implementation in (existing) open source SDC software would be needed. Hence, Eurostat asked the Centre of Excellence on SDC to provide these implementations. In the current extended abstract, we will quickly describe the suggested methods and introduce the implementations that have been developed.


Reference:
CPS08-002
Session:
Ensuring privacy when accessing to data
Presenter/s:
Peter-Paul de Wolf
Presentation type:
Oral presentation
Room:
JENK
Chair:
Aleksandra BUJNOWSKA, Eurostat, Luxembourg, (Email)
Date:
Thursday, 14 March
Time:
10:00 - 11:00
Session times:
10:00 - 11:00