In recent years there have been several projects at the European level to study the problems of Statistical Disclosure Control. Several new methods have been developed. These methods have not only been described in research papers but have also been made available through the development of the ARGUS software twins:µ-ARGUS for the SDC of microdata and τ-ARGUS for tabular data. This approach of coordinated methodological development and the development of the tools have proved to be very successful and the methodology included in the ARGUS-software acts as a de facto standard in Europe.
Nevertheless there are still open issues. As several methods exist for the protection of microdata, this might lead to an uncoordinated application of methods by all NSIs in Europe. This might be a reasonable solution for isolated NSIs, but at the European level there is the ambition of joining national protected datafiles to European databases. This asks for coordination. Recently IStat has proposed a framework for microdata protection. This combines a coordinated approach while leaving open the flexibility of the NSIs to opt for national solutions. Case studies are needed to test the feasibility of this approach as well as the new way of working together where "national level works for European level" and the burden of creating a new microdata file for research shared amongst member states.
At the tabular side the current τ-ARGUS is very well capable of protecting (large) tables. However there is a need for extending the methodology and τ-ARGUS allowing for the protecting of linked tables. In a previous ESSNet some first steps have been made. These have resulted is an extension of τ-ARGUS to protect a certain class of linked tables. However this will not cover the most general case. Also special features to make it more easily useable to Eurostat are still missing. Alternative heuristics and new methodologies beyond the traditional cell suppression approach have been proposed. Case studies here are needed to make concrete examples of real applications of τ-ARGUS to European surveys to demonstrate the heuristics, new methods and open issues.
On the software side there is a strong need to secure the future of the software tools for SDC. The ARGUS twins are being used by many NSIs. At this moment the development of ARGUS strongly depends on the availability of a few developers at Statistics Netherlands, but that is not enough to guarantee for the future. Also it is needed to join the forces in Europe and enlarge the group of developers; an open source solution should be investigated. Especially the Austrian initiative to build open source SDC tools should be considered in this respect. Although the current tools can solve many SDC-tasks, several problems are reported by NSIs when they try to embed ARGUS in their production systems. There are problems like different operation systems, metadata and file formats
These issues will certainly not all be solved in a one year project, but we foresee to detail the current structure of ARGUS, describe the architecture for the future software and present a roadmap for the future.
The main purpose of this ESSNet will be to pave the way for future cooperation in Europe. Therefore the main outcome of this project will be case studies investigating new directions to be realised by future projects and an architecture for making the ARGUS software more open, allowing to enlarge the team of developers.
Tasks in the ESSnet
|