Preservation of confidentiality in small area statistics in Austria
Keywords
Principle 5, data dissemination, confidentiality, quality, grided statistics
Constrains on data dissemination mostly occur where data may be confidential or extra manpower is needed to prepare data according to customer wishes. Generally, as stressed in its mission statement, Statistic Austria’s tries to provide data information in an easily accessible and comprehensible manner for further processing, following scientific principles and using various communication channels.
Having customer wishes in mind, who need tailor-made information or to whom place matters the standard regional units for dissemination are not always detailed enough. Yet, the highest responsibility must be taken to prevent the identification of individuals (persons, business enterprises etc.). The solution was found in offering the service of customer defined polygons and data on the basis of grids.
Grid statistics and the service of customer defined polygons require manpower; therefore, they are commercial products and have to be paid for. As these products contain a lot of information on the location of the statistical data, the data portfolio is restricted, and customers of grid statistics and customer defined polygons have to sign specific terms of use. With signing the terms of use customers agree to avoid misuse, updates are regulated and the usage rights defined for the different uses, internal use and commercial use.
In addition these small area products have to be evaluated for quality and confidentiality to prevent the identification of individuals. This is taken care for by applying the method of target record swapping for some of the sensitive variables in the base data. In addition, confidentiality thresholds apply based on the absolute count (e.g. of population, buildings, dwellings etc.) in each grid cell and customer defined polygon respectively.