In our Study on Cross-Border Data Collection, we use administrative data as a source. This kind of source allows us, in principle, to publish exact numbers of cases. However, one would like to avoid statistical disclosure, namely the identification of individuals from the results. Statisticians have developed several Statistical Disclosure Control methods to prevent this. These methods can be quite complex and, as a consequence, labour intensive. In addition, there is a risk that relevant information will be suppressed. In order to minimise the workload, we propose an alternative simple approach to deal with this issue.
In our project we ask for tables in which the exact numbers are rounded to tens. This evidently prevents any risk of disclosure. We also think that we will not lose any essential information. Since we aim to collect the numbers of cross-border workers, the unit of data collection is persons. An accuracy of tens of persons is already quite high. These numbers can be used to calculate shares or make aggregations with high reliability. In addition, less than 5 persons is a small number and can therefore be considered as neglectable. To further reduce the potential negative impact of rounding errors when adding up rounded numbers, we also ask data producers to provide numbers for the most commonly used and logical aggregates. This means, for instance, providing also the aggregates for NUTS2, NUTS1 and country level, in addition to the NUTS3 information, and asking for totals for the breakdowns.
Curious about our project?
The Cross-Border Data Collection Project is commissioned by DG REGIO and carried out by a consortium led by ÖIR GmbH, in collaboration with Statistics Netherlands, and supported by INSEE and Statistics Denmark. For more information on the project or our methods, please reach out to our team at crossborderdata@cbs.nl.
- Kérjük, jelentkezzen be, ha szeretne észrevételeket közzétenni.