Background
Microdata are in principle confidential, and can be shared only inside the European Statistical System for the production of European Statistics and can be shared under certain conditions with research organisations for research purposes.
Public use files, this is microdata files that can be shared with the general public, were first developed at the European level by the Centre of Excellence on Statistical Disclosure Control, before the launch of the DIGICOM project. These microdata are protected in such a way that no longer information on individual persons can be derived. This goes with an considerable loss in the analytical value of the data.
Methodologically, two different protection approaches were applied to the two most frequently demanded microdata sets:
- the labour force survey (EU-LFS)
where data were treated with a combination of traditional protection methods, notably through deleting or aggregating variables - statistics on income and living conditions (EU-SILC)
where data were constructed synthetically with a model, using the original distributions as an input.
Shortly after the ESSC endorsement of the DIGICOM business case, a proposed strategy for ESS public use files (PUFs) was discussed by the 7th meeting of the Working Group on Statistical Confidentiality. In its conclusions, the group foresaw further involvement of users in the evaluation of the usefulness of the PUFs in the framework of the DIGICOM project.
Two use cases were envisaged from the beginning:
- Use in education;
- Use by researchers preparing for access to scientific use files (these are partially-anonymised files, prepared for scientific use).
In a questionnaire to users of PUFs, a significant part of the registered users were actually colleagues from the European Statistical System, curious to see the files. The other users were researchers, teachers and students. The survey confirmed the usefulness of the files for education purposes.