Alternative approaches to microdata protection Alternative approaches to

Alternative approaches to microdata protection • Alternative approaches to microdata protection: problem statement • The Scientific Use Files (SUF) are often not detailed enough for some specific research projects • Researchers want to work with original data (secure use files) • the individual output checking is too much workload the (Eurostat) staff 1

Alternative approaches to microdata protection • standard SUF will never completely fit for specific projects, no matter if they are anonymized with traditional or with perturbation methods or if it's synthetic data • with Synthetic data it is very elaborative to produce a "secure" SUF with good quality • for most of the projects the SUF is an adequate way of access • for specific research projects the SUF would be a dead end • we can’t produce tailor-made SUFs for specific projects with all MS 2

Alternative approaches to microdata protection • the all-purpose solution for SUFs is very important • this will not work for all projects • there will always be projects that still need original microdata. • we can also not proceed with SUFs forever as the only way of access to European microdata nowadays • researchers will keep demanding more detailed microdata 3

Alternative approaches to microdata protection • my suggestion would be that researchers are completely responsible for the safe output. The will get rules and support (like the guidelines on output checking) to produce a final secure output • only final !!! output will be checked if the rules are applied and if the output is secure. If not, the output will be rejected • Suppression or modification of values will not be done by Eurostat staff 4

Alternative approaches to microdata protection • to start this you could carefully open the safe centre at Eurostat for a "pilot dataset" like EU-SILC or LFS • the burden for researchers is still quite high to travel to Luxembourg • it might be that there are not too many requests • BUT, basically there is a mode of access to secure use files of European microdata • this step is important to get experiences with the process of output checking (amount, workload and researcher education) 5