Eurostat disseminates a high number of statistical tables. All these tabular data are treated for primary and secondary confidentiality. Primary confidentiality concerns tabular cell data, the dissemination of which would allow the disclosure of attributes. The main reasons for primary confidentiality are 1) too few contributors in a cell 2) the dominance by one or two contributors in a cell . Secondary confidentiality refers to some safe tabular cell data to which statistical disclosure methods need to be applied to prevent the disclosure of primary confidential cells through recalculation (based on the additive nature of the tables or the hierarchical structure of the data).
In the tables disseminated on the Eurostat website, a cell is confidential if it is primary or secondary confidential. A cell is primary confidential if
- the extrapolated number of contributors is positive but less than a specified number that is referred to as the "minimum threshold". In other words, a minimum number of contributors in a cell is required for the data in that cell to be published;
- a well-informed intruder, i.e., the second largest contributor, can estimate the value of the largest contributor within a chosen percentage, p%, of the true value of the largest contributor.
A cell is secondary confidential if it can be used with marginal totals to disclose primary confidential cells.
A confidential value is replaced with ":c" in the disseminated tables.
For non-confidential cells:
- until IFS 2020, the extrapolated number of holdings and all extrapolated values of the cell variables in cells are rounded to the closest multiple of 10.
- starting with IFS 2023, the extrapolated number of holdings and all values of variables in cells falling in interval [1,7] are rounded to 5, while if they are at least 8 they are rounded to the nearest multiple of 10.
Because of the confidentiality treatment, the sum of the individual cells does not systematically match with the value of the "total" cell.