How does the census work? Confidentiality during the census

When publishing the results of the census, those responsible must ensure that the information provided by individual persons or other entities cannot be inferred, in order to safeguard each citizen’s fundamental right to informational self-determination and to comply with the Federal Statistics Act (BStatG ). At the same time, the published data should provide as much general information as possible. This is ensured by maintaining an appropriate level of confidentiality.

Various methods are used to ensure confidentiality in a census whenever evaluations are based on demographic data, building and housing data, household data or family data. No confidentiality procedure is required for data that is extrapolated from a sample survey (e.g. employment, education and training data), as the data is extrapolated and subsequently rounded to a multiple of 10 – or blocked if the number of cases is too low – to immediately prevent any information from being inferred about individual members of the population. The unaltered original population figures are shown for all administrative territorial units (without a demographic breakdown).

In the 2011 census, statistical confidentiality was ensured using the “SAFE” method for the secure anonymisation of individual data. However, the Statistical Offices of the Federation and the Länder have decided to change the method used for the 2022 census, where statistical confidentiality will be ensured using a stochastic (i.e. random) overlap based on the “cell key method”.

The cell key method

The SAFE and cell key methods are both based on the same principle: Some case figures shown in the tables are slightly altered from their original values. In contrast to the SAFE method, however, individual data is not altered in the cell key method, but rather slight changes to the original result are only made when the results are generated. During this process, some results or table fields (= cells) are permanently assigned a small “overlapping value” (e.g. +2 or -1). Instead of the original result being shown for those cells, the sum of the original result and the “overlapping value” is published. The results are overlapped in this way on the condition that individual information is protected and the meaningfulness of the original results is preserved as much as possible. In comparison with the SAFE method, the results altered using the cell key method generally deviate less from their original values and the definitive data can be evaluated with even more flexibility.

Application of the cell key method to the results of the 2011 census

As part of the relaunch of the 2011 census database, the cell key method was retrospectively applied to the results of the 2011 census. This will allow comparisons to be drawn between the census results from 2011 and those from the 2022 census, which will be generated using the exact same confidentiality method (i.e. with almost no differences in the methodology).

Some of the results available in the 2011 census database may differ slightly from the previous publications for the 2011 census due to the change in procedure. However, the change in the method used to ensure confidentiality does not compromise the quality of previous publications or the quality of the results available in the database.