Appling for sensitive data
What are sensitive variables?
By default, the FDZ provides data sets as scientific use files (SUFs). In these data sets, all information has been removed from the original data that would allow a potential re-identification of individual persons, as well as information that is particularly worthy of protection. This concerns information on ethnic origin, political opinions, religious or ideological beliefs, health data or data on sexual life or sexual orientation. We do not pass on this information. You can recognise these variables by an addition in the variable label: "(zur Anonymisierung geleert (FDZ))".
For sensitive variables that can be coarsened in a meaningful way so that a re-identification of individual study participants can be excluded and at the same time the analysis potential of the data set can be increased, the FDZ creates additional variables. These contain the information from sensitive variables at a higher level of aggregation, so that they can be made available in the scientific use file. This means, for example, that if there are values on variables that come from five or fewer persons, we recode them into a new variable. Typically, these are variables such as country of origin, language spoken at home, occupation of the parents, diagnosis of special educational needs, etc. These variables can also be recognised by an addition in the variable label: "(zur Anonymisierung gruppiert (FDZ))".
In the blank data sets for our studies (to be found on the respective study pages), you can therefore already find out which variables are affected before applying for the data and, if necessary, be advised by us if you have any questions on this issue.
How can I access sensitive variables?
Sensitive variables can either be recoded by us according to your wishes, so that they no longer contain expressions of five or fewer persons, or they can be accessed by remote computing (see below).
In education monitoring studies, the federal state (Länder) variable is also a sensitive variable. According to our rules of procedure, it can be requested for the following purposes:
(a) as a covariate for control purposes only
b) for the purpose of adding contextual characteristics or other third-party variables
c) for comparisons between aggregated groups of federal states (Länder)
d) to describe the sample (e.g. distribution of participants across federal states (Länder) and across school types within federal states (Länder)).
If necessary, we create grouped variables for data users who want to work with federal state information, suitable for the respective question. You can describe the desired variable(s) directly in your application text. Please note that none of the groups may contain only one federal state.
Further information on special cases of the use of federal state variables can be found in the section Novel Länder comparisons (i.e., novel comparisons across German federal states [Länder]).
For analyses with sensitive data that cannot be made available in scientific use files (SUFs), data users can use the remote computing system JoSuA. This does not give direct access to the data, but allows to submit syntaxes via an online portal. Users will receive the output of the commands back. Please note that you are still not allowed to publish results for groups of five or less or individual ferderal states (Länder). Remote computing via JoSuA can be used by one person per project for 4 months (extension possible if required).
You can find more information on remote computing on our FAQ page.
Please describe in your application for data usage whether and which sensitive data you would like to analyse.