Safe outputs discuss the rules that make an output unsafe, and the basics of how to protect outputs while maintaining usefulness.
We have designed a website designated to output checking including the rules that can make an output unsafe and how to protect outputs while maintaining usefulness: outputchecking.org
Examples of questions/issues being addressed:
- If the aim of access is to produce statistics, is there any residual risk by, for example, showing outliers?
- If the aim of the access is to produce data for onward transmission, how do we make sure that the released data are appropriate for the next use?
Possible ways to achieve safe outputs:
- Output checking (see outputchecking.org)
- Minimum group sizes required before releasing statistics
- Dynamic disclosure control, using tools like SACRO and sdcTable to automate checks.