4. Conclusion
The aim of quality assurance is to avoid generating outliers, and its development since the turn of the century is closely linked to lab- oratory accreditation according to ISO 17025 [10] – another aspect that Andersen fails to mention! The approach based on En numbers and the use of T-statistic is an optimal way of demonstrating the competence necessary for accreditation.
Failure to take into account relevant work in the field leads An- dersen to conclusions that cannot be upheld; at the same time, the current approach fulfils Andersen’s success criteria!
It is recommended not to exclude outliers, but to present all data as shown in Fig. 1. A consensus value should, however, be based only on mutually compatible data in statistical control.