Abstract
Current methods of testing laboratories for their proficiency in reporting correct measurement results are liable to substantial errors of the 2nd kind. This means that laboratories with deflated uncertainties are accepted as proficient, even though their reported measurement results pave the way for erroneous conclusions. Only by using En numbers based on an accepted reference value with the lowest possible uncertainty can the risk of recognizing incorrect measurement results be kept at an acceptable level.
Based on an actual set of PT-data for the concentration of Pb in water the paper compares proficiency test results obtained by methods using En numbers with methods based on z-scores.
Original language | English |
---|---|
Journal | Accreditation and Quality Assurance |
Volume | 13 |
Issue number | 12 |
Pages (from-to) | 703-709 |
ISSN | 0949-1775 |
DOIs | |
Publication status | Published - 2008 |
Keywords
- En numbers
- Fiddle factor
- z-scores
- Proficiency testing
- Errors of the second kind
- Type II error