1. Science
  2. Publications
  3. Information Processing Systems
  4. 3(110)'2013
  5. SPECIFICITY OF STATISTICAL DESIGN FOR PROFICIENCY TESTING RESULTS IN ACCORDANCE WITH IUPAC INTERNATIONAL HARMONIZED PROTOCOL

SPECIFICITY OF STATISTICAL DESIGN FOR PROFICIENCY TESTING RESULTS IN ACCORDANCE WITH IUPAC INTERNATIONAL HARMONIZED PROTOCOL

N.Y. Efremova
Annotations languages:

The paper presents the key features and benefits of the statistical design for proficiency testing results and homogeneity checks of samples which is proposed by IUPAC International Harmonized Protocol against the statistical design provided by the international standard ISO 13528. The paper describes an example, in which the assigned value and its uncertainty for proficiency testing are determined as consensus values from real measurement results by making a kernel density estimate of the distribution and bootstrapping the data, recommended by the IUPAC Protocol. In the paper the point of view, proposed in the IUPAC Protocol to the measurement uncertainties represented by the proficiency testing participants and the possibility of taking into account of their values in statistical design for proficiency testing results is discussed.
Keywords: proficiency testing, statistical design, performance statistics, bootstrapping, kernel density, assigned value, measurement uncertainty