Content
Overview
Key comparisons are interlaboratory comparisons carried out regularly between National Metrology Institutes (NMIs) within the framework of the CIPM Mutual Recognition Arrangement (MRA). Key comparisons enable the mutual recognition of calibrations, measurements, and test certificates of the NMIs and mark a major step in supporting international trade, commerce and regulatory affairs. The final report and the supporting technical data of each key comparison are stored and made publicly available at the key comparison data base KCDB of the Bureau International des Poids et Mesures (BIPM). Fig. 1 shows a typical example of key comparison data.
Research
The analysis of key comparisons can be seen as a Meta-Analysis in which the results reported by the participating laboratories are assessed. Fixed effects or random effects models are popular models applied in meta-analysis. One focus of research of PTB’s Working Group 8.42 is the development of Bayesian methods for the estimation of fixed and random effects models. This includes the assignment of vaguely informative or appropriate non-informative priors and the exploration of the properties of the resulting inference. Other topics of research are situations in which the common measurand shows a drift, or where the explanatory power of key comparisons is assessed in dependence on the stability of the common measurand and the uncertainties reported by the participating laboratories.
Software
Publications
Publication single view
Article
Title: | Analysis of key comparison data: assessment of current methods for determining a reference value |
---|---|
Author(s): | C. Elster and A. Link |
Journal: | Measurement Science and Technology |
Year: | 2001 |
Volume: | 12 |
Issue: | 9 |
Pages: | 1431 |
DOI: | 10.1088/0957-0233/12/9/308 |
Web URL: | http://stacks.iop.org/0957-0233/12/i=9/a=308 |
Tags: | 8.42,KC |
Abstract: | The degree of equivalence of national measurement standards is established by means of key comparisons. The analysis of data from a key comparison requires the determination of a reference value, which is then used to express the degree of equivalence of the national measurement standards. Several methods for determining a reference value are available and these methods can lead to different results. In this study current methods for determining a reference value are compared. In order to quantitatively assess the quality of performance, the methods are applied to a large set of simulated key comparison data. The simulations refer to several realistic scenarios, including correlated measurements. Large differences in the results can occur and none of the methods performs best in every situation. We give some guidance for selecting an appropriate method when assumptions about the reliability of quoted uncertainties can be made. |