Evaluation of Inter-Laboratory Comparison Results: Representative Examples

The primary purpose of inter-laboratory comparisons is to verify the CMC claims of the participating labora-tories. The most commonly used evaluation criterion, the normalized error |Eni|, has flaws, particularly if the comparison uncertainty ucomp,i is large relative to the reference standards being compared. According to the recommendations of CIPM Working Group for Fluid Flow, the uncertainty ucomp,i should be considered by including the uncertainty of the transfer standard and the repeatability of the calibrations. A review of previous comparison reports shows the importance of this task for many other measurands beside the field of fluid flow. In this paper, we propose the use of comparison uncertainty ucomp,i as a better tool for assessing the power of the comparison. We applied the comparison evaluation criteria to recent comparison results to illustrate their benefits over the |Eni|<= 1 criterion. A probability-based approach evaluates the comparison results and we recommend its usage in future comparisons.