Combining inconsistent data from interlaboratory comparisons. http://dx.doi.org/10.1088/0026-1394/44/5/019

Abstract

The most widely used procedure for finding the value of a quantity from data obtained from interlaboratory comparisons involves calculating the inverse-variance weighted mean of the laboratories’ estimates. This method produces adequate results if the data are consistent. However, sometimes a consistency test reveals the possible existence of outliers that nevertheless have to be included in the evaluation task. In this paper the Bayesian understanding of probability is used to derive three methods that may be applied in the case when the data appear to be inconsistent. The performances of these methods are compared through a simulated example