On multiple-method studies

摘要

In this paper we review statistical models that describe measurements from a multiple-method study such as in the development of a reference material. We also review requirements for the so-called GUM compliance, as this appears to be an important criterion for choosing a model and a method for assigning a value to a measurand of interest and calculating its uncertainty. For each modelling approach, we identify groups of competing methods and evaluate their status with respect to their statistical characteristics and GUM compliance.