As a rule, measurement standards are materialisations of a single property value, and interlaboratory comparison or proficiency testing schemes compare single property values obtained by different laboratories on the same reference object. The picture changes when more than one single value are to be determined for the reference object (e.g. a couple of analyte concentrations for one matrix), and the values are suspected to be correlated, and a real problem arises when functionally related properties are to be determined. In this paper, some aspects of the applicability of available multivariate data assessment techniques with respect to interlaboratory comparisons will be discussed.
展开▼
