Estimation of reproducibility and measurement uncertainty of a viscosity test method from proficiency test data

Abstract

Evaluation of laboratory test results in a proficiency test should preferably be done against a predetermined standard deviation. This standard deviation may be derived from the reproducibility of a standardized test method. However, in practice, the respective test method may lack precision data. The often used alternative, the use of a characteristic function, of which the Horwitz function is the most well known, is only applicable for chemical parameters where the concentration of an analyte is measured. For physical parameters, like viscosity, the concentration-dependent functions cannot be used. In this article, an example is presented on the estimation of the reproducibility (reproducibility limit R) and the measurement uncertainty of a viscosity test method from historical proficiency test data.