Fractional nonconformance based conformity testing

Abstract

Conformity testing, also known as evaluation of conformity or compliance testing, is exercised to assure that an entity meets a specific requirement and/or regulatory standard. Measurement and sampling uncertainties, including metrological traceability, become crucial for the declaration of conformity, especially when the measured value is close to the set limiting value. The fractional nonconformance (FNC) approach was recently introduced to deal with measurement errors for acceptance sampling purposes. This approach can be extended to develop new conformity evaluation protocols. In this study, we develop a new fractional nonconformance based conformity testing procedure, and compare its performance with the conformity testing procedure prescribed in the ISO Standard 10576. The proposed procedure explicitly distinguishes the limiting value from a specification by linking the false alarm rate and the limiting value using a baseline distribution. Our research shows that the two-stage conformity testing procedure using the fractional nonconformance statistic reduces the probability of incorrect declaration of conformity or inconclusive result for nonconforming entities when the number of test samples is larger than one. This means that the new FNC based conformity testing is more powerful than the current procedure set by the ISO standard.