Risk of erroneously deciding conformity of measuring instruments

Due to the unavoidable measurement uncertainty associated with every measurement and calibration result, conformity testing always implies the risk of taking erroneous decisions. Whereas end-users of measuring instruments are concerned mainly with instruments that have successfully passed a conformity test but actually indicate outside their specifications, instrument manufacturers are interested primarily in a low fail-error probability. In this paper, a straightforward Bayesian approach to evaluating pass- and fail-error probabilities in deciding conformity of measuring instruments is presented. It is based mainly on the expression of both knowledge of the respective calibration result of a measurement and the distribution of error values of the instrument population by means of appropriate probability distributions. The risk-calculating approach is explained using the example of conformity testing of a precision scale.