Risks in a sausage conformity assessment due to measurement uncertainty, correlation and mass balance constraint

Abstract

A technique for evaluation of risks of false decisions in conformity assessment of the chemical composition of sausages was developed based on a multivariate Bayesian approach, taking into account measurement uncertainty, correlation and mass balance constraint of the regulated contents of the sausage components. As a case study, a dataset of test results of chemical compositions of the sausage “Braunschweigskaya” (measured contents of fat, protein, moisture and salt) was used for evaluations of the risks. A total global consumer's risk of 0.006 and a total global producer's risk of 0.017, characterizing the production process in general, were evaluated using Monte Carlo simulations. The difference in the risk values indicates a clear preference of the consumer's interests over the producer's interests. The total specific consumer's and producer's risks, related to a (specific) sausage batch, were evaluated using normal approximations in the Bayesian model. The risk values obtained were much more significant when measured contents approached their tolerance/specification limits or exceeded them. The codes, written in the R programming environment, for calculations of both the total global and the specific risks are provided as electronic supplementary material to this paper.