Conclusion

The choice of an uncertainty assessment method should be guided by the user’s needs—for example, to meet the requirements of a standard—taking into account all available information: accuracy conditions, previous experimental results, expert opinions, prior information about the measurand, etc.

The Bayesian approach thus appears to be an option worth considering on par with traditional approaches to uncertainty propagation and distribution propagation. Indeed, this article has shown that the Bayesian approach can be applied to any “traditional” uncertainty assessment and that it generally yields the same results as the Monte Carlo method when non-informative priors are chosen.

The major contribution of Bayesian statistics to metrology is to treat the measurand and all influencing quantities as random variables to which a degree of belief...