Comparison of GUM Supplement 1 and Bayesian analysis using a simple linear calibration model

Abstract

Recent work referred to two approaches for doing a Bayesian analysis for simple linear calibration and pointed out that there could be a difference between the results of applying the procedure of GUM Supplement 1 and one of those approaches. It will be shown that the difference between the two Bayesian approaches reflects the use of two different priors. It will be shown that the results obtained by GUM Supplement 1 are those of a Bayesian analysis with commonly used priors and a measurement equation that satisfies the principles of the Guide to the Expression of Uncertainty in Measurement (GUM).