2. Bayesian Approach to the Full Statistical Model

The Bayesian approach to assessing measurement uncertainty is particularly well-suited when prior information about the measurand is available $Y$ (for example: the sign of the measured quantity) or when the relationship between quantities is represented by a statistical model.

Instead of being modeled to inform an input variable X of the measurement model in the Bayesian GUM, the data $d_1,\dots ,d_N$ are modeled in terms of their relationship to the measurand Y and the other input variables $Z\text{=}\left(Z_1,\dots ,Z_p\right)$...