NIST produces a number of food-matrix Standard Reference Materials. A problem that frequently occurs in their analysis is one of assessing compatibility of data obtained by a laboratory with the specified values and uncertainty estimates given in the certificate of such a reference material. This certificate can be interpreted as the parameter assignment for measurand's distribution whose random nature leads to larger uncertainties. In particular, the non-informative, probability matching priors and the default data-dependent priors for lab's bias are suggested. These prior distributions are compared via the frequentist coverage probability of corresponding credible intervals. Several statistical approaches to this problem are suggested and numerically implemented in the R language.
See the special publication listed in the major accomplishments section.
NIST Special Publication 260-181, 2014