Statistical Estimation of Degrees of Equivalence for Key Comparison experiments

Develop Bayesian statistical methods for the estimation of Degrees of Equivalence between laboratories participating in Key Comparison experiments. Degrees of Equivalence are measures of agreement between a single laboratory and the Key Comparison Reference Value, and also measures of agreement between pairs of laboratories.

The Mutual Recognition Agreement (MRA) requires the calculation of Degrees of Equivalence and of their uncertainty. The unilateral Degrees of Equivalence (DoA) are defined as the difference between the Key Comparison Reference Value and the result of a particular laboratory. For many experiments, this definition does not have a clear statistical interpretation making the uncertainty evaluation difficult to formulate in terms of a statistical model. This work starts with a statistical model of the experiment and develops measures of equivalence based on the model. These then lead to estimates and corresponding uncertainties which are then related to the DoA.

• Collaboration with scientists both within NIST, and internationally, on Key Comparison data analysis.
• Publication of several articles in key metrological journals.
• Presentations at international metrology conferences and workshops.