A new role of proficiency testing in nuclear analytical work

The most recent definition of measurement result requires a statement of uncertainty whenever results obtained by nuclear or other quantitative methods of analysis are reported. Proficiency testing (PT) therefore must include the ability of laboratories to present not only unbiased quantity values, but reliable estimates of their uncertainty Hence, a reference value with the smallest possible uncertainty is needed to ascertain the proficiency of laboratories reporting results with lower than average uncertainty. A strategy based on the T-statistic is proposed leading to an accepted reference value that fully reflects the uncertainties reported by participants in a PT-scheme and permits calculation of En numbers to distinguish whether or not measurement results are consistent with the accepted definition of the measurand. The strategy is applied to PT-data from a recent international laboratory intercomparison of uranium isotopic ratios with very low reported uncertainties. In the paper this example is used to present the situation in the nuclear field.