GUM-based Decisional Criteria to Make Decisions in Presence of Measurement Uncertainty

This paper deals with the issue of decision-making involving measurement data. Often, measurement results must be put in comparison with limits so to take a decision on specification conformity or limit overcoming. This task is made complex due to measurement uncertainty. In fact, a measurement result is represented by an interval of values that could reasonably be attributed to the measurand. Additional information concerning probability density function and level of confidence may be available. As a consequence, the comparison concerns a limit and an interval of values having a probability distribution. This becomes more difficult when the limit also comes from a measurement, so even it is affected by uncertainty. In this case, two intervals of values have to be compared. The author aims to propose simple and intuitive decisional criteria to guide user during the decision-making process. Three main cases have been faced: the case of comparison with a simple limit; the case of conformity assessment with a conformance interval; and the case of comparison between two measurement results. The suggested criteria are based on the theory of the uncertainty evaluation reported in the Guide to the Expression of Uncertainty in Measurement. In detail, since a measurement result can be described by a statistical distribution, this information is used to compute the probability of specification conformity or the probability of limit overcoming. Different probability density functions have been considered. The final aim of this work is to provide a full overview on decisional problems where measurement results are compared in order to overcome the current lack of the state of the art. At the same time, the author wants to provide intuitive and easy to use decisional criteria to encourage their use even by users not experienced with decision-making and with the effects of uncertainty on the decisional process.