Measurement Uncertainty in Decision Making

Because all measurements have uncertainty, any measurement used to make a pass or fail decision, whether during R&D, manufacturing, calibration, or operation of a product, has some probability of being incorrect. Formulas for false accept and false reject probabilities in the context of manufacturing are discussed, including their relationship to the simplified but commonly used Test Uncertainty Ratio metric for assessing the adequacy of a measurement for making a particular decision. Methods of guardbanding a measurement to reduce the false accept risk are also given for both symmetric and asymmetric tolerances. Finally, a discussion of decision risk and decision rules in a calibration lab setting are provided.