A unified approach to uncertainty for quality improvement

To improve quality, process outputs must be measured. A measurement with no statement of its uncertainty gives no meaningful information. The Guide to the expression of Uncertainty in Measurement (GUM) aims to be a universal framework for uncertainty. However, to date, industry lacks such a common approach. A calibration certificate may state an Uncertainty or a Maximum Permissible Error. A gauge study gives the repeatability and reproducibility. Machines have an accuracy. Processes control aims to remove special cause variation and to monitor common cause variation. There are different names for comparable metrics and different methods to evaluate them. This leads to confusion. Small companies do not necessarily have experts able to implement all methods. This paper considers why multiple methods are currently used. It then gives a common language and approach for the use of uncertainty in all areas of manufacturing quality.