Assessing the Role of Distributions and Control Charts in Metrology

Abstract

In this paper we suggest a radical change in the existing metrology paradigm for measurement data analysis, namely a rejection of searching the appropriate Distribution Function (DF) in favor of using nonparametric methods for uncertainty estimation in all circumstances. This way we can avoid the limiting and improbable assumptions that often lead to invalid conclusions. We also propose the use of Shewhart Control Charts (ShCC) as an indispensable method to assess and monitor the current state of a measurement system. This approach has many advantages and does not require any assumptions about specific type of DF.