Nonparametric test on process capability

The study of process capability is very important in designing a new product or service and in the definition of purchase agreements. In general we can define capability as the ability of the process to produce conforming products or deliver conforming services. In the classical approach to the analysis of process capability, the assumption of normality is essential for the use of the indices and the interpretation of their values make sense but also to make inference on them. The present paper focuses on the two-sample testing problem where the capabilities of two processes are compared. The proposed solution is based on a nonparametric test. Hence the solution may be applied even if normality or other distributional assumptions are not true or not plausible and in the presence of ordered categorical variables. The good power behaviour and the main properties of the power function of the test are studied through Monte Carlo simulations.