Diagnosis of special causes is one of the primary concerns of quality practitioners. Identification of special causes can be greatly aided by knowledge of when and in what manner the process change occurred. Such a task can be especially complicated when the special cause affects only a portion of a subgroup. This paper will discuss the behavior of x ̄-R control charts in such situations, and presents two methods for identifying change point and structure. Monte Carlo results show that the simpler of the two methods correctly identifies change structure 81% of the time, over the range of changes considered. A more complex method yields 62% correct identification of charge structure, but also gives an estimate of the charge point and is more robust.
ASJC Scopus subject areas
- Computer Science(all)