Investigation of the Real Meaning of the Stability Index and Its Empirical Analysis

Abstract

Assessment of process stability is a key to ensuring the high quality of any product, service, activity, etc. The main tool for doing this is a Shewhart control chart. However, it becomes difficult to analyze hundreds or thousands of control charts at a time for some up-to-date complex processes. In order to solve this problem, various indicators of process stability have been recently proposed with the Stability Index (SI) being the latest one, and, in many opinions, the best one. We have thoroughly analyzed different suggestions and found many of them ambiguous and sometimes even harmful to practitioners. Some examples of possible problems with the SI are provided below. It is argued that abnormal values of the SI can be caused not only by process instability, but by non-homogeneity, non-randomness, nonnormality, and autocorrelation as well. It is proved that the SI can be helpful for practitioners as an index to attract primary attention to the process. A proposal to rename the SI and widen significantly its application is given. Finally, some ways that practitioners could use this index more effectively are discussed.