Open AccessThe Capability Index when Some Assumptions are not Satisfied: Analysis and Empirical Comparisons
Author(s) -
Pablo Fernández,
Juan Francisco Muñoz Rosas,
Encarnación Álvarez Verdejo
Publication year - 2020
Publication title -
estudios de economía aplicada
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.123
H-Index - 6
eISSN - 1697-5731
pISSN - 1133-3197
DOI - 10.25115/eae.v34i3.3064
Subject(s) - conventional pci , estimator , process capability index , process capability , confidence interval , statistics , normality , standard deviation , index (typography) , mathematics , sample (material) , computer science , econometrics , engineering , work in process , operations management , medicine , world wide web , chemistry , chromatography , psychiatry , myocardial infarction
The process capability index (PCI) evaluates the ability of a process to produce items with certain quality requirements. The PCI depends on the process standard deviation, which is usually unknown and estimated by using the sample standard deviation. The construction of confidence intervals for the PCI is also an important topic. The usual estimator of the PCI and its corresponding confidence interval are based on various assumptions, such as normality, the fact that the process is under control, or samples selected from infinite populations. The main aim of this paper is to investigate the empirical properties of estimators of the PCI, and analyze numerically the effect on confidence intervals when such assumptions are not satisfied, since these situations may arise in practice.