Article ID: | iaor2000796 |
Country: | United Kingdom |
Volume: | 26 |
Issue: | 3 |
Start Page Number: | 425 |
End Page Number: | 435 |
Publication Date: | Jun 1998 |
Journal: | OMEGA |
Authors: | Prasad Sameer, Bramorski Tom |
Keywords: | statistics: general |
Many organizations are mandating the use of process capability indices to measure and reduce variability. Managers may find the use of such indices as unreliable when the interaction of the dependency structure and outliers mask potential assignable causes of variation. A monitoring procedure is proposed that effectively classifies a series' variability into three components: (1) the underlying correlation structure, (2) the outliers, and (3) the unexplained variance. Indices are proposed to distinguish these components. This approach can help a manager identify the appropriate corrective actions to be taken to reduce variability.