00405 9aa7a79db8fc5ef26a978b643a86f73

409

Process Capability: Engineering and Statistical Issues

capability index over another. Each newly proposed index is based upon some perceived shortcoming of one of the existing indices. Unfortunately, focusing on the issue of one process capability index versus another tends to overshadow and sidetrack the far morę important issues. For example, are these process capability indices really needed? Do they provide the proper information for decision making? Have the data been collected properly and have all required assumptions been verified?

In this paper, we urge that practitioners exercise restraint in rushing to calculate process capability indices. We suggest that they first focus on the larger and morę important issue of establishing and stating clearly the reason for these calculations. Then, based on these reasons, they establish an appropriate data collection scheme. Only after successfully conducting a complete process capability study should a process capability index be considered. We take the position that an index should only be called a process capability index if it is constructed from data collected in a legitimate process capability study, and if the resulting data have been appropriately analyzed.

Below, we present a review of the commonly used indices along with the reąuired assumptions and discuss how they can be verified. We also discuss some of the important issues conceming how the data for a process capability study should be collected. We discuss some of the fundamental issues needed for gaging a process' capability. As advocates of total quality, we realize that we may raise morę questions than we answer. However, these questions need to be resolved, possibly through the development of generał methodologies, procedures and standards.

Finally, we recognize that the exhortation Just Say "No!" to process capability indices is not sufficient. Hence, we provide a check list to be used prior to the calculation of any such index. We also urge practitioners to adopt statistically valid methods and techniques such as process capability studies, to measure a process' capability so that proven results through total quality can be achieved.

Two quality handbooks, the Western Electric (1956) Statistical Quality Control Handbook and Juran and Gryna's (1988) Quality Control Handbook provide substantial commentary and directions for assessing the capability of a process, and we will reference these extensively. A cursory review of text books on quality indicated that this topie was not typically treated in sufficient depth for its application. While we are aware that many companies have created their own manuals for quality, and specifically for process capability, we will not make specific reference to any of these.

Furthermore, when we initiated this effort, it quickly became apparent to us that there was an obvious need for a standard in this area. At that time,