00428 R6477e3a6b2ece2e9df3395dc73b468

432

Pignatiello & Ramberg

wise to follow it with a formal process capability study for systematic evaluation.

In addition to the four types of studies and programs elaborated by Mentch, many customers require other types of studies to document ‘fcapability.” Depending upon the quantity of products produced, these studies can require anywhere from a smali proportion of production to 100% evaluation. Typically these customer-driven requirements specify a minimum number of data points to be reported and a reporting period.

Presumably these requirements are for the purpose of anticipating futurę problems, as well as for providing benchmarks for continuous improvement programs. Except when a process capability study is conducted, it is not elear how to estimate the variability of these estimates. The problem is that without knowing the variability of such estimates, decisions based on them may simply be due to the innate variability present in them. Hence, these calculated indices should not be called capability indices.

Conclusions

In this paper we have stated some important issues regarding process capability indices and process performance indices. We noted the existence of an emerging ANSI-Z1 standard in this area. We discussed reasons why process capability indices are used and have given some of the morę commonly used indices. We have also discussed the statistical assumptions that are required for using these indices and have provided some methods for validating these assumptions. We have also indicated how very capable processes can yield estimates of process capability indices that are quite Iow. We have also discussed the veiy critical issue of how data is collected for process capability studies.

We believe that process capability and process performance are important and critical matters that cannot be simply summarized by blindly calculating an index. To that practice, we Just Say "No!" and urge practitioners to view process capability and process performance assessments as the complex issues they are.

References

Bigelow, J. S. (1992). Standard Method for Calculating Process Capability and Performance Measures, Drafl 0.3, ANSI-Z1 Committee on Quality Assurance, c/o James S. Bigelow (Chair), Polymers Group - Americas, 13501 Kąty Freeway, Houston, Texas 77079-1398.