Q: Why is a standard capability analysis determined to be best represented by 30 pieces?
I have answered this question by explaining it best represents a normal distribution. But I wonder if this is traceable to an industry standard?
A: You are right that most people associate 30 pieces with the conventional quantity for performing a capability study. Although I don’t know the origin of this number, I can tell you the following:
- The number 30 has nothing to do with whether or not the population is normally distributed.
- In many applications, the number 30 is insufficient to properly model the process. For example, automotive industry standards published by the Automotive Industry Action Group (AIAG) in their statistical process control (SPC) and production part approval process (PPAP) documents define 100 pieces as the appropriate sample size for an initial capability study (based on 20 subgroups of five or 25 subgroups of four).
I hope you find this helpful.
Denis J. Devos, P.Eng
A Fellow of the American Society for Quality
Devos Associates Inc.
For more on this topic, please visit ASQ’s website.