Non-Normal Cpk - What do I do? (Poor 6 Sigma training)

[@Old.Vet.]

😀Thank you!

[@paulallen5321]

You're welcome...

[@gl4c.4]

What sample size would you use in a process step where descriptive data is only used? 1000 sample size minimum? What if you make only 50 products a year?

[@paulallen5321]

Hi Karol - then you have a problem. However if you see a defect in just 50 samples you really do have high defect rate, so act on that. But when the defects disappear it doen't mean you defect rate is zero, it just means you don't have enough data to know whats going on...sample sizes are really difficult to get around. It's not a made up theory, it;s a law of mathematics, best thing to do is concentrate on inputs and the process as well controlled as possible...