Sampling size conundrum

vipul — Sat, 10 Jun 2023 20:50:05 GMT

Hi All,

How can I prove that there is practically 0 defects (0.05 ppm), without taking millions of samples? I have a pass/fail test.

Has anyone else within this community faced a similar issue?

Thank you in advance!

Re: Sampling size conundrum

statman — Mon, 04 Apr 2022 13:13:51 GMT

Well that's a loaded question. First, you can't "prove" anything (perhaps you can change your degree of confidence in your conclusions). Second, how do you know the failure rate is .05ppm? How do you know your test is "good" (measurement error)? Do you know if the process is consistent?

You could take either a probabilistic approach or an analytical approach to building your confidence in the low failure rate. I usually don't care much about whether the rate is "correct" or not. I am usually applying statistical method to discover what factors affect the rate.

topic Sampling size conundrum in Discussions

Sampling size conundrum

Re: Sampling size conundrum