On 11 Jan 2001 16:28:50 -0800, [EMAIL PROTECTED] (Eric Scharin)
wrote:
> All -
>
> I have a problem which I am hoping someone can help me solve. (No, this is
> not a problem from class, but a real-life problem from industry.)
>
> A product has a specification of <= 1 defect/10 cm^2. The product has an
> overall surface area of X cm^2. While this specification was not obtained
> based on an analysis of the alpha & beta risks for acceptance sampling, the
> specification itself implies an alpha & beta level. Is there a simple way
> to calculate how the alpha and beta risks would be affected by a change in
- is there a simple way to spell out the alpha and beta risks?
If you cut the sampling in half, then there *might* be no change in
something-or-other. Are you talking about a risk, per-test? or are
you trying to relate to a production line?
This seems to be a problem with 2 equations and 6 unknowns.
Or something similarly undefined.
> sample from 10 cm^2 to 5 cm^2? Specifically, if someone accidentally
> sampled 5 cm^2 and found 1 defect, what could you say about the status of
> the product?
I don't know. When I see one defect, what am I supposed to say about
the status of the product? "I see a defect!"
I think your real-life problem probably needs a real-life consultant.
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================