Re: Need help with deciding sample n sub sample. size for nested design(DOE)

From: Richard Ulrich (Rich.Ulrich_at_comcast.net)
Date: 10/25/04 

Date: Sun, 24 Oct 2004 23:36:58 -0400

This is an under-parameterized problem, with additional
confusion of terminology.

On 23 Oct 2004 08:52:08 -0700, shankardasm@bluebottle.com (shanky)
wrote:

> Hi Guys:
>
> I am doing a nested design experiment. I know the amount it is going
> to cost me to have 1 subsample within the sample tested(example $5).
> Also I know the total amount i have for conducting tests(example
> $1000).

Let us say -- There are 2 suppliers. We can test k devices
once each (at $5), or we can do extra tests on a device
(nesting) at a cheaper rate. (If this is also at $5, then WHAT
reason exists for doing multiple tests?)

What reason exists for doing multiple tests, anyway?

>
> The criteria for making the decision is variablity should not be more
> than +/- 0.2 units. There are 2 suppliers. I want to decide upon the
> supplier.

What is varying? What is being tested?
Nothing was said about any accuracy of a single test.
SO there is no way state any numerical answer.

>
> I would like to know how should i calculate the number of samples from
> each supplier and the subsamples within each sample for each supplier.
>
> I just know that Supplier* Sample*Sub sample= (1000/5).
>

Hope this helps.

http://www.pitt.edu/~wpilib/index.html

Relevant Pages
Quantcast