From: TYC
Date: 1/22/2006
Time: 8:56:34 PM
My supplier presented the following formula for determining the sample size recently. The aim is to detect a 500 PPM problem in their daiy production of 400 units.
Method: Simple random sampling without replacement
sample size = [N*p*(1-p)]/[(N-1)*D^2 + p*(1-p)]
where N=400, p= 500 ppm, D = not too sure but I think is is something associated with 99.5% confidence interval.
The result sample size is 8!!! Which I find highly unbelieveable.
Can some help to shed some light on this?
Alternatively, can someone advise how we can determine an appropriate sample size in this case?
Thanks in advance.