From: Stan Hilliard
Date: 27 May 1999
Time: 16:05:08
Where did you find this rule. SQR(n)+1 is a rule of thumb that I use to obtain a reasonable number of cells in making a histogram.
I do not think it is appropriate to use this formula for the sample size of an acceptance sampling plan. I have a number of reasons for this opinion, including:
1) you would have different protection for different lot sizes (N). That is, the probability/confidence/risk of acceptance/rejection of a lot depend on N.
2) there is no one probability/confidence/risk statement that you can make that describes the performance of the rule. Thus, you cannot describe the decision-making aspects of the rule to the business decision-makers.
I recommend that you use a more sophisticated procedure that discriminates between good and bad lots. See:
www.samplingplans.com/modern3.htm
Stan Hilliard, shilliard@samplingplans.com