From: Stan Hilliard
Date: 07 May 2003
Time: 12:29:41
1) The method of making the oc curve for an attribute plan with C=0 is the same as it is for any other value of C. Use the binomial distribution or the poisson approximation to the binomial.
2) The binomial and poisson are used to calculate the oc curves in Mil-Std-105E, so they would be the same.
3) You can test the accuracy of any oc curve by simulating its decisions with random numbers. You could use MS Excel to do the simulation. 5,000 decisions at each point should give an accurate measure of the oc curve's reliability.
In general, think that C=0 sampling plans sound better than they actually are. To see this for a particular plan, look at the consumer's point rather than the producer's point.