On the Use of Prior Distributions in Acceptance Sampling *

The use and interpretation of prior and posterior distributions in Bayesian analyses is discussed with particular reference to whether the prior represents personal probability or the distribution of a random variable. One situation in which the latter applies is acceptance sampling, the parameter of interest varying from lot to lot. Two criteria for assessing or determining a sampling plan, which incorporate the information represented by the prior distribution into the producer's and consumer's risks, are presented and illustrated.