The beta distribution is assumed to govern the selection of M
unobservable probabilities p/sub k/(K = 1, 2,..., M). M binomial experiments of
n trials each are performed with the probability of a success on any trial on the
k/sup th/ experiment being p/sub k/. from the observed results of these
experiments, a 100 gamma -percent confidence region is derived for the
parameter of the underlying
... [Show full abstract] beta distribution using a normal approrimation. An
example is included which applies the model to lot acceptance sampling by
attributes. (auth)