A SAMPLING THEOREM FOR FINITE DISCRETE DISTRIBUTIONS Available to Purchase

There are many phenomena in information science which, when quantified, yield variants of the hyperbolic frequency distribution and are characterized by long straggling tails. Conventional statistical sampling theory offers no general techniques for the confident analysis of finite discrete distributions of this kind. A sampling theorem, based on the binomial probability distribution, but otherwise distribution‐free, is derived and some of its applications are illustrated numerically. One application answers Kendall's question: How many additional periodicals would be found on doubling the period of publication within which periodicals contributing to a specified topic are enumerated? The effects of sampling truncated distributions are also briefly discussed.