Sample Size Required to Observe at Least k Rare Events

Abstract

Tables based upon Poisson distribution are presented which specify total sample size adequate to ensure ( p > .95) that a certain specified number of rare events will be observed. The tables are useful in planning sampling surveys where special interest is in obatining at least a specified number of cases of special research interest. The tables can also be used as basis for rejecting the hypothesis that occurrence rate for rate event exceeds some specified value in the population.