All Count Distributions are Not Alike

Abstract
Although the Poisson distribution is probably the most common distribution associated with count (or defects per item) data, it is by no means the only one. This article will present some situations which would lead to other distributions, in particular situations where: (1) defects occur in clusters; (2) defects are the result of two underlying causes; and where (3) probability of no defects is not related to the distribution of counts or the occurrence of no defects cannot be recorded.