Categorical Data Analysis Using The SAS® System, 2nd Edition

12.1. Introduction

Categorical data often appear as discrete counts that are considered to be distributed as Poisson. Examples include colony counts for bacteria or viruses, accidents, equipment failures, insurance claims, and incidence of disease. Interest often lies in estimating a rate or incidence (bacteria counts per unit volume or cancer deaths per person-months of exposure to a carcinogen) and determining its relationship to a set of explanatory variables. Poisson regression became popularized as an analysis method in the 1970s and 1980s (Frome, Kutner, and Beauchamp 1973; Charnes, Frome, and Yu 1976; and Frome 1981), although Cochran pointed out the possibilities in a 1940 paper (Cochran 1940), along with the suggestion of the appropriateness ...