Logistic Regression Using SAS, 2nd Edition

Chapter 9

Regression for Count Data

9.1 Introduction

9.2 The Poisson Regression Model

9.3 Scientific Productivity Example

9.4 Overdispersion

9.5 Negative Binomial Regression

9.6 Adjustment for Varying Time Spans

9.7 Zero-Inflated Models

 

9.1 Introduction

In this chapter, we examine Poisson regression and negative binomial regression, which are two methods that are appropriate for dependent variables that have only non-negative integer values: 0, 1, 2, 3, etc. Usually these numbers represent counts of something, like number of people in an organization, number of visits to a physician, or number of arrests in the past year. While such data are fairly common in the social sciences, there is another reason why Poisson regression is important: ...