Logistic Regression Using SAS®: Theory and Application

2.1. Introduction

A great many variables in the social sciences are dichotomous—employed vs. unemployed, married vs. unmarried, guilty vs. not guilty, voted vs. didn’t vote. It’s hardly surprising, then, that social scientists frequently want to estimate regression models in which the dependent variable is a dichotomy. Nowadays, most researchers are aware that there’s something wrong with using ordinary linear regression for a dichotomous dependent variable, and that it’s better to use logit or probit regression. But many of them don’t know what it is about linear regression that makes dichotomous variables problematic, and they may have only a vague notion of why other methods are superior.

In this chapter, we focus on logit analysis (a.k.a. ...