ABSTRACT
This chapter discusses the ways to estimate regression models when one has dichotomous dependent variables (also called binary, or dummy variables). Models for dichotomous dependent variables are used to estimate the probability of the occurrence of an event. It is also important to understand the distinction between probability and likelihood. A probability is estimated from a “population” of which one knows its “parameters”, while a likelihood estimates the values of the parameters for which the observed result best fits them. Logistic models have grown tremendously in popularity in Political Science. The Logit and Probit models differ in their link functions. Logit bears its name because the function is given by the natural logarithm of the odds. Logit models, on the other hand, are interactive by nature: the effect of an independent variable on the probability of occurrence of the dependent variable, depends on the values of the other covariates of the model.
