Categorical and Dummy Variables in Regression Models

Abstract

In the application of regression analysis there are many situations where either the dependent variable or one or more of the regressors are categorical variables. When one or more categorical variables are used as regressors, a financial modeler must understand how to code the data, test for the significance of the categorical variables, and, based on the coding, how to interpret the estimated parameters. When the dependent variable is a categorical variable, the model is a probability model. Keywords: Categorical variables; dummy variable; marginalization; Chow test; dependent variable; linear probability model; probit regression model; logit regression model; logistic distribution; Leverage World