Introduction

Dummy variables (also known as binary, indicator, dichotomous, discrete, or categorical variables) are a way of incorporating qualitative information into regression analysis. Qualitative data, unlike continuous data, tell us simply whether the individual observation belongs to a particular category. We stress understanding dummy variables in this book because there are numerous social science applications in which dummy variables play an important role. For example, any regression analysis involving information such as race, marital status, political party, age group, or region of residence would use dummy variables. You are quite likely to encounter dummy variables in empirical papers and to use them in your own work.

This chapter first defines dummy variables, then examines them in a bivariate regression setting, and finally considers them in a multiple regression setting. We stress the interpretation of coefficient estimates in models using dummy variables; discussion of issues related to inference is deferred until the second part of this book.

Dummy variables are another way in which the flexibility of regression can be demonstrated.