Chapter 10 Qualitative (Dummy) Variables

Quantitative variables are easy to model and interpret because they take on numerical values that are readily dealt with by computers. Qualitative variables, however, are variables that do not naturally deliver numerical values. In other words, they are more like categories. Examples of qualitative variables are:

• Gender (male, female)

• Marital status (yes, no)

• Ethnicity (white, Hispanic, Asian, etc.)

Qualitative variables are made operational for regression analysis by creating dummy variables. A dummy variable can only take on two values (i.e., 0 or 1) and should be thought of as a switch.

• 1 implies the switch is on, meaning that the designated trait is present for an individual observation.

• 0 implies the switch is off, meaning that the trait is absent for an individual observation.

We can consider two different types of dummy variables depending on if we model the presence or absence of a trait to impact the intercept of the model or the relationship (or slope) between the dependent variable and other independent variables. We will cover these in turn.