Linear Regression Model: Qualitative Variables as Predictors

In a regression model, qualitative variables are used as regressors to estimate influences of qualitative aspects on the response variable. Examples of qualitative variables include gender, religion, geographic region, caste, political affiliation, or marital status. Qualitative information are normally represented by categorical variables in the form of numerical codes without having any numerical meaning. The categorical variables can be quantified by artificially constructed binary form representing the presence or absence of an attribute. This artificially generated binary variable is called the dummy variable. The dummy variable also has no numerical meaning, and it is used to represent mutually exclusive categories. The interpretation of the coefficient of a dummy variable is different from that of a numerical continuous variable.