ABSTRACT
This chapter discusses: how a qualitative variable can be represented as a predictor in a multiple linear regression model via indicator (dummy) variables; the characteristics of a commonly used coding scheme for indicator variables known as reference cell coding; the fact that the interpretation of regression coefficients for indicator variables depends on the particular coding scheme used although certain results from the multiple regression analysis are identical regardless of the prescribed coding scheme used for the indicator variables; a step-by-step approach for evaluating a multiple linear regression model with a two-way interaction between a qualitative and a continuous predictor, where the qualitative variable is represented by an indicator variable using reference cell coding and the continuous predictor has been centered; and SAS programs that implement this step-by-step approach.
