ABSTRACT
One of the true powers of the linear model framework is that predictors can be both continuous (such as pitch, which is measured on the continuous scale of Hertz) or categorical (such as active versus passive voice, or male versus female sex). This chapter introduces the reader to how categorical predictors can be incorporated a linear model analysis. Two “coding schemes” for categorical predictors are discussed in this chapter, including “treatment coding” (R’s default) and “sum-coding”. In a hands-on example, the reader is guided through an analysis of the emotional properties of taste and smell words (such as sweet, bitter, pungent, and fragrant).
