Linear models become much more interesting once multiple predictors are included (“multiple regression”). This chapter provides an extensive example of a dataset on iconicity (the degree to which a word sounds like what it means, such as onomatopeic forms like “beeping” and “squealing”). Multiple predictors are included into a linear model to look at how iconicity is affected by various word-level properties. The chapter also clearly shows why it is important to standardize one’s variables if the comparison of the relative impact of different predictors is of interest. The chapter concludes with a discussion of linear modeling assumptions, including normality and homoscedasticity (“constant variance”/“equal variance”).