This dissertation focuses on automatic race categorization (ARC), or the tendency for people to perceive others as falling into discrete racial groups rather than perceiving more continuous racial variation. Traditional research approaches treat race categorization as a binary process and assume that people perceive race as falling into discrete categories. In contrast, this research conceives of race categorization as a continuum where a person can perceive the boundary between races as strong and distinct, weak and non- existent, or anything in-between. First, I introduce a new method of measuring ARC (Sedlins, Malahy, Plaks, & Shoda, 2012). The data support the idea that people tend to see racial continua as falling into discrete categories (chapter 1: studies 1-3) and provide evidence that there are significant individual differences in how strongly people tend to perceive discrete race boundaries (chapter 1: study 1). The present research then examines several individual difference predictors of ARC strength: political ideology (Chapter 2: studies 1-2), beliefs about genetic variation (Chapter 3: studies 1-2), and multiracial salience (Chapter 4: studies 1-2). Finally, the present research provides evidence that strength of ARC predicts race bias. Those with strong ARC (i.e., perceiving races as highly discrete; strong race categorizers) show greater racial bias than weaker race categorizers (Chapter 3: study 2). Together these studies provide a new approach and method to research race categorization and suggest new ways to approach prejudice and discrimination in intergroup contexts. These findings are discussed in terms of their implications for psychological science and social policy.