The Blurry Boundary between Stereotyping and Evaluation in Implicit Cognition

Introduction Does the distinction between cognition and affect apply to implicit mental states? In the literature on intergroup relations, cognition and affect correspond to stereotypes— beliefs about social groups, or, more concretely, representations of traits associated with social groups (Stangor, 2009)—and prejudice—negative affective or evaluative responses toward others on the basis of social group members (McConahay & Hough, 1976; Dixon et al., 2012). While the distinction between explicit stereotypes and explicit prejudices is widely supported (Fiske & Linville, 1980; Martin & Halverson, 1981; Wol, 1982; Augoustinos & Innes, 1990; Duckitt, 1992; Dovidio et al., 1996; Fiske, 1998; Mackie & Smith, 1998; Fiske et al., 2002), its applicability to implicit mental states is controversial. Understanding the relationship between stereotypes and prejudice is important as it can help illuminate the nature of the mental representations underlying implicit mental states. While some theories propose that stereotyping and prejudice are functionally integrated (Greenwald et al., 2002; Gawronski & Bodenhausen, 2006, 2011), others propose that the cognitive and affective components of implicit states represent separate constructs—namely implicit stereotypes (IS) and implicit evaluations (IE)—which reflect different mental processes and neural systems (Amodio & Devine, 2006, 2009; Amodio & Ratner, 2011). That stereotyping and prejudice are at least conceptually distinct in implicit cognition has been clear ever since Anthony Greenwald and Mahzarin Banaji (1995) coined these terms. But the question of their functional integration or independence leads to a number of significant conclusions. Most centrally, functional integration supports a “onetype” model of implicit states, according to which the core construct of the field is implicit attitudes as such (Fazio, 1990; Strack & Deutsche, 2004; Gawronski & Bodenhausen, 2006, 2011; Petty, 2006). Functional independence, on the other hand, supports a “two-type” model, according to which there are two fundamental kinds of implicit mental states: ISs and IEs (Amodio & Devine, 2006, 2009; Forbes & Schmader, 2010; Amodio & Ratner, 2011). Considering the evidence for a one-type vs. two-type model bears upon a number of other important questions. First, understanding the relationship between stereotyping and prejudice in implicit mental states may help to improve the predictive validity of indirect measures of attitudes. The predictive validity of the most widely used indirect measure—the Implicit Association Test (IAT; Greenwald et al., 1998)—has been questioned, and one potential limitation of the standard race-evaluation IAT is that it does not distinguish the relative contributions to test performance of liking/disliking social groups from beliefs about social groups. Second, given the centrality of the