Dixon et al. provide important direction to the psychological literature on improving intergroup relations by shifting the research goal from “prejudice reduction” to “social change.” The authors review the history of prejudice conceived as an individual's negative attitude or emotion and argue that a model of collective action is preferable to one of individual prejudice reduction for achieving intergroup equity and justice. However, their proposed model of collective action is also founded on ideas about the primary role of individual attitudes and emotions, “including anger and a sense of relative deprivation” (sect. 3.1, para. 11). In this way, Dixon et al. overlook an equally strong historical idea within psychology, which is that social change and the achievement of intergroup equity and justice will not result from a bottom-up change in attitudes. In the words of Thomas Pettigrew (Reference Pettigrew1991), theorizing social change as a project in changing individual attitudes is a “reductionist view that [denies] social and structural factors” (p. 10).

What are the social and structural factors that are useful for a psychological model of social change? Prior to Allport's (Reference Allport1954) seminal work on intergroup relations and prejudice, psychologists theorized that intergroup social norms are critical determinants of intergroup behavior. A social norm refers to group members' perceptions of the group consensus regarding the typical or appropriate treatment of another group. An early and continuing tradition of research in psychology conceptualizes prejudice as a perceived social norm, rather than as an individual's personal attitude or emotion (e.g., Crandall & Stangor Reference Crandall, Stangor, Dovidio, Glick and Rudman2005). An individual's attitude may interact with a social norm; for example, when personal attitudes are aligned with the perceived norm, the norm gives individuals social permission to act. When the attitudes are not aligned, the perceived norm encourages repression of attitudinal expression. Including social norms in a model of social change is critical because behavioral conformity to the perceived group consensus is a normal, universal process (Crandall & Stangor Reference Crandall, Stangor, Dovidio, Glick and Rudman2005).

Indeed, research has consistently suggested that perceptions of social norms regarding the appropriate treatment of an outgroup are more powerful predictors of intergroup behavior than individual attitudes toward that group (e.g., Blanchard et al. Reference Blanchard, Crandall, Brigham and Vaughn1994; Paluck Reference Paluck2009a). Furthermore, norms are important psychological constructs for investigators interested in the mobilization of collective behavior (e.g., Latané Reference Latané1980) because they define the standards of behavior for a group, not just for an individual (Paluck & Shepherd, under review). This research suggests that to change intergroup relations, the critical target is not what an individual personally thinks or feels about another group, but rather what he or she perceives other members of his or her group think or feel. Examples of social norms interventions, or where these new social norms can originate, include media messages depicting a social consensus of nondiscrimination or referent group leaders announcing support for stigmatized group members.

Psychologists have also theorized the role of structural factors for improving intergroup relations. Some psychologists have even argued that the top-down process of change starting with institutional change has been the most successful method for promoting intergroup justice and equity (Pettigrew Reference Pettigrew1991). This model of change starts with formal regulations and other types of institutional arrangements that channel behavioral action (Lewin Reference Lewin and Cartwright1951) created by nations, organizations, or communities, which affect collective patterns of behavior and perceived social norms, and finally individual level variables like attitudes and emotions. Top-down theories were cited in response to arguments that an authoritarian personality was responsible for racism and discrimination. Theorists argued that “direct structural change to which individuals (even authoritarian personalities) must accommodate is both more practical and effective … individual personality and attitude changes remain important for the successful completion of the change cycle. But they are usually not initially causal” (Pettigrew Reference Pettigrew1991, pp. 10–11).

A recent empirical review of the organizational diversity literature, cited by Dixon et al., provides support for this argument. The review (Kalev et al. Reference Kalev, Dobbin and Kelly2006) suggests that an organization's institutional diversity initiatives, such as appointing managers responsible for maintaining diversity, are responsible for advancing minorities and women into higher and better paying positions, not individually directed attitude change efforts like diversity training. In contrast to attitudinal change, diversity in managerial positions is the kind of material and behavioral outcome to which Dixon et al are referring as the goal for psychological intergroup relations research.

Theories of individual attitudes and emotions can be fruitfully combined with theories of social norms and structural factors in intergroup relations research. For example, our research program has focused on measuring the success of interventions that target the collective norms of a group, and this research has shown that changes in perceived norms regarding harassment and bullying change behaviors in schools over the course of one year (Paluck & Shepherd, under review). Future research should investigate whether attitudes, which do not change in the first year, eventually follows these normative and behavioral changes.

Additionally, psychologists have identified a number of individual level attitudes and emotions that might speed the process of social change once institutions or social norms have started to shift. For example, a belief that institutional changes are inevitable and permanent weakens people's resistance to and increases support for the changes (Laurin et al. Reference Laurin, Kay and Fitzsimons2012). Additionally, disconfirming fears or anxieties about outgroup members is predicted to improve intergroup interaction in times of social transition (Pettigrew Reference Pettigrew1991). Making positive examples of outgroup members accessible through storytelling or through the mass media, as a form of extended intergroup contact, are two examples of such interventions (Cameron & Rutland Reference Cameron and Rutland2006; Kenrick & Paluck, in progress).

Without negating the importance of individual level variables like attitudes and emotions, research suggests that psychologists reweight the importance of social and structural factors in theories of social change. I applaud the authors for their timely review and provocative reformulation of critical questions about improving intergroup relations. It seems time for psychologists to revisit the historical debate of whether we as a discipline should re-weight bottom-up versus top-down models of social change, and of what role psychology can play in either effort.