Self-categorization theory for beginners

Anti-immigration rhetoric, racism, and xenophobia have been increasingly widespread around the world over the last decade. People compartmentalize one another based on identity markers, such as gender or color. On the other hand, researchers have shown that people did not even need to be racist or xenophobic for racist and xenophobic outcomes to emerge in society (Axelrod, 1997; Salzarulo, 2006). What is more, history has shown us that education is not a cure for racism and xenophobia. As Milgram’s experiment revealed, people tend to obey authority and while following orders of a superior they might act against their own better judgment (Milgram, 1974). People’s reluctance to confront those who abuse power explains how the Holocaust was organized and carried out as ordinary German citizens obeying authority followed orders that they did not approve of (Billikopf Encina, 2014). How can we make sense of such phenomena as social scientists? Understanding how polarization happens and how it results in racial, religious, ethnic, and gender-based discrimination in society is fundamental to prevent and cope with identity-based acts of violence. In this essay, I will summarize a very complicated social psychology theory, self-categorization theory (SCT).

Self-categorization theory (SCT) assumes that identities are defined in society through interactions within and among groups. Identities prescribe how individuals should feel and behave. Individuals possess multiple identities, which get activated in different contexts. Identity activation is an interactive process. People place themselves and others into some social categories based on some underlying attributes that are salient. These attributes operate as a checklist; that is, individuals need not feature all at once. Rather, people perceive themselves to be similar to a class of attributes — which constitute in-group — in comparison to another class of attributes — featured by an out-group.

The gist of the SCT is to suggest that in-group and out-group are both defined as a result of self-categorization. Self-categorization is an interactive process whereby people compare intra-group differences to differences from other groups. This comparison has two steps: First, individuals reduce others to a stereotype — a checklist of salient attributes. Second, they establish a checklist of salient attributes defining their in-group, which is called prototype. In other words, prototype refers to a set of attributes that in-group members exclusively feature and which differentiates them from out-groups. During this comparison of in-group to out-group, individuals underscore within-group similarities and exaggerate their differences from other groups.

As a result of the self-categorization process, individuals began to perceive themselves as a group member rather than an individual with distinct attributes. This process called de-personalization has two mechanisms: Perceived fit or the meta-contrast principle refers to the degree to which the perceived in-group differences are less than the perceived differences between the in-group and out-groups. In other words, meta-contrast suggests that intra-group differences are weighed against inter-group differences. Normative fit represents the extent to which similarities and differences between group members overlap with the meaning of group membership (Lange, Kruglanski, & Higgins, 2012).

Comparison of in-group to out-group need not be confined to one dimension. Crossed categorizationrefers to comparisons done with respect to more than one dimensions. For example, people can compare themselves to others with respect to gender and with respect to race without having to choose between the two. Crossed categorization refers to comparisons, where both dimensions are simultaneously used, e.g., Black women, white men. Crossed categorization reduces inter-group conflict. In that, crossing one categorization with another produces four groups: an in-group on both dimensions, a double out-group; and two crossed conditions (in-group-out-group and out- group-in-group) (Hewstone, Islam, & Judd, 1993).

Crossed categorization has been studied in three different ways. In the additive model, each activated category has a positive or negative influence. This model adds them all up and finds that some crossed category can be neutral, positively perceived, or negatively perceived (Echabe & Guede, 2006; Singh, Yeoh, Lim, & Lim, 1997). The hierarchical model suggests that of the crossed categories, some are more important than others. For example, in a context where both class and ethnicity are activated, one of them, say, class might be more impactful than the other. Finally, in the interactive model, the activated categories interact with one another without adding up or being hierarchical. Their interactions generate new categories, such as rich & black, where neither class or race predominates the other. Importantly, the novel categories that emerge also trigger the rise of new biases.

Axelrod, R. (1997). The Dissemination of Culture A Model with Local Convergence and Global Polarization. Journal of Conflict Resolution, 41(2), 203–226.

Billikopf Encina, G. (2014). Milgram’s Experiment on Obedience to Authority. Retrieved from

Echabe, A. E., & Guede, E. F. (2006). Crossed-categorization and stereotypes: class and ethnicity, Abstract. Revue internationale de psychologie sociale, Tome 19(2), 81–101.

Hewstone, M., Islam, M. R., & Judd, C. M. (1993). Models of crossed categorization and intergroup relations. Journal of Personality and Social Psychology, 64(5), 779–793.

Lange, P. A. M. V., Kruglanski, A. W., & Higgins, E. T. (Eds.). (2012). Handbook of theories of social psychology. Los Angeles: SAGE.

Milgram, S. (1974). Obedience to Authority: An Experimental View. Harper & Row.

Salzarulo, L. (2006). A Continuous Opinion Dynamics Model Based on the Principle of Meta-Contrast. Journal of Artificial Societies and Social Simulation, 9(1).

Singh, R., Yeoh, B. S. E., Lim, D. I., & Lim, K. K. (1997). Cross-categorization effects in intergroup discrimination: Adding versus averaging. British Journal of Social Psychology, 36(2), 121–138.

Computational social scientist, artist (I draw!), and a math and philosophy lover