Thinking about others in terms of their group memberships is known as social categorization—the natural cognitive process by which we place individuals into social groups. Social categorization occurs when we think of someone as a man (versus a woman), an old person (versus a young person), a Black person (versus an Asian or White person), and so on (Allport, 1954/1979).  Just as we categorize objects into different types, so do we categorize people according to their social group memberships. Once we do so, we begin to respond to those people more as members of a social group than as individuals.

Imagine for a moment that two college students, Farhad and Sarah, are talking at a table in the student union at your college or university. At this point, we would probably not consider them to be acting as group members, but rather as two individuals. Farhad is expressing his opinions, and Sarah is expressing hers. Imagine, however, that as the conversation continues, Sarah brings up an assignment that she is completing for her women’s studies class. It turns out that Farhad does not think there should be a women’s studies program at the college, and he tells Sarah so. He argues that if there is a women’s studies program, then there should be a men’s studies program too. Furthermore, he argues that women are getting too many breaks in job hiring and that qualified men are the targets of discrimination. Sarah feels quite the contrary—arguing that women have been the targets of sexism for many, many years and even now do not have the same access to high-paying jobs that men do.

You can see that an interaction that began at individual level, as two individuals conversing, has now turned to the group level, in which Farhad has begun to consider himself as a man, and Sarah has begun to consider herself as a woman. In short, Sarah is now arguing her points not so much for herself as she is as a representative of one of her ingroups—namely, women—and Farhad is acting as a representative of one of his ingroups—namely, men. Sarah feels that her positions are correct, and she believes they are true not only for her but for women in general. And the same is true of Farhad. You can see that these social categorizations may create some potential for misperception, and perhaps even hostility. And Farhad and Sarah may even change their opinions about each other, forgetting that they really like each other as individuals, because they are now responding more as group members with opposing views.

Imagine now that while Farhad and Sarah are still talking, some students from another college, each wearing the hats and jackets of that school, show up in the student union. The presence of these outsiders might change the direction of social categorization entirely, leading both Farhad and Sarah to think of themselves as students at their own college. And this social categorization might lead them to become more aware of the positive characteristics of their college (the excellent rugby team, lovely campus, and intelligent students) in comparison with the characteristics of the other school. Now, rather than perceiving themselves as members of two different groups (men versus women), Farhad and Sarah might suddenly perceive themselves as members of the same social category (students at their college).

Perhaps this example will help you see the flexibility of social categorization. We sometimes think of our relationships with others at the individual level and sometimes at the group level. And which groups we use in social categorization can change over time and in different situations. You are more likely to categorize yourself as a member of your college or university when your rugby or football team has just won a really important game, or at your graduation ceremony, than you would on a normal evening out with your family. In these cases, your membership as a university student is simply more salient and important than it is every day, and you are more likely to categorize yourself accordingly.

Spontaneous Social Categorization

Social categorization occurs spontaneously, without much thought on our part (Crisp & Hewstone, 2007). Shelley Taylor and her colleagues (Taylor, Fiske, Etcoff, & Ruderman, 1978) showed their research participants a slide and tape presentation of three male and three female college students who had supposedly participated in a discussion group. During the presentation, each member of the discussion group made a suggestion about how to advertise a college play. The statements were controlled so that across all the research participants, the statements made by the men and the women were of equal length and quality. Furthermore, one half of the participants were told that when the presentation was over, they would be asked to remember which person had made which suggestion, whereas the other half of the participants were told merely to observe the interaction without attending to anything in particular.

After they had viewed all the statements made by the individuals in the discussion group, the research participants were given a memory test (this was entirely unexpected for the participants who had not been given memory instructions). The participants were shown the list of all the statements that had been made, along with the pictures of each of the discussion group members, and were asked to indicate who had made each of the statements. The research participants were not very good at this task, and yet when they made mistakes, these errors were very systematic.

As you can see in Table 11.1, “Name Confusions,” the mistakes were such that the statements that had actually been made by a man were more frequently wrongly attributed to another man in the group than to another woman, and the statements actually made by a woman were more frequently attributed to other women in the group than to a man. The participants evidently categorized the speakers by their gender, leading them to make more within-gender than across-gender confusions.

Interestingly, and suggesting that categorization is occurring all the time, the instructions that the participants had been given made absolutely no difference. There was just as much categorization for those who were not given any instructions as for those who were told to remember who said what. Other research using this technique has found that we spontaneously categorize each other on the basis of many other group memberships, including race, academic status (student versus teacher), social roles, and other social categories (Fiske, Haslam, & Fiske, 1991; Stangor, Lynch, Duan, & Glass, 1992).

The conclusion is simple, if perhaps obvious: Social categorization is occurring all around us all the time. Indeed, social categorization occurs so quickly that people may have difficulty not thinking about others in terms of their group memberships (see Figure 11.4).

The Benefits of Social Categorization

The tendency to categorize others is often quite useful. In some cases, we categorize because doing so provides us with information about the characteristics of people who belong to certain social groups (Lee, Jussim, & McCauley, 1995). If you found yourself lost in a city, you might look for a police officer or a taxi driver to help you find your way. In this case, social categorization would probably be useful because a police officer or a taxi driver might be particularly likely to know the layout of the city streets. Of course, using social categories will only be informative to the extent that the stereotypes held by the individual about that category are accurate. If police officers were actually not that knowledgeable about the city layout, then using this categorization heuristic would not be informative.

The description of social categorization as a heuristic is also true in another sense: we sometimes categorize others not because it seems to provide more information about them but because we may not have the time (or the motivation) to do anything more thorough. Using our stereotypes to size up another person might simply make our life easier (Macrae, Bodenhausen, Milne, & Jetten, 1994). According to this approach, thinking about other people in terms of their social category memberships is a functional way of dealing with the world—things are complicated, and we reduce complexity by relying on our stereotypes.

The Negative Outcomes of Social Categorization

Although thinking about others in terms of their social category memberships has some potential benefits for the person who does the categorizing, categorizing others, rather than treating them as unique individuals with their own unique characteristics, has a wide variety of negative, and often very unfair, outcomes for those who are categorized.

One problem is that social categorization distorts our perceptions such that we tend to exaggerate the differences between people from different social groups while at the same time perceiving members of groups (and particularly outgroups) as more similar to each other than they actually are. This overgeneralization makes it more likely that we will think about and treat all members of a group the same way. Tajfel and Wilkes (1963) performed a simple experiment that provided a picture of the potential outcomes of categorization. As you can see in Figure 11.5, “Perceptual Accentuation,” the experiment involved having research participants judge the length of six lines. In one of the experimental conditions, participants simply saw six lines, whereas in the other condition, the lines were systematically categorized into two groups—one comprising the three shorter lines and one comprising the three longer lines.

Lines C and D were seen as the same length in the noncategorized condition, but line C was perceived as longer than line D when the lines were categorized into two groups. From Tajfel (1970).

Tajfel found that the lines were perceived differently when they were categorized, such that the differences between the groups and the similarities within the groups were emphasized. Specifically, he found that although lines C and D (which are actually the same length) were perceived as equal in length when the lines were not categorized, line D was perceived as being significantly longer than line C in the condition in which the lines were categorized. In this case, categorization into two groups—the “short lines group” and the “long lines group”—produced a perceptual bias such that the two groups of lines were seen as more different than they really were.

Similar effects occur when we categorize other people. We tend to see people who belong to the same social group as more similar than they actually are, and we tend to judge people from different social groups as more different than they actually are. The tendency to see members of social groups as similar to each other is particularly strong for members of outgroups, resulting in outgroup homogeneity—the tendency to view members of outgroups as more similar to each other than we see members of ingroups (Linville, Salovey, & Fischer, 1986; Ostrom & Sedikides, 1992; Meissner & Brigham, 2001). Perhaps you have had this experience yourself when you found yourself thinking or saying, “Oh, them, they’re all the same!”

Patricia Linville and Edward Jones (1980) gave research participants a list of trait terms and asked them to think about either members of their own group (e.g., Blacks) or members of another group (e.g., Whites) and to place the trait terms into piles that represented different types of people in the group. The results of these studies, as well as other studies like them, were clear: people perceive outgroups as more homogeneous than their ingroup. Just as White people used fewer piles of traits to describe Blacks than Whites, young people used fewer piles of traits to describe elderly people than they did young people, and students used fewer piles for members of other universities than they did for members of their own university.

Outgroup homogeneity occurs in part because we don’t have as much contact with outgroup members as we do with ingroup members, and the quality of interaction with outgroup members is often more superficial. This prevents us from really learning about the outgroup members as individuals, and as a result, we tend to be unaware of the differences among the group members. In addition to learning less about them because we see and interact with them less, we routinely categorize outgroup members, thus making them appear more cognitively similar (Haslam, Oakes, & Turner, 1996).

Once we begin to see the members of outgroups as more similar to each other than they actually are, it then becomes very easy to apply our stereotypes to the members of the groups without having to consider whether the characteristic is actually true of the particular individual. If men think that women are all alike, then they may also think that they all have the same positive and negative characteristics (e.g., they’re nurturing, emotional). And women may have similarly simplified beliefs about men (e.g., they’re strong, unwilling to commit). The outcome is that the stereotypes become linked to the group itself in a set of mental representations (Figure 11.6). The stereotypes are “pictures in our heads” of the social groups (Lippman, 1922). These beliefs just seem right and natural, even though they are frequently distorted overgeneralizations (Hirschfeld, 1996; Yzerbyt, Schadron, Leyens, & Rocher, 1994).

Our stereotypes and prejudices are learned through many different processes. This multiplicity of causes is unfortunate because it makes stereotypes and prejudices even more likely to form and harder to change. For one, we learn our stereotypes in part through our communications with parents and peers (Aboud & Doyle, 1996) and from the behaviors we see portrayed in the media (Brown, 1995). Even five-year-old children have learned cultural norms about the appropriate activities and behaviors for boys and girls and also have developed stereotypes about age, race, and physical attractiveness (Bigler & Liben, 2006). And there is often good agreement about the stereotypes of social categories among the individuals within a given culture. In one study assessing stereotypes, Stephanie Madon and her colleagues (Madon et al., 2001) presented U.S. college students with a list of 84 trait terms and asked them to indicate for which groups each trait seemed appropriate (Figure 11.7, “Current Stereotypes Held by College Students”). The participants tended to agree about what traits were true of which groups, and this was true even for groups of which the respondents were likely to never have met a single member (Arabs and Russians). Even today, there is good agreement about the stereotypes of members of many social groups, including men and women and a variety of ethnic groups.

Once they become established, stereotypes (like any other cognitive representation) tend to persevere. We begin to respond to members of stereotyped categories as if we already knew what they were like. Yaacov Trope and Eric Thompson (1997) found that individuals addressed fewer questions to members of categories about which they had strong stereotypes (as if they already knew what these people were like) and that the questions they did ask were likely to confirm the stereotypes they already had.

In other cases, stereotypes are maintained because information that confirms our stereotypes is better remembered than information that disconfirms them. When we see members of social groups perform behaviors, we tend to better remember information that confirms our stereotypes than we remember information that disconfirms our stereotypes (Fyock & Stangor, 1994). If we believe that women are bad drivers and we see a woman driving poorly, then we tend to remember it, but when we see a woman who drives particularly well, we tend to forget it. This illusory correlation is another example of the general principle of assimilation—we tend to perceive the world in ways that make it fit our existing beliefs more easily than we change our beliefs to fit the reality around us.

And stereotypes become difficult to change because they are so important to us—they become an integral and important part of our everyday lives in our culture. Stereotypes are frequently expressed on TV, in movies, and in social media, and we learn a lot of our beliefs from these sources. Our friends also tend to hold beliefs similar to ours, and we talk about these beliefs when we get together with them (Schaller & Conway, 1999). In short, stereotypes and prejudice are powerful largely because they are important social norms that are part of our culture (Guimond, 2000).

Because they are so highly cognitively accessible, and because they seem so “right,” our stereotypes easily influence our judgments of and responses to those we have categorized. The social psychologist John Bargh once described stereotypes as “cognitive monsters” because their activation was so powerful and because the activated beliefs had such insidious influences on social judgment (Bargh, 1999). Making things even more difficult, stereotypes are strongest for the people who are in most need of change—the people who are most prejudiced (Lepore & Brown, 1997).

Because stereotypes and prejudice often operate out of our awareness, and also because people are frequently unwilling to admit that they hold them, social psychologists have developed methods for assessing them indirectly. In the Research Focus box following, we will consider two of these approaches—the bogus pipeline procedure and the Implicit Association Test (IAT).

Although in some cases the stereotypes that are used to make judgments might actually be true of the individual being judged, in many other cases they are not. Stereotyping is problematic when the stereotypes we hold about a social group are inaccurate overall, and particularly when they do not apply to the individual who is being judged (Stangor, 1995). Stereotyping others is simply unfair. Even if many women are more emotional than are most men, not all are, and it is not right to judge any one woman as if she is.

In the end, stereotypes become self-fulfilling prophecies, such that our expectations about the group members make the stereotypes come true (Snyder, Tanke, & Berscheid, 1977; Word, Zanna, & Cooper, 1974). Once we believe that men make better leaders than women, we tend to behave toward men in ways that makes it easier for them to lead. And we behave toward women in ways that makes it more difficult for them to lead. The result? Men find it easier to excel in leadership positions, whereas women have to work hard to overcome the false beliefs about their lack of leadership abilities (Phelan & Rudman, 2010). This is likely why female lawyers with masculine names are more likely to become judges (Coffey & McLaughlin, 2009) and masculine-looking applicants are more likely to be hired as leaders than feminine-looking applicants (von Stockhausen, Koeser, & Sczesny, 2013).

These self-fulfilling prophecies are ubiquitous—even teachers’ expectations about their students’ academic abilities can influence the students’ school performance (Jussim, Robustelli, & Cain, 2009).

Of course, you may think that you personally do not behave in these ways, and you may not. But research has found that stereotypes are often used out of our awareness, which makes it very difficult for us to correct for them. Even when we think we are being completely fair, we may nevertheless be using our stereotypes to condone discrimination (Chen & Bargh, 1999). And when we are distracted or under time pressure, these tendencies become even more powerful (Stangor & Duan, 1991).

Furthermore, attempting to prevent our stereotype from coloring our reactions to others takes effort. We experience more negative affect (particularly anxiety) when we are with members of other groups than we do when we are with people from our own groups, and we need to use more cognitive resources to control our behavior because of our anxiety about revealing our stereotypes or prejudices (Butz & Plant, 2006; Richeson & Shelton, 2003). When we know that we need to control our expectations so that we do not unintentionally stereotype the other person, we may try to do so—but doing so takes effort and may frequently fail (Macrae, Bodenhausen, Milne, & Jetten, 1994).

Social Psychology in the Public Interest

Stereotype Threat

Our stereotypes influence not only our judgments of others but also our beliefs about ourselves, and even our own performance on important tasks. In some cases, these beliefs may be positive, and they have the effect of making us feel more confident and thus better able to perform tasks. Because Asian students are aware of the stereotype that “Asians are good at math,” reminding them of this fact before they take a difficult math test can improve their performance on the test (Walton & Cohen, 2003). On the other hand, sometimes these beliefs are negative, and they create negative self-fulfilling prophecies such that we perform more poorly just because of our knowledge about the stereotypes.

One of the long-standing puzzles in the area of academic performance concerns why Black students in the United States perform more poorly on standardized tests, receive lower grades, and are less likely to remain in school in comparison with White students, even when other factors such as family income, parents’ education, and other relevant variables are controlled. Claude Steele and Joshua Aronson (1995) tested the hypothesis that these differences might be due to the activation of negative stereotypes. Because Black students are aware of the (inaccurate) stereotype that “Blacks are intellectually inferior to Whites,” this stereotype might create a negative expectation, which might interfere with their performance on intellectual tests through fear of confirming that stereotype.

In support of this hypothesis, Steele and Aronson’s research revealed that Black college students performed worse (in comparison with their prior test scores) on math questions taken from the Graduate Record Examination (GRE) when the test was described to them as being “diagnostic of their mathematical ability” (and thus when the stereotype was relevant) but that their performance was not influenced when the same questions were framed as “an exercise in problem solving.” And in another study, Steele and Aronson found that when Black students were asked to indicate their race before they took a math test (again activating the stereotype), they performed more poorly than they had on prior exams, whereas the scores of White students were not affected by first indicating their race.

Steele and Aronson argued that thinking about negative stereotypes that are relevant to a task that one is performing creates stereotype threat—performance decrements that are caused by the knowledge of cultural stereotypes. That is, they argued that the negative impact of race on standardized tests may be caused, at least in part, by the performance situation itself. Because the threat is “in the air,” Black students may be negatively influenced by it.

Research has found that the experience of stereotype threat can help explain a wide variety of performance decrements among those who are targeted by negative stereotypes. For instance, when a math task is described as diagnostic of intelligence, Latinos and particularly Latinas perform more poorly than do Whites (Gonzales, Blanton, & Williams, 2002). Similarly, when stereotypes are activated, children with low socioeconomic status perform more poorly in math than do those with high socioeconomic status, and psychology students perform more poorly than do natural science students (Brown, Croizet, Bohner, Fournet, & Payne, 2003). Even groups who typically enjoy advantaged social status can be made to experience stereotype threat. White men performed more poorly on a math test when they were told that their performance would be compared with that of Asian men (Aronson, Lustina, Good, Keough, & Steele, 1999), and Whites performed more poorly than Blacks on a sport-related task when it was described to them as measuring their natural athletic ability (Stone, 2002).

Stereotype threat is created in situations that pose a significant threat to self-concern, such that our perceptions of ourselves as important, valuable, and capable individuals are threatened. In these situations, there is a discrepancy between our positive concept of our skills and abilities and the negative stereotypes suggesting poor performance. When our stereotypes lead us to be believe that we are likely to perform poorly on a task, we experience a feeling of unease and status threat.

Research has found that stereotype threat is caused by both cognitive and affective factors. On the cognitive side, individuals who are experiencing stereotype threat show an impairment in cognitive processing that is caused by increased vigilance toward the environment and attempts to suppress their stereotypical thoughts. On the affective side, stereotype threat creates stress as well as a variety of affective responses including anxiety (Schmader, Johns, & Forbes, 2008).

Stereotype threat is not, however, absolute—we can get past it if we try. What is important is to reduce the self-concern that is engaged when we consider the relevant negative stereotypes. Manipulations that affirm positive characteristics about oneself or one’s group are successful at reducing stereotype threat (Alter, Aronson, Darley, Rodriguez, & Ruble, 2010; Greenberg et al., 2003; McIntyre, Paulson, & Lord, 2003). In fact, just knowing that stereotype threat exists and may influence performance can help alleviate its negative impact (Johns, Schmader, & Martens, 2005).

Media Attributions

• “CHILLIN” by sabrina’s stash is licensed under a CC BY 2.0 licence.