Gender Stereotypes and Gender Differences

The Psychology of Women and Gender: Half the Human Experience + - Nicole M. Else-Quest, Janet Shibley Hyde 2018

Gender Stereotypes and Gender Differences

“Man should be trained for war and woman for the recreation of the warrior.”

Nietzsche, Thus Spoke Zarathustra (1883)

Some people believe that this is the postfeminist era, that gender stereotypes have vanished, and that women (and men) can be anything they want to be. But then why are people so offended by a man who behaves in a feminine manner? Why are some people so upset by transgender and genderqueer folks, to the point that they pass legislation about which bathrooms they can use? Gender stereotypes and gender roles are still in force in contemporary culture, and violations of those roles and stereotypes seem very serious to some people.

Gender Stereotypes

Stereotypes About Men and Women

Gender stereotypes are simply a set of shared cultural beliefs about men’s and women’s behavior, appearance, interests, personality, and so on. Research shows that even in modern American society, and even among college students, there is a belief that men and women do differ psychologically in many ways. A list of these gender-stereotyped traits is given in Table 3.1.

Gender stereotypes: A set of shared cultural beliefs about men’s and women’s behavior, appearance, interests, personality, and so on.

How do researchers collect evidence of these stereotypes? In the study whose results are shown in Table 3.1, the researchers recruited a sample of undergraduates who were ethnically diverse (Ghavami & Peplau, 2013). Participants read instructions that said,

We are all aware of cultural stereotypes of social groups. These may be ideas that you learned from movies, saw in magazines. . . . For example, people often perceive models as beautiful, tall, but dumb. Note that these characteristics may or may not reflect your own personal beliefs about these groups. In the space below, list at least 10 characteristics that are part of the current cultural stereotypes of [the target group] as a group rather than a specific individual you may know. Please note that we are not asking for your personal beliefs, but rather those held by people in general.

Respondents received different groups for the target group, for example, women, Asian Americans, or Black women. It turns out that, if you ask the question this way, people will give you the stereotypes. Not every participant uses the same words to describe a quality, so the research team then groups synonyms together to come up with the lists. For example, one person might say wealthy and another might say rich. Those would be clustered together.


Source: Data from Ghavami & Peplau (2013).

As shown in Table 3.1, women are believed to be emotional, caring, and talkative, stereotypes that have been around for decades. The same is true of the stereotypes of men as strong, intelligent, and leaders. Gender stereotypes have changed little from the early 1980s (Haines et al., 2016).

Although these gender stereotypes persist in modern American culture, the evidence indicates that attitudes about gender roles have changed considerably over the last 30 or more years. Some data on this point are shown in Table 3.2. For example, the attitude that the man should be the achiever outside the home and the woman should tend to the home and family went from 66% in 1977 to 32% in 2012 (Smith et al., 2012).

The data seem to present a paradox. Table 3.1 shows evidence of continuing old-fashioned gender stereotypes, whereas Table 3.2 shows that gender-related attitudes have changed considerably. How can that be? The answer is that both findings are actually compatible. Americans are deeply committed to the principles of equality and justice, and when feminists posed the issue of women’s rights within that framework, many people were persuaded and changed their attitudes in the equal rights direction shown in Table 3.2. Yet gender stereotypes are powerful and seem harmless, and people privately believe that women are emotional and submissive and men are intelligent and leaders. Sarah is a woman, so she is emotional and submissive, and she has every right to run for president.

Why do people stereotype others? Social psychologists have uncovered two basic goals: comprehension and self-enhancement (Kunda & Spencer, 2003). As for the comprehension goal, when we meet a new person, we tend to fill in a lot of assumed information about that person so that we can understand them until we have more actual information. For example, breadwinner is a key aspect of the male role. When we meet a man, we are likely to invoke that stereotype and ask an opening question such as “What kind of work do you do?” Our first question is not “Are you a househusband?” When people stereotype for comprehension purposes, the stereotypes can be positive or negative. They are just trying to understand more about the person than they actually know. Of course, they may make errors in the process. We might actually be talking to a househusband, and he is offended by our assumption that he holds a job outside the home.


Source: Data from Smith et al. (2012).

In contrast, when we stereotype for self-enhancement purposes, the stereotypes tend to be negative. We make ourselves feel better by denigrating another group. For example, if we say (or think), “Teenagers are so irresponsible,” by implication we, as adults, are highly responsible.

In contrast to the conscious stereotypes shown in Table 3.1, implicit stereotypes are learned, automatic associations between social categories (e.g., women) and other attributes (e.g., nurse but not mathematician or scientist; Rudman & Glick, 2008). These stereotypes are not necessarily conscious. The method used to measure these implicit stereotypes is the Implicit Association Test (IAT), which measures an individual’s relative strength of association between different pairs of concepts (Nosek et al., 2005). The key to measuring these associations is reaction time, which is measured on a computer in milliseconds. We react quickly to two concepts that we associate strongly and more slowly to two concepts that we do not associate strongly. One of the advantages of this measure is that people can’t fake their reaction times. For example, they cannot hide their socially unacceptable stereotyped ideas. (If you want to try the IAT yourself, you can do it online at

Implicit stereotypes: Learned, automatic associations between social categories (e.g., female) and other attributes (e.g., nurse but not mathematician).

In one important experiment, Mahzarin Banaji and her colleagues measured the association between gender and math (Nosek et al., 2002). In the practice phase, participants placed one finger on the left key of a keypad and another finger on the right key. They were instructed to press the left key if the word they saw on the screen in front of them was in the category math (e.g., algebra, equation) or if it was in the category pleasant (e.g., peace, love). They were to press the right key for topics in the arts (e.g., drama, poetry) or words that were unpleasant (e.g., hatred). After following this pattern for many trials, the instructions changed and they had to press the left key for math words and unpleasant words, and the right key for arts and pleasant words. All of this was practice for the real task. In the first phase of it, participants pressed the left key if the words were in the math category or the male category (e.g., male, boy) and the right key if the words were in the arts category or the female category (e.g., female, girls). Then in the second phase the pairings were reversed, so participants pressed the left key for the math category and the female category and the right key for the arts category and the male category. Implicit stereotyping is indicated if people respond faster to the male and math pairing than they do to the female and math pairing, and that is exactly what participants do! That is, people have an implicit association between math and male but not female.

In a similar study, researchers demonstrated an implicit association between science and male but not female (Carli et al., 2016). We will return to this finding in Chapter 8, when we consider why women are underrepresented in the STEM (science, technology, engineering, and mathematics) fields.

In an ambitious study, the researchers collected data on the implicit association between male and science for people from 34 nations around the world, using an online version of the IAT (Nosek et al., 2009). Implicit stereotyping is stronger in some nations than in others. The researchers also tapped international data on the gender gap in the science knowledge of eighth graders in these countries, using an approach similar to the one used by Eagly and Wood (1999; see the “Social Role Theory” section in Chapter 2). They found that the correlation, across nations, between implicit stereotyping of science as male and the gender gap in science performance was an amazing r = 0.60! That means that, to the extent that people in a country stereotype boys and men as better at science, boys in that country perform better than girls do on standardized science tests. The researchers believe that implicit stereotypes and the gender gap in science performance contribute to a vicious cycle. Implicit stereotypes held by adults and youth in a country discourage girls from studying science. And then, when girls do not do well in science (because they have studied it less), that strengthens the implicit stereotypes.

Intersectionality and Gender Stereotypes

As explained in earlier chapters, one of the essential principles of feminist theory is intersectionality. Here we will examine the intersection of gender and ethnicity when it comes to gender stereotypes. An intersectional approach tells us that gender stereotypes may not be the same in different ethnic groups.

Table 3.3 shows stereotypes about women and men from different ethnic groups in the United States (Ghavami & Peplau, 2013). Consistent with an intersectionality hypothesis, the gender-ethnic stereotypes contain distinct elements that do not represent adding gender stereotypes to ethnic stereotypes, or ethnic stereotypes to gender stereotypes. For example, Black women are stereotyped as athletic, but that stereotype is not found for Middle Eastern women, Latinx women, White women, or Asian American women. Latinx men are described as arrogant, as are White men and White women, but none of the other gender-ethnic groups. Essentially, then, within an ethnic group, men and women have some stereotyped traits in common, but also some that differ. For example, both Latinx men and Latinx women are described as hardworking, but Latinx men are described as arrogant whereas Latinx women are not. White women and Asian American women are stereotyped as intelligent, but women from the other ethnic groups are not. Given this intersectional analysis, it is difficult to talk simply about gender stereotypes in the United States, because those stereotypes are so specific to particular ethnic groups. Given the dominance of Whites in American society, what we have traditionally thought of as gender stereotypes are probably stereotypes about White men and women.

Stereotype Threat

Stereotypes are more than just abstract ideas. They can really hurt. Psychologist Claude Steele discovered a phenomenon he calls stereotype threat, and it documents the kind of subtle damage that stereotypes can inflict (Steele, 1997; Steele & Aronson, 1995). Stereotype threat can be defined as a situation in which there is a negative stereotype about a person’s group, and the person is concerned about being judged or treated negatively on the basis of that stereotype (Spencer et al., 2016).

Stereotype threat: A situation in which there is a negative stereotype about a person’s group, and the person is concerned about being judged or treated negatively on the basis of that stereotype.

Steele’s original work concerned ethnic stereotypes—specifically, the stereotype that African Americans are intellectually inferior. In one experiment, he administered a test of verbal intelligence to Black and White college students, all of whom were highly talented Stanford students. Half of each group were told that the test was diagnostic of intelligence and half were told it was not diagnostic of intelligence. The Black students who were told that the test measured intelligence performed worse than the Black students who were told it didn’t, whereas White students’ performance was unaffected by the instructions they received. The effect for the Black students demonstrates stereotype threat.

Other researchers then quickly moved to test whether stereotype threat applies to gender stereotypes—in particular, the stereotype that women are bad at math (Brown & Josephs, 1999; Quinn & Spencer, 2001; Schmader & Johns, 2003; Walsh et al., 1999). In one experiment, male and female college students with equivalent math backgrounds were tested (Spencer et al., 1999). Half were told that the math test had shown gender differences in the past and half were told that the test had been shown to be gender fair—that men and women had performed equally on it. Among those who believed that the test was gender fair, there were no gender differences in performance, but among those who believed it showed gender differences, women underperformed compared with men. Stereotypes about women and math hurt women’s performance. In another experiment by the same researchers, women performed worse on the math test even when there was no mention of gender differences. The stereotype about women and mathematics is so much a part of the culture, it did not even have to be primed by the experimenters. The stereotype is simply there for women any time they encounter difficult mathematics problems.


Source: Data from Ghavami & Peplau (2013).

Meta-analyses show that the size of the stereotype threat effect on women’s math performance ranges between d = 0.17 and d = 0.36 (Spencer et al., 2016). The size of the stereotype threat effect for African Americans and Latinx on intellectual tests is somewhat larger, around d = 0.50 (Spencer et al., 2016).

What about the intersection of gender and ethnicity in stereotype threat? The case of Asian American women and mathematics is particularly interesting. As women, they are stereotyped as being bad at math, but as Asian Americans they are stereotyped as being good at math. Research shows that when Asian American women’s ethnic identity is primed (highlighted), they perform better on math problems, and when their gender identity is primed, they perform worse, compared with a control group that has had neither identity primed (Shih et al., 1999).

Latinx women face a different set of challenges than Asian American women, because both their gender and their ethnic group are stereotyped as weak in math. In one experiment, Latinx men and women and White men and women were randomly assigned to either a stereotype threat condition (they were told that the math test they were about to take was diagnostic of their “actual abilities and limitations”) or a no-threat condition (no reference to their ability was made; Gonzales et al., 2002). They then completed a difficult math test. The results are shown in Figure 3.1. Notice that the performance of Latinx men is hurt somewhat under the stereotype threat (diagnostic) condition, but the performance of Latinx women is hurt more. Being the object of two negative stereotypes seems to hurt their performance twice as much. Notice also that stereotype threat actually helps the performance of White men, something that has been found in numerous studies and has been termed “stereotype lift” (Walton & Cohen, 2003).

Figure 3.1 Results of a study measuring the performance of Latinx and White women and men on a difficult math test under stereotype threat and no-threat conditions.


Source: Gonzales et al. (2002).

What are the psychological processes that underlie stereotype threat effects? Two processes have been documented, and they may be in play for different individuals in different situations (Spencer et al., 2016).

1. Underperformance due to extra pressure to succeed: People in a stereotype threat situation are usually motivated to disconfirm the negative stereotype about their group. This leads to extra pressure to perform well, a pressure not experienced by others. This kind of pressure leads people to exert more effort, and sometimes that helps their performance (e.g., on easy tasks), but in other cases (e.g., difficult tasks) the pressure becomes highly stressful and hurts performance. The pressure can also deplete working memory capacity, which is needed for difficult intellectual tasks.

2. Underperformance due to threats to self-integrity and belonging: Stereotype threat can threaten people’s sense of self-worth. To protect themselves, people may engage in various kinds of self-handicapping, such as setting lower goals for themselves so that they don’t fail, which only makes them reach lesser goals. Stereotype threat can also reduce a person’s sense of belonging, reducing their motivation and perhaps leading them to withdraw from the situation.

Can anything be done to counteract stereotype threat? Researchers have devised and tested a number of interventions designed to reduce stereotype-threat effects (Spencer et al., 2016). One intervention strategy involves guiding people to reconstrue (think differently about) a threatening situation as less threatening. For example, people may be taught to think differently about their anxiety. People can also be given coping strategies to deal with threatening situations and to maintain their self-integrity. For example, in one classroom study, a self-affirmation exercise with middle-schoolers closed the race gap in school performance by 40% (Cohen et al., 2006). These techniques need to be deployed widely in our schools.

Stereotypes About Trans Individuals

Gender stereotype research rests on the gender binary and tells us about stereotypes about women and stereotypes about men. There is little or no research on stereotypes about trans individuals, but we can glean some ideas from related research.

In one study, transgender individuals were asked whether they had had experiences of prejudice, discrimination, and stereotypes related to being transgender (Mizock & Hopwood, 2016). One stereotype is that trans individuals are gay or lesbian; that is, some people conflate sexual orientation and gender identity. Such people may not even be aware of gender diversity, and they react to a trans individual with anti-gay prejudice. In addition, gender-binary privilege and stereotypes can be highly salient for transgender individuals. For example, a trans woman has grown up enjoying male privilege but, after transitioning, loses that privilege in ways that may be quite disconcerting. As one participant in the study said,

The second they feminize their appearance, they’ve lost their true male privilege, but they haven’t lost their voice. Women are so accustomed to being repressed, that these bold [trans] women who still retain their voice seem male to them. And that’s not male [to be outspoken], it’s them sticking up for themselves. (quoted in Mizock & Hopwood, 2016, p. 99)

Other research has shown that trans individuals are stereotyped as mentally ill (Reed et al., 2015), much as gays and lesbians were a few decades ago. Importantly, much of the prejudice against trans people seems to result from their violation of traditional gender role stereotypes. For example, on the Genderism and Transphobia Scale, one item is “I have behaved violently towards a woman because she was too masculine,” and another is “I have teased a man because of his feminine appearance or behavior” (Hill & Willoughby, 2005). Thus cisgenderism—prejudice against people who are outside the gender binary—is deeply rooted in resentment toward those who violate gender role stereotypes.

Psychologists have begun to design interventions to reduce prejudice toward transgender people. In one case, researchers designed an intervention aimed at humanizing and perspective taking (Tompkins et al., 2015). Participants (college students) in the intervention condition viewed a video about Jazz, a transgender child, to humanize trans children for the participants. They then completed a writing assignment designed to promote perspective taking, in which they imagined that they were transgender and wrote a letter coming out to their parents as transgender. Participants in the control condition were presented with DSM criteria for gender dysphoria and heard a videotaped lecture by an expert. The intervention was successful at reducing trans prejudice. This study is a good first step along the way to developing successful interventions that could be used, for example, with all students entering college.


At this point, studies of psychological gender differences number in the thousands. These studies are all based on the assumption of a gender binary—that is, that there are only two genders, female and male. Later in the chapter, we will consider how trans individuals might fit into this research in the future.

With these thousands of studies, we should have a thorough understanding of which behaviors show gender differences and which do not. Unfortunately, things are a bit more complicated than that. Often the results of different studies contradict each other. For example, some studies of gender differences in infants’ activity levels find that boys are more active, whereas others find no gender difference. In such cases, what should we conclude?

Another problem is that sometimes a single study that finds a gender difference will be picked up by the media and included in textbooks, and the five other studies of the same behavior that found no gender difference will be ignored. It seems likely that this occurs particularly when a finding of gender differences confirms the stereotypes held by the scientists and the general public. Moreover, scientists and the general public are fascinated by findings of gender differences; they tend to find results indicating gender similarities to be, well, boring.

Meta-analysis is a technique that allows researchers to bring order out of this seeming chaos of sometimes contradictory studies (Lipsey & Wilson, 2001). Meta-analysis is a statistical method that allows the researcher to statistically combine the results from all previous studies of the question of interest to determine what, taken together, the studies say. In conducting a meta-analysis, the researcher goes through three steps:

Meta-analysis: A statistical technique that allows a researcher to combine the results of multiple research studies on a particular question.

1. The researcher locates all previous studies on the question being investigated (e.g., gender differences in aggression). This step is typically done using computerized searches of databases such as PsycINFO or Web of Science.

2. For each study, the researcher computes a statistic that measures how big the difference between male participants and female participants was and what the direction of the difference was (male participants scoring higher or female participants scoring higher). This statistic is called d. The formula for it is Image

where MM is the mean or average score for male participants, MF is the mean score for female participants, and s is the average standard deviation of the male scores and the female scores. If you’ve studied statistics, you know what a standard deviation is. For those of you who haven’t, the standard deviation is a measure of how much variability there is in a set of scores. For example, if the average score for people on test Q is 20 and all the scores fall between 19 and 21, then there is little variability and the standard deviation would be small. If, on the other hand, the average score for people is 20 and scores range from 0 to 40, then there is great variability and the standard deviation will be large. The d statistic, then, tells us, for a particular study, how big the difference between the male and female means was, relative to the variability in scores. If d is a positive value, then male participants scored higher; if d is negative, female participants scored higher; and if d is zero, there is no difference.

3. The researcher averages all the d values over all the studies that were located. When all studies are combined, this average d value tells what the direction of the gender difference is (whether male participants score higher or female participants score higher) and how large the difference is.

Although there is some disagreement among experts, a general guide is that a d of 0.20 is a small difference, a d of 0.50 is a moderate difference, and a d of 0.80 is a large difference (Cohen, 1969).

Numerous meta-analyses of gender differences are now available, most of them based on large numbers of studies. They are a much more reliable source than a single study. Meta-analyses, whenever available, will form the basis for the conclusions presented in this chapter and throughout this textbook.

It is also worth noting that meta-analysis can be used for synthesizing not just research on gender differences, but any research that uses a two-group design. For example, is cognitive-behavioral therapy effective, compared with a control group? How big is the effect?

Psychological Gender Differences

In this section, we will consider some of the scientific research on gender differences, focusing on aggressive behavior, impulsivity, activity, self-esteem, helping behavior, and anxiety. Discussions of other gender differences are found across the rest of the chapters in this book.

Aggressive Behavior

One of the most consistently documented psychological gender differences is in aggressive behavior, with boys and men being more aggressive than girls and women. Psychologists generally define aggression as behavior intended to harm another person. This gender difference holds up for many different kinds of aggression, especially physical aggression (Archer, 2004).

Aggression: Behavior intended to harm another person.

Developmentally, this gender difference appears about as early as children begin playing with each other, around the age of 2 (Alink et al., 2006; Baillargeon et al., 2007). The difference continues consistently throughout the school years. Of course, as people get older they become less aggressive, at least in the sense of physical aggression. It is rare to see adults rolling around on the floor as they punch each other, compared with the frequency with which that occurs on an elementary school playground. Less information is available on gender differences in adult aggression, but we do know that the great majority of crimes of violence are committed by men (although female crime is on the increase).

According to a meta-analysis, d = 0.55 for physical aggression, which is a moderate difference (Archer, 2004). For verbal aggression, d = 0.09, or close to no difference.

Recently there has been much publicity about “mean girls.” The idea is that girls do not express their aggression physically the way boys do, but rather are mean to each other, spreading degrading rumors or excluding someone from a social group. Psychologists call this type of aggression indirect aggression or relational aggression (Crick & Grotpeter, 1995; Werner & Crick, 2004). Are girls really the mean ones, the relationally aggressive ones? Meta-analysis shows that the gender differences are not as large or as consistent as one would expect from the publicity. Girls scored higher, but the gender difference was small, whether assessed by peer ratings (d = —0.19, notice that the negative sign means that girls scored higher) or teacher reports (d = —0.13; Archer, 2004). Boys are nearly as mean as girls are.

Relational aggression: Behavior intended to hurt others by damaging their peer relationships. Also termed indirect aggression.

What causes the gender difference in aggression? Researchers debate between nature and nurture. The nature team attributes gender differences in physical aggressiveness to the greater size and muscle mass of male bodies and/or differences in the levels of the hormone testosterone. These factors will be discussed in detail in Chapter 10.

Photo 3.1 Gender differences in aggressiveness appear early.



On the nurture side, a number of environmental forces might produce the observed gender difference. First, aggressiveness is a key part of the male role in our society, whereas aggressiveness is a violation of the female role. Following the logic of cognitive-developmental theory, as soon as children become aware of gender roles, girls realize that they are not supposed to be aggressive and boys know that they should be. As explained in Chapter 2, this reasoning does not work very well in explaining how gender differences develop so early, but it may be helpful in explaining gender differences among older children. Second, children imitate same-gender adults more than other-gender adults, and they see far more aggression in men than in women, particularly on TV and in the movies. In short, boys imitate men, who are aggressive, and girls imitate women, who are unaggressive. Third, boys receive more rewards for aggression and less punishment for it than girls do. These reinforcements and punishments might be in a physical form, such as spanking, or in a verbal form, such as comments from adults like “Boys will be boys” in response to a boy’s aggression. Boys may also be rewarded in the form of status or respect from their peers for being aggressive, whereas girls receive no such reward from their peers and may even find that other girls don’t want to play with them if they are aggressive. Research actually indicates that boys are punished more for aggression than girls are by both parents and teachers, thus posing a problem for this explanation. Yet psychologists believe that some kinds of punishments for aggression may actually increase a child’s aggression rather than decrease it. Therefore, the punishments that boys receive may make them more aggressive.

One interesting experiment tested the first hypothesis stated above, that gender roles are a powerful force creating gender differences in aggression (Lightdale & Prentice, 1994). The researchers used the technique of deindividuation to produce a situation that removed the influences of gender roles. Deindividuation refers to a state in which the person has lost their individual identity; that is, the person has become anonymous. Under such conditions people feel no obligation to conform to social norms such as gender roles; deindividuation essentially places the individual in a situation free of gender roles. Half the participants were placed in an individuated condition (the opposite of deindividuation), by having them sit close to the experimenter, identify themselves by name, wear large name tags, and answer personal questions. Deindividuated participants sat far from the experimenter and were simply told to wait quietly. All participants were also told that the experiment required information from only half the participants, whose behavior would be monitored, and that the other half would remain anonymous. Next, the participants played a video game in which they first defended and then attacked by dropping bombs. The number of bombs dropped was the measure of aggressive behavior.

Deindividuation: A state in which a person has become anonymous and has therefore lost their individual identity—and therefore the pressure to conform to gender roles.

The results indicated that, in the individuated condition, men dropped significantly more bombs (31.1, on average) than women did (26.8, on average). In the deindividuated condition—that is, in the absence of gender roles—there were no significant gender differences. In short, the significant gender differences in aggression disappeared when the influences of gender roles were removed.


Stereotypes hold that men are impulsive risk takers and that women are less so. Impulsivity refers to the tendency to act spontaneously and without careful thought (Cross et al., 2011). There are actually multiple aspects of impulsivity: reward sensitivity (being especially likely to do something because it will feel good right now), sensation seeking, risk taking, and impulse control (the opposite of impulsivity, i.e., being able to control one’s actions).

A meta-analysis found that men did indeed score higher than women on risk taking (d = 0.38) and sensation seeking (d = 0.22). There were no gender differences in reward sensitivity or impulse control, though.

Men’s greater tendency toward risk taking has negative implications for their health and life expectancy. (We will discuss this point in more detail in Chapter 16.) Yet there are also times in life when one has to take some calculated risks to achieve more—for example, taking the risk to start one’s own business or taking the risk to apply to a high-status graduate program. Women may be disadvantaged in situations like this if they are less willing to take the leap.


Psychologists have debated whether gender differences in activity level exist. Certainly, if you ask the average parent or teacher, they will tell you that boys are more active, and most child psychology textbooks share this view.

A meta-analysis found that d was approximately 0.50—that is, that there is a moderate gender difference, with boys and men having the higher activity level (Eaton & Enns, 1986). Among infants, d = 0.29; it was 0.44 for preschoolers and 0.64 for older children and young adults. Thus a small difference is present from infancy, and the difference gets larger with age, at least among children. This meta-analysis was based on samples of children from the general population. In samples of hyperactive children, about 80% to 90% are boys (Biederman et al., 2002).

What causes this gender difference, and why does it get larger from infancy to childhood? One possibility is that the small difference in infancy is magnified by social interactions, especially when boys increasingly play actively with other boys and not with girls, something called the gender segregation effect (discussed in Chapter 7). Essentially, boys egg each other on to more and more active play. Another possibility has to do with the developmental precocity of girls. Girls are somewhat ahead of boys in development, including brain development. As children grow older, they learn to control their activity more. It might be, then, that the lower activity level of girls actually represents a greater ability to control activity because of their being somewhat more mature than boys.


Popular best sellers like Mary Pipher’s (1994) Reviving Ophelia have spread the word that girls have major self-esteem problems beginning in early adolescence—and, by implication, that boys do not. Again, we have a meta-analysis available to speak to this issue (Kling et al., 1999; see also Major et al., 1999).

Self-esteem: The level of global positive regard that one has for oneself.

Averaged across all samples, the average effect size was d = 0.21. Male participants scored higher, on average, but the difference was small—certainly not the large gender gap that one would expect from the popular press.

Pipher and others have also argued that the pattern of gender differences changes developmentally with age. Elementary school girls may have self-esteem equal to that of boys, but the problems begin in early adolescence. To test this hypothesis, effect sizes were computed by age-group in the meta-analysis. The results showed that, for elementary school children (ages 7 to 10), d = 0.16; for middle school children (ages 11 to 14), d = 0.23; and for high schoolers, d = 0.33. That is, in early adolescence the gender difference is still small, and it grows larger in high school. Interestingly, for adults between the ages of 23 and 59, d = 0.10, and for those who are 60 and over, d = 0.03. In other words, the gender difference is close to zero in adulthood.

What about the intersection of gender and ethnicity? Is this gender difference found in all U.S. ethnic groups? For White samples the effect size was 0.20, whereas for Black samples it was —0.04. (For other ethnic groups, too few studies were available to compute effect sizes.) These results are a vivid illustration of the ways in which psychology has been a psychology of White people. The much-publicized gender difference in self-esteem is present for White Americans (and even then, it is small), but it is not found for Black Americans.

In sum, boys and men on average score higher on self-esteem measures, but the gender difference is small and may be true only for White Americans. The gender difference is tiny in the elementary school years and largest in the high school years, but even then it is not huge. On the other hand, it probably is large enough to be concerned about, in which case at least two of Mary Pipher’s (1994) explanations ring true—that girls in the United States are subjected to a sex-saturated media environment that objectifies them and undermines their self-esteem and that they are victims of peer sexual harassment. Adolescent girls deserve a better environment.

To this point we have been talking about general self-esteem, one’s global self-evaluation. We can also talk about domain-specific self-confidence—that is, confidence in specific areas such as athletics or academics. Another meta-analysis examined gender differences in domain-specific self-confidence (Gentile et al., 2009). The results indicate that male participants had more self-confidence in the domains of physical appearance (d = 0.35) and athletics (d = 0.41). Female participants had more self-confidence in the areas of behavioral conduct (d = —0.17) and morals or ethics (d = —0.38). At the same time, gender similarities were found in the areas of academics and social acceptance. If we consider an even more specific domain, math self-confidence, boys score higher than girls in the United States (d = 0.26) and in most other nations, with exceptions in a few places such as Eastern Europe (specifically, Estonia and Russia; Else-Quest et al., 2010).

Self-confidence: A person’s belief that they can be successful at a particular task or in a particular domain such as athletics or academics.

These gender differences in self-confidence can be important in people’s lives. People with low self-confidence avoid engaging in challenging tasks. Thus this gender difference in math self-confidence may have important effects on women’s career choices, a point to be discussed further in Chapter 8.

Before leaving the topic of self-confidence, we need to consider one additional issue, namely, interpretation of the results. The objective, statistical result is that, for example, male students are more likely than female students to rate themselves at the top of their class in math ability. To use the terminology of Chapter 1, a typical interpretation of this result is that boys have more math self-confidence than girls or that girls are lacking in math self-confidence. This is a female-deficit interpretation. Would it be possible to make a different interpretation that would still be consistent with the data? An alternative interpretation is that boys’ estimates are too high (rather than girls’ being too low) and that boys are unrealistically overconfident. This alternative is just as reasonable an interpretation of the gender difference, but it implies a problem for boys. As it turns out, with tasks such as these it is possible to decide which interpretation is more accurate, because we can find out how students actually did on the math exam or in the course. In fact, boys and men do tend to overestimate their performance by about as much as girls and women underestimate theirs, although some studies find female participants’ estimates to be accurate and male participants’ to be inflated (Beyer, 1999; Cole et al., 1999). Therefore, there is some truth in each interpretation—men are probably somewhat overconfident and women somewhat underconfident.

Helping Behavior

Social psychologists have studied helping behavior extensively. Who do you think is more likely to help another person, women or men? A meta-analysis of studies of gender differences in helping behavior found that d = 0.34 (Eagly, 2009; Eagly & Crowley, 1986). The positive value indicates that men, on average, helped others more than women did and that the gender difference is somewhere in the small to moderate range. This finding may be somewhat surprising because helping or nurturing is an important part of the female role. In Table 3.1, caring is one of the stereotypes about women. To probe the findings more deeply, Alice Eagly and Maureen Crowley (1986) examined the kinds of situations that produced more helping by men and those that produced more helping by women. They noted that some kinds of helping are part of the male role and some are part of the female role. Helping that is heroic or chivalrous—rescuing your comrade injured in battle—falls within the male role, whereas nurturance and caretaking fall within the female role.

Focus 3.1 Who Is More Narcissistic: Men or Women?

Narcissism has been in the headlines a great deal recently, with the 2016 presidential election and with claims that millennials are a narcissistic bunch. Are there gender differences in narcissism?

Generally narcissism refers to a personality trait characterized by an excessive focus on oneself, along with a grandiose, exaggerated sense of one’s own talents, an extreme need for admiration, and a lack of empathy for others. Who would we predict would be more narcissistic, men or women?

Narcissism: A personality trait characterized by an excessive focus on oneself, along with a grandiose, exaggerated sense of one’s own talents, an extreme need for admiration, and a lack of empathy for others.

Using social theory (discussed in Chapter 2), we would consider which aspects of gender roles would be associated with narcissistic traits. Men are expected to be very self-confident, and arrogant is one of the male stereotypes shown in Table 3.1, so that would lead men to have an exaggerated sense of their own talents. Caring and being empathic are part of the female role, so men would be likelier to lack empathy. In general, then, the male role is more consistent with narcissistic traits than the female role is.

A meta-analysis found that men are more narcissistic, but only by a small amount, d = 0.29 (Grijalva et al., 2015). But it turns out that personality inventories include three facets of narcissism. For the Exploitative/Entitlement scale, d = 0.29, men scoring higher, and for Leadership/Authority, men also score higher, d = 0.20. However, for Grandiose/Exhibitionism, d = 0.04, that is, there is no gender difference. There is also a newly discovered type of narcissism, vulnerable narcissism, which is characterized by low self-esteem, neuroticism, and introversion, all of which involve an excessive focus on the self. For this type of narcissism, there is also no gender difference, d = —0.04. Overall, then, we see gender similarities in some aspects of narcissism, but men scoring higher on the Exploitative/Entitlement facet.

The studies in the meta-analysis involved samples of the general population and a broad range of scores on personality scales. If we look at the extreme cases of diagnosable narcissistic personality disorder, men outnumber women, 7.7% to 4.8% (Stinson et al., 2008).

Why is narcissism important? A certain amount of narcissism probably helps propel people into leadership roles, which is advantageous. However, more extreme narcissism of the Exploitative/Entitlement type, where men score higher, can be disadvantageous or even disabling. The evidence indicates that people who are high on Exploitative/Entitlement are more likely to engage in antisocial behaviors at work, suffer poor adjustment to college, and experience poor relationship satisfaction (Grijalva et al., 2015). These are not outcomes that are desirable for the individual or for those who interact with them.

Consistent with these predictions from an analysis of social roles, Eagly and Crowley (1986) found that the tendency for men to help more was especially pronounced when the situation might involve danger (such as stopping to help a motorist with a flat tire). The tendency was also stronger when the helping was observed by others (rather than when the person needing help and the research participant were alone together). Helping that involves danger and that carries with it a crowd of onlookers has great potential for heroism, and that kind of helping is part of the male role.

The plot thickens, because social psychologists have spent most of their time studying precisely these kinds of helping behaviors—the ones that occur in relatively short-term encounters with strangers. They have devoted little research to the kind of caretaking and helping that is characteristic of the female role—the kind of behavior that more often occurs in the context of a long-term relationship, such as a mother helping her child. Therefore, the gender difference found in the meta-analysis, showing that men help more, is probably no more than an artifact of the kinds of helping that psychologists have studied and the kinds of helping that they have overlooked.

Photo 3.2 Research shows that which gender helps more depends on the situation or context.


© & ©

These results are consistent with a pattern: Gender differences are highly dependent on the situation or context in which they are observed (Zakriski et al., 2005).


Most studies show that girls and women are more fearful and anxious than boys and men, although the difference is not large. One large, cross-national study found d = —0.38 for self-reports of general anxiety (Löckenhoff et al., 2014). Because this study was based on self-reports, what we know is that girls and women are more willing to admit that they have anxieties and fears. It is possible that these self-reports reflect higher levels of anxiety in female respondents than in male respondents. But it is also possible that men experience the same levels of anxiety as women and that women are just more willing to admit them. Women are stereotyped as being more anxious than men (Löckenhoff et al., 2014). This stereotype would encourage women to admit their feelings and men to pretend not to have them. At this point, however, studies have not been able to resolve the issue.

The Gender Similarities Hypothesis

Our culture is dominated by the differences model, the belief that women and men are very different from each other. As one best-selling book title has it, Men Are From Mars, Women Are From Venus (Gray, 1992)—men and women are so different, it’s like they are from different planets. Then there’s Boys and Girls Learn Differently! A Guide for Teachers and Parents (Gurian, 2011), which has been read and taken seriously by thousands of parents and teachers.

Yet the scientific evidence from meta-analyses leads to a very different conclusion: men and women are actually quite similar. The gender similarities hypothesis states that men and women are similar on most, but not all, psychological variables (Hyde, 2005a). That is, women and men are more similar than they are different.

Gender similarities hypothesis: The hypothesis that men and women are similar on most, but not all, psychological variables.

Evidence for the gender similarities hypothesis comes from a review of 46 meta-analyses of psychological gender differences (Hyde, 2005a; for an update with the same conclusion, see Zell et al., 2015). Across those meta-analyses, 30% of the effect sizes were close to 0 (d ≤ 0.10) and an additional 48% were small (d between 0.11 and 0.35). That is, 78% of the gender differences were small or smaller. We have seen some of these patterns of gender similarities already in this chapter. For example, self-esteem is viewed, in the culture, as showing a large gender difference, with girls and women having low self-esteem, yet a meta-analysis shows the gender difference to be small. We will continue to see other evidence of gender similarities in other chapters of this book.

The original statement of the gender similarities hypothesis did note that there are a few exceptions. One is aggressive behavior, where the gender difference is moderate in size, although not large. Another exception is some aspects of sexuality. We will consider these findings in Chapter 12.


Beginning in the 1970s, feminist psychologists sought to create new models of human behavior that would overcome gender stereotypes. One prominent alternative that emerged was androgyny, the combining of masculine and feminine characteristics in an individual (Bem, 1974). But before that, there was a history of several decades of psychologists measuring masculinity—femininity, so let’s consider that first.

Androgyny: The combination of masculine and feminine psychological characteristics in an individual.

Psychologists’ Traditional Views of Masculinity—Femininity

Psychologists’ traditional view—beginning roughly in the 1930s—was that masculinity and femininity were at opposite ends of a single scale. We would call that a unidimensional (one-dimensional) bipolar (two opposite ends) continuum. It is shown in Panel 1 of Figure 3.2.

Figure 3.2 Progressive conceptualizations of masculinity—femininity.


Source: Created by the authors.

One of these traditional scales was part of the California Psychological Inventory (Gough, 1957). It is simply a test on which you respond true or false about yourself to a series of items such as “I am somewhat afraid of the dark.” Then a score is computed, over all the items, that places you at some point along the bipolar continuum.

Items were chosen for this scale in a simple way: They showed gender differences, with a much different percentage of men and women responding true to each one. Women are somewhat more likely to say true to “I am somewhat afraid of the dark,” so that’s why the item is on the scale. The implicit assumption, then, is that “femininity” is the quality of women that differentiates them from men.

In the 1970s, feminist psychologists raised a number of criticisms of masculinity—femininity scales (e.g., Constantinople, 1973). The basic issue is whether femininity and masculinity are really opposites of each other, so that the more feminine a person is, the less masculine they are. Is it possible to be both strongly feminine and strongly masculine? Of course it is. Enter androgyny.

The Concept of Androgyny

Most of us know people who have both masculine and feminine qualities. An example would be a woman who is both very nurturing to her children and a high-status leader on the job. The research on androgyny was designed to acknowledge and describe such people.

Androgyny means having both masculine and feminine psychological characteristics. It is derived from the Greek roots andro, meaning “man,” and gyn, meaning “woman” (as in gynecologist). As shown in Figure 3.2, Panel 2, the concept of androgyny is based on a two-dimensional model of masculinity—femininity, in contrast to the traditional one-dimensional model. One of the dimensions is femininity, ranging from low to high, and the other is masculinity, ranging from low to high. With this conceptualization, a person could have a high score on both femininity and masculinity, and would therefore be androgynous. In Figure 3.2, those people would fall in the upper right quadrant.

Measuring Androgyny

Psychologist Sandra Bem (1974) constructed a test to measure androgyny (see also Spence & Helmreich, 1978). It consists of 60 adjectives or descriptive phrases. Respondents are asked to indicate, for each, how well it describes them on a scale from 1 (never or almost never true) to 7 (always or almost always true). Of the 60 adjectives, 20 are stereotypically feminine, 20 are stereotypically masculine, and 20 are neutral filler items—that is, not gender typed. In Table 3.4 we show only the stereotypically feminine and masculine items, because only they are relevant to computing scores. Items 1, 3, 5, and so on are masculine, and items 2, 4, 6, and so on are feminine.

How did Bem choose feminine items? She did so on the basis of characteristics that are considered socially desirable for women in our culture, and similarly for masculinity. She did this by asking a sample of people to list qualities that are socially desirable for women and qualities that are socially desirable for men.

A person who scores high (above the median) on both the masculinity and the femininity scales is classified as androgynous. A person who scores high on femininity but low on masculinity would be categorized as feminine. Someone who scores low on femininity and high on masculinity would be categorized as masculine. And someone who scores low on both scales would be called undifferentiated.


Source: Adapted from Bem (1974, 1977) and Hyde & Phillis (1979).

Criticisms of Androgyny

Androgyny as a concept and a scale was an advance over previous masculinity—femininity scales because it acknowledged the possibility that people might be high on both masculinity and femininity. Nonetheless, criticisms can be raised. The major one is that the scale is now over 40 years old and cultural ideals of masculinity and femininity have changed a lot over those decades. We can’t really know whether the same items would describe femininity and masculinity today.

Androgyny and Transgender

How does the concept of androgyny relate to the transgender spectrum and to the experience of trans or genderqueer individuals (Halberstam, 2012)? According to one view, the androgyny of the 1970s might be seen as a forerunner of the concept of genderqueer today. That is, the androgynous individual blends masculinity and femininity as the genderqueer individual does. To dig a bit deeper, though, androgyny is really about personality traits and behaviors (independent, forceful, and so on), whereas genderqueer and transgender are about identity (I identify as a man, or I identify as a woman, or I identify as neither or both). Most people who score as androgynous on the Bem scale have definite cisgender identities. At that point, the link between androgyny and genderqueer breaks down.

Experience the Research: How Accurate Are People’s Beliefs About Gender Differences?

Ask four people you know to provide you, individually, with some data. When you interview them, tell them that you want to determine how accurate people are in estimating the size of some psychological gender differences. Have them fill out the following form, explaining what they are to do. Be sure that they understand that they can give any number they want; for example, they do not have to answer just 0.20 or 0.50, but can give an answer like 0.35. Also be sure that they understand the importance of the difference between negative numbers and positive numbers. Negative numbers on this scale mean that women score higher than men, and positive numbers mean that men score higher than women.

1. Aggressive behavior among preschoolers


Your estimate: ______

2. Performance on math tests by elementary school children


Your estimate: ______

3. Approval of casual sex (i.e., two people engaging in sexual intercourse when they are only casually acquainted)


Your estimate: ______

How accurate were your respondents? Meta-analyses show that, among preschoolers, boys are more aggressive, d = 0.58. For math performance by elementary school children, d = 0.0; that is, there is no gender difference. For approval of two people engaging in sexual intercourse when they are only casually acquainted, d = 0.45; men are more approving (see Chapter 12).

Source: Based on Swim (1994). Copyright © by the American Psychological Association.

Chapter Summary

Gender stereotypes are a set of shared cultural beliefs about the characteristics of men and women. People stereotype for two reasons: comprehension and self-enhancement. Implicit gender stereotypes are learned automatic associations between a gender category (e.g., female) and other attributes (e.g., mathematics). An intersectional approach considering the intersection of gender and ethnicity finds that gender stereotypes vary across U.S. ethnic groups.

Stereotype threat occurs in a situation in which there is a negative stereotype about a person’s group and the person is concerned about being judged or treated negatively because of the stereotype. Stereotype threat can hurt women’s math performance and can doubly hurt Latinx women’s math performance.

Little research is available on stereotypes about trans individuals, yet it is clear that there is prejudice against them. Psychologists have begun to design interventions to reduce anti-trans prejudice.

Meta-analysis is a statistical method for synthesizing the results of numerous studies on a particular question (e.g., gender differences in aggressive behavior). The effect size, d, measures how large a gender difference is. A d value of 0.20 is small, 0.50 is moderate, and 0.80 is large.

Gender differences are found in aggressive behavior, with boys and men being more aggressive; d = 0.55 for physical aggression. A deindividuation experiment showed that gender differences in aggression can be erased, though, if the force of gender roles is removed.

In regard to impulsivity, boys and men are higher in risk taking, d = 0.38. Boys and men also are higher in activity level, and the gender difference grows larger with age, from d = 0.29 in infancy to d = 0.64 in older children and young adults. Boys also account for 90% of hyperactive children.

Despite stereotypes that women are the ones with low self-esteem, d = 0.21, meaning that men have higher self-esteem but the difference is small. From an intersectional perspective, the gender difference is found for U.S. White samples, but not Black samples.

Research on gender differences in helping behavior provides evidence of the ways in which gender differences depend on the situation and context. Men help more in situations that involve danger, whereas women help more in situations that involve nurturing in the context of a long-term relationship.

Self-report studies find that women and girls are more anxious than men and boys are by a small to moderate amount.

The gender similarities hypothesis states that men and women are similar on most, but not all, psychological variables. Across 46 meta-analyses of gender differences, 78% of the effect sizes were small or very close to zero.

Androgyny refers to a combination of masculine and feminine characteristics in an individual. It is measured using a two-dimensional scale, with one dimension measuring femininity and the other measuring masculinity. In considering the relationship between being androgynous and being genderqueer, androgyny refers to personality characteristics and behaviors, whereas genderqueer refers to identity.

Suggestions for Further Reading

Hyde, Janet S. (2005). The gender similarities hypothesis. American Psychologist, 60, 581—592. This article represents the original statement of the gender similarities hypothesis and is accessible reading for college students.

Tompkins, Tanya L., Shields, Chloe N., Hillman, Kimberly M., & White, Kadi. (2015). Reducing stigma toward the transgender community: An evaluation of a humanizing and perspective-taking intervention. Psychology of Sexual Orientation and Gender Diversity, 2, 34—42. This article reports one of the first interventions designed to reduce anti-trans prejudice (cisgenderism). Because it is an early study, some aspects of it could be improved. How would you improve it?