In the 2004 U.S. presidential election, a majority of men (54%) voted to reelect George W. Bush, but a minority of women (48%) supported Bush’s reelection. Four years earlier, the gender gap in presidential vote choice was even wider: Women were 10 percentage points less likely than men to vote for George W. Bush. The gender gap in presidential voting was first identified by the news media in the 1980 contest between Ronald Reagan and Jimmy Carter when 54% of men and 46% of women voted for Reagan (Borquez, Goldenberg, & Kahn, 1988). In each year since 1980, the gender gap in vote preference has been apparent, but it has varied in size from about 4 percentage points in 1992 to about 11 percentage points in 1996 (Center for the American Woman and Politics, 2004).

The gender gap has been identified in lower level races as well. There was a 7% gender gap in U.S. House races in 2004; 52% of women and 45% of men voted for the Democratic congressional candidate in their district. Likewise, in the U.S. Senate races in 2004, 54% of women and 46% of men voted for the Democratic senatorial candidate.Footnote 1

Numerous scholars have also identified a gender gap in policy preferences and priorities. For example, men and women differ in their views about the necessity of war and the use of force to resolve foreign conflicts (Conover & Sapiro, 1993; Hurwitz & Smithey, 1998). In addition, women are more likely than men to support higher increases in spending for social welfare, education, health care, and the environment (Chaney, Alvarez, & Nagler, 1998; Howell & Day, 2000). Finally, women are more pessimistic about the likelihood of a terrorist attack, they are less satisfied with current gun laws, and they are less sanguine about the moral and ethical climate of the country (Saad, 2003).

We should point out that some scholars have claimed that gender differences in political attitudes and behavior have been exaggerated in the literature (see, for example, Bourque & Grossholtz, 1974; Sapiro, 2003). For example, Sapiro (2003) explained that literature reviews of the gender gap tend to offer an unrepresentative view by “cherry-picking” findings of differences and ignoring findings of similarities.

Nonetheless, the bulk of the evidence on voting preferences and policy preferences indicates that men and women currently view politics in the United States differently. What factors help to explain these differences? Three principal explanations have been offered to explain men’s and women’s different political views. First, the gender gap may be caused by differences in the economical vulnerability of men and women. Women, for example, continue to earn less than men do, and women are more likely than men to live in poverty. Because women are more economically vulnerable, they are more likely than men to be dependent on the welfare state. Therefore, women are more supportive of the welfare state and more compassionate toward people who are hurt by economic downturns. Box-Steffensmeier, De Boef, and Lin (2004) found some support for this hypothesis as they demonstrated that the gender gap grows as the economy deteriorates and economic vulnerability increases. Nonetheless, when researchers control for economic vulnerability, some gender differences persist, especially differences in the use of force, views about morality, and attitudes toward gun control regulations (Chaney et al., 1998).

A second explanation for the gender gap offered by scholars rests on the notion that women are more likely than men to adopt a feminist ideology that rejects hierarchy, domination, and the use of force (Conover, 1988; Shapiro & Mahajan, 1986; Tolleson-Rinehart, 1992). Adherence to this ideology may help explain to gender differences in views about the welfare state as well as views about the use of force (Conover, 1988; Conover & Sapiro, 1993; Howell & Day, 2000). Again, some evidence has been offered to support feminist explanations of the gender gap. For example, Conover (1988) found that feminist women are less supportive then men are of military intervention and are more supportive than men are of increased spending on the elderly, food stamps, child care, and affirmative action programs for women. Furthermore, feminist women and non-feminist women differ in their support for conventional war, their support for increased spending on defense, and their support for government welfare programs (Conover, 1988). However, not all of the gender differences in political views can be explained by feminist ideology. For example, non-feminist women and men differ in their willingness to increase spending for the elderly, for medicare, and for the poor (Conover, 1988). Therefore, adherence to a feminist ideology only partially explains gender differences in political views.

The final explanation for the gender gap in political views draws on differences between girls and boys and women and men in terms of socialization and experiences. For example, some scholars have argued that women do more parenting than men do and that their experiences as mothers make them more empathetic, more caring, and more attentive in their relationships (Chodorow, 1978; Conover, 1988; Conover & Sapiro, 1993). This mothering experience leads women to be more caring toward disadvantaged groups which leads to greater support for welfare programs.Footnote 2

The evidence to support the socialization explanation for the gender gap is largely indirect. In particular, when researchers have found a residual gender gap after examining the impact of economic vulnerability and feminism, as well as other factors, they have often concluded that the residual gap must be a function of differences in socialization and experiences of men and women (see, for example, Chaney et al., 1998).

In the present study, we utilized a more direct test of the socialization explanation for the gender gap. In particular, we examined the political views of eighth-grade boys and girls. If boys and girls differ in their views about policy and parties at an early age, these differences probably reflect gender differences in socialization patterns. It is unlikely, for example, that eighth-grade boys and girls differ systematically in their economic vulnerability. Similarly, 13 and 14 year old boys and girls are unlikely differ in their adherence to a feminist ideology because it is unlikely that adolescents are cognitively sophisticated enough to have developed ideological views. Instead, if differences emerged in the political predispositions of boys and girls, we can conclude that, at least for the moment, these differences are caused by gender socialization experiences.

Hooghe and Stolle (2004) used a similar method to explain gender differences in political participation. In particular, Hooghe and Stolle studied 14-year-olds as a way of testing the claim that gender differences in resources explain differences in adult participation rates. Hooghe and Stolle argued that gender differences in the availability of resources, such as time, money, and cognitive skills, are unlikely among 14-year-olds. Therefore, if gender differences in anticipated political participation do not appear among these students, “...this strengthens explanations that rely on resource arguments to explain gender differences in [adult] political participation levels” (Hooghe & Stolle, 2004, p. 3). As Hooghe and Stolle expected, they found no evidence of a gender gap in participation among adolescents.

A large number of political socialization scholars have examined the gender gap in political interest and political knowledge among children as young as fourth-grade (e.g., Greenstein, 1965; Hess & Torney, 1967). For example, Greenstein (1965), in his study of fourth to eighth graders in New Haven, found that boys were more informed than girls about politics. Similarly, Hess and Torney (1967), in their study of third to eighth graders, found that girls held a more immature picture of political authorities than boys did. However, more recent studies of adolescents have not shown young women to be less interested or less knowledgeable about politics than young men are (e.g., Hahn, 1998; Hepburn & Napier, 1982–1983; Torney-Purta, Lehmann, Oswald, & Schulz, 2001).

In our study, we examined the political preferences of eighth graders. We focused on eighth graders for several reasons. First, eighth-grade is the formal beginning of civics training; the vast majority of middle schools require civics courses in eighth-grade (Niemi & Hepburn, 1995). In addition, in eighth-grade, students are beginning to develop their views toward politics, although few students have crystallized attitudes about politics (e.g., Sears & Valentino, 1997; Valentino & Sears, 1998; for an exception see Torney-Purta et al., 2001). Finally, eighth-grade students are starting to discover their political and self-identities and are beginning to understand their roles in their communities (Hooghe & Stolle, 2004; Tolo, 1998).

Theory and Hypotheses

There is an extensive body of literature on the process of gender role socialization (for a review, see Ruble & Martin, 1998). A key feature of gender role socialization is encouraging girls to be connection-oriented, to focus on goals of maintaining social relationships, and to be empathetic with others (i.e., communal), whereas boys are encouraged to develop an agentic style emphasizing status, independence, and dominance (e.g., Eagley & Mitchell, 2004; Hibbard & Buhrmester, 1998; Hutchings, Valentino, Philpot, & White, 2004; Lang-Takac & Osterweil, 1992; Moskowitz, Suh, & Desaulniers, 1994).

The communal dimension of gender role socialization is similar to the “ethic of caring” discussed by Gilligan (1982). As Gilligan and others have pointed out, girls’ socialization experiences stress connections in relationships and promote concern for others from early childhood (Chodorow, 1978; Gilligan, 1982). Girls are encouraged to identify with and to be dependent on their mothers. Therefore, girls and women tend to value strong nurturing relationships with others, and they are more likely to internalize a responsibility to care for others (Hutchings et al., 2004). Boys, on the other hand, are encouraged to be independent and autonomous (Chodorow, 1978), and are expected to be assertive and aggressive in their relationships (Hurwitz & Smithey, 1998).

We hypothesized that these role orientations influence gender differences in political orientations. For example, girls are encouraged to value nurturance and caring for others (e.g., Eagley & Mitchell, 2004). Therefore, girls may be more likely to want to protect society’s most vulnerable citizens which would make girls more supportive of social programs aimed at helping others, such as welfare, health care, and education (Hutchings et al., 2004).

Boys, on the other hand, because they are socialized to value independence, may place less value on government programs aimed at helping others. Similarly, boys are often taught that aggression is an appropriate means of securing instrumental rewards (Carlo, Raffaelli, Laible, & Mayer, 1999; Ruble & Martin, 1998). Therefore, boys may be more supportive than girls of aggressive policies to solve programs (e.g., war to resolve international conflicts).

Girls may be more supportive than boys of the Democratic Party because the Democratic Party is more closely associated with maintaining or increasing funding for government programs aimed at helping the disadvantaged. Boys, on the other hand, may be more attracted to the Republican Party because the Republican Party has actively tried to curtail funding of social programs since the Reagan presidency. In addition, girls may be less attracted to the Republican Party because they may view Republicans as more supportive of military solutions to international problems given contemporary events (Barone & Cohen, 2004).

Finally, we expected that boys and girls would differ in their views regarding women’s role in politics. Previous studies of young adults have shown that women are more equalitarian than men in their views regarding women’s role in the political arena (e.g., Baldi, Perie, Skidmore, Hahn, & Greenberg, 2001; Hahn, 1998; Jennings, 2006; Torney-Purta et al., 2001).

Why might boys and girls differ in their views regarding women’s place in politics? To begin, the discussion of women’s role in politics continues to be largely absent in today’s civics curriculum. In civics courses, teachers rarely discuss barriers to women’s holding office, women politicians are rarely mentioned in these classes, and there is little (if any) discussion of women’s role in grassroots movements (Hahn, 1998). Furthermore, research has shown that the news media, especially newspapers and news magazines, differ in their treatment of men and women politicians (Woodall & Fridkin, 2006). In particular, women politicians receive less press attention and more negative press coverage than men do. In addition, the news media are more likely to focus on women politicians’ personal life, marital status, appearance, and personality in their coverage (Woodall & Fridkin, 2006).

In sum, sexist messages continue to permeate the academic and media environment of children. Girls, however, may be more likely than boys to reject messages regarding gender inequality. Furthermore, girls may be more likely than boys to attend to and remember positive messages about women in politics, even if these messages are not as prevalent. Recent research has shown that women and girls differ from men and boys in their attention to women in politics. For example, Campbell and Wolbrecht (2006) found that the presence of visible female candidates was associated with increases in the anticipated political involvement of girls, but not boys. Similarly, Atkeson (2003), who studied adults, found that the presence of prominent female candidates had a positive effect of the political engagement of women, but had no effect on the engagement of men (see also Hansen, 1997; Koch, 1997; Verba, Burns, & Schlozman, 1997).

Gender differences in views about women’s roles in politics may also be driven by gender differences in media exposure. Boys and girls use the media differently. Therefore, they are exposed to different messages. For instance, boys are much more likely than girls to play video games (Roberts, Foehr, Rideout, & Brodie, 1999; Woodard & Gridina, 2000). Dietz (1998) found that the portrayal of women characters, when they appear in video games, is largely sexist. Women are most often portrayed as sex objects, as victims, and as contributing less than men.

Boys and girls also differ in the music they prefer (Roberts et al., 1999). Boys, for example, are more likely than girls to listen to rap and R&B radio stations. Girls, on the other hand, listened more often to Top 40 stations and country stations. Further, the lyrics in these genres differ markedly; rap music portrays women negatively as “ho’s” and “bitches,” and lyrics often depict women as victims of violence (Armstrong, 2001). In sum, because boys and girls attend to different media messages and because girls are more likely to reject the sexist messages prevalent in the today’s media and academic environment, we expected boys and girls to differ in their views about women in politics.

Method

Participants

We interviewed eighth-grade students from six middle schools in Maricopa County, AZ in the springs of 2003 and 2004. This study was part of a larger study of civic engagement. To select the schools to be included in this study, we stratified middle schools by school district and randomly selected one middle school from each school district.

Maricopa County is a large metropolitan area (population approximately 3.8 million) and is the fourth most populous county in the nation. On several important dimensions, the demographic characteristics of Maricopa County mirror the demographics of the United States. For example, 75% of the population of the U.S. are European American and 77% of the population of Maricopa county are European American; the mean per capita income for U.S. residents is $21, 587, whereas the mean per capita income for Maricopa County is $22,251; the percentage of high-school graduates is 80% for the U.S. and 82% for Maricopa County; 18% of people in the U.S. and 26% of Maricopa County residents speak a language other than English in their homes; 14% of the U.S. and 12% of the population of Maricopa County live below the poverty level.Footnote 3

The sample consisted of 439 eighth-grade students. Fifty-three percent of the students who completed the survey were girls, and 47% were boys. In addition, 40% of the students who completed the survey were European Americans, 35% were Latina/Latino, 14% were Native American, and 9% were African American. The remaining 7% of the students were coded as “other” in terms of their racial and ethnic background.

Measures

Self-report questionnaires were used to gather information about students’ political attitudes. The questionnaire assessed students’ views on women in politics, priorities about issues, and political party preferences.

Political party identification

To measure political party identification, we relied on the following American National Election Study measure (American National Election Studies, 2005): “Generally speaking, do you usually think of yourself as a Republican, a Democrat, an Independent, or what? In answering the question, please place yourself on following scale: Strong Republican; Weak Republican; Independent, but leaning to Republican Party; Independent; Independent, but leaning to Democratic Party; Weak Democrat; Strong Democrat; Other.”

Political spending priorities

We borrowed three measures from the American National Election Study (ANES, 2005) to assess political spending priorities” (1) “Tell me if you would like to see increased government spending for health care, even if it meant paying more taxes, if you would like to see it decreased, or if you would leave it the same”; (2) “Tell me if you would like to see increased government spending for public education, even if it meant paying more taxes, if you would like to see it decreased, or if you would leave it the same”; (3) “Tell me if you would like to see increased government spending for the War on Terrorism, even if it meant paying more taxes, if you would like to see it decreased, or if you would leave it the same.” To answer these questions, students were given the following alternatives: increasing spending, keep spending the same, decrease spending.

Attitudes toward women’s role in politics

We assessed views about women’s roles in politics with the following three item index borrowed from the General Social Survey (Davis, Smith, & Marsden, 2002): (1) “Most men are better suited emotionally for politics than most women”; (2) “Women should run for public office and take part in the government just as men do”; (3) “Men are better qualified to be political leaders than women.” Participants responded to each statement on a 5-point Likert scale (1 = agree strongly to 5 = disagree strongly).

Procedure

Social studies teachers in the six participating schools gave each student a survey to fill out during their class after the teachers had secured permission from the students’ parents. Teachers in each school allocated sufficient time for each student to complete the survey. The survey took an average of 30 min to complete. However, the amount of time necessary to complete the class survey varied because students’ literacy levels varied-especially across schools.

We offered students an incentive to participate; those students who completed questionnaires received a $5.00 movie-pass. This incentive, according to each of the teachers who helped with our study, was an effective inducement for participation. Overall, completion of the survey ranged from 75 to 90% across the six schools.

Results

Do boys and girls differ in their views of politics? We began by examining the political party identification of eighth-grade students. We found that 57% of the students provided an answer to the party identification question. Boys and girls did not differ in their willingness to identify their party affiliation. However, boys and girls did differ in their party preferences. The data in Fig. 1 reveal that boys were 10% more likely than girls to identify with the Republican Party. Similarly, girls were significantly more likely than boys to identify with the Democratic party; 26% of the girls identified with the Democratic Party, whereas only 13% of the boys identified did. These differences are statistically significant, F(1, 296) = 2.70, p < .10.

Fig. 1
figure 1

Gender differences in party identification in eighth graders.

Full size image

Furthermore, the gender gap in party preference persisted when we looked at European Americans and minority students. For example, among European American students, girls were almost twice as likely as boys to identify with the Democratic party (17 vs. 9%), and, among minority students, girls also were more likely than boys to identify with the Democratic party (24 vs. 15%). These data suggest, at least in metropolitan Phoenix, that the gender gap in party identification is established early—4 years before students have reached voting age.

In addition to party differences, we also hypothesized that gender role socialization would lead boys and girls to differ in their policy priorities. We asked students to indicate whether they would favor increasing or decreasing spending for a series of programs, including education, health care, protecting the environment, and fighting the “war on terrorism.”

According to the data presented in Fig. 2, eighth-grade boys and girls had significant and substantive differences in their policy preferences. Across each of the social programs, girls were more likely than boys to favor increased federal spending. These differences are statistically significant for two of the three issues: the environment, F(1, 370) = 7.60, p < .01, and education, F(1,421) = 2.81, p < .10. For example, almost 60% of the girls thought that spending to protect the environment should be increased, whereas only about 40% of the boys favored increased spending for the environment. Also, as expected, boys were significantly more likely than girls to favor increased spending for the war on terror, F(1, 404) = 17.67, p < .01. Only one-third of girls supported spending more on fighting terrorism, whereas more than 40% of boys agreed with increased funding in this area.

Fig. 2
figure 2

Gender differences in policy views among eighth graders.

Full size image

The gender differences in spending priorities persisted when we controlled for the ethnic and racial background of the students. For example, among European Americans, 43% of boys and 33% of girls favored increased spending for the war on terror. This gender difference was also found among minority students; 42% of the minority boys endorsed increased spending on terrorism, whereas only 32% of minority girls endorsed increase spending.

Finally, we looked at whether boys and girls differed in their views about women’s roles in politics. As expected, we found a substantial gender gap in students’ views about women’s roles in politics. We found statistically significant gender differences across each of the three questions that assessed attitudes toward women in politics: (1) men are better suited emotionally for politics F(1, 404) = 79.56, p < .01,; (2) women should run for public office, F(1, 441) = 59.67, p < .01; and (3) men are better qualified to be political leaders F(1, 434) = 150.48, p < .01.

The significant gender difference in views toward women is illustrated in Fig. 3. For instance, 34% of the eighth-grade boys agreed that “most men are better suited emotionally for politics than most women,” whereas only 12% of the eighth-grade girls agreed. Similarly, 32% of boys agreed that “men are better qualified to be political leaders than women,” whereas only 6% of eighth-grade girls agreed with that statement.

Fig. 3
figure 3

Gender differences in views about women in politics among eighth graders.

Full size image

The gender differences concerning views of women’s role in politics did not disappear when we looked within ethnic and racial groups. For example, 95% of the European American girls agreed with the statement that “women should run for public office and take part in government just as men do,” whereas only 76% of the European American boys agreed with the statement. Among minority students, 97% of the girls agreed that women should run for public office, whereas only 71% of the boys agreed.

Discussion

The gender gap in adults’ views about politics has been a fixture in American politics since, at least, the 1980s. However, the explanation as to why the gender gap exists remains a source of controversy. In the present study, by examining the political views of eighth-grade boys and girls, we eliminated two of the three primary explanations: gender differences in economic vulnerability and gender differences in adherence to a feminist ideology. In particular, it is unlikely, that eighth-grade boys and girls differ in their economic vulnerability, and, 13 and 14 year old boys and girls are unlikely to vary in their adherence to a feminist ideology as adolescents are not cognitively sophisticated enough to have developed an ideology.

Therefore, we contend that the consistent and significant differences in boys’ and girls’ views about politics are probably caused by gender differences in socialization. For example, we found that girls were more likely than boys to favor increased spending for social programs, such as education and the environment, whereas boys were more likely to favor increased spending to fight terrorism. We believe that these differences reflect gender differences in socialization patterns, such that girls are taught to be nurturing and caring, while boys are taught to be assertive and aggressive (Chodorow, 1978; Hurwitz & Smithey, 1998; Hutchings et al., 2004).

The youth gender gap in spending priorities and party identification echoes the adult gender gap in political views (Chaney et al., 1998; Howell & Day, 2000). For example, Eichenberg (2003), who examined polling data from the Persian Gulf conflict in 1991 to the current wars in Afghanistan and in Iraq, found that women were consistently less likely than men to support the use of force.

We also found that boys and girls differ in their views regarding women’s roles in politics. Gender differences in young people’s views about women in politics have been documented by other researchers (e.g., Fridkin, Kenney, Crittenden, & Herrera, 2003; Hahn, 1998; Jennings, 2006; Torney-Purta, 2001–2002). For example, Jennings (2006) found that young men were significantly less progressive than young women in their attitudes toward women’s roles in politics. In addition, Jennings found that gender differences in views of women were greatest among the youngest generation of his four wave panel study (1965, 1973, 1982, and 1997). Although adolescent boys and girls differ in their views about women, adults do not. Men’s and women’s attitudes toward women in politics are remarkably similar (Davis et al., 2002).

The findings regarding the early gender gap in political views raises a number of interesting and important questions. For example, what factors influence the gender gap in policy views? Do girls model their mothers’ messages, whereas boys model their fathers’ messages? And, what about the gender differences in views about women’s role in politics? What factors, for example, lead boys to view women as less qualified to run for public office and less emotionally fit for politics? Are these views affected by the media sources to which boys choose to attend? For example, does the preference for rap music among boys—with its negative view of women—contribute to boys’ views of women as less politically competent? And, will these gender differences in views about women’s role in politics persist over time, or will these differences disappear as children grow and experience different socialization agents and events (e.g., Sears & Valentino, 1997)?

The specific socializing mechanisms that lead boys and girls to develop different views about politics need to be explored in future research. In particular, we suggest that researchers conduct large-scale panel studies of several thousand students where eighth graders are interviewed and then re-interviewed when the students are seniors in high school and beyond. At the same time, it is important to interview the parents, friends, and teachers of these students. Furthermore, it is crucial to measure additional socialization agents, such as the entertainment and news media favored by the students. Given the growth of new media (e.g., internet, cable, satellite), as well as young people’s heavy reliance on traditional media for entertainment (e.g., television, radio, video games), it is important to examine how political messages presented in both entertainment media and news media influence young people’s political attitudes and behaviors. With such a study, we can more fully understand the origins, dynamics, and sustainability of the gender gap in citizens’ views about politics.