Although an overwhelming majority of women politicians worldwide cite the media as barriers to women's election to office (Inter-Parliamentary Union 1997), the role of the news media has not been examined systemically in comparative research on women and politics. To date, studies of campaigns and the news media have been cast in country-specific terms and largely center on American elections. In the United States, research suggests that men and women candidates are often covered differently by the press. By taking a cross-national perspective, we may uncover the conditions that shape the portrayal of candidates in the news media.
The news media play a central role in contemporary election campaigns (Scammell and Semetko Reference Scammell and Semetko2000). News coverage of campaigns influences voters' views of the candidates in myriad ways, including recognition of the candidates, views of the candidates' plans and positions, and assessments of the candidates' personality, as well as overall evaluations of the competing contestants (e.g., Ansolabehere, Behr, and Iyengar Reference Ansolabehere, Behr and Iyengar1993). Candidates' campaigns have become increasingly media centered across postindustrial democracies (Norris Reference Norris2000). In order to wage successful campaigns, candidates need to garner favorable news coverage for their chosen messages. Therefore, gender differences in candidate coverage may influence voters' evaluations of male and female candidates, candidates' choice of campaign strategies, and ultimately, people's views regarding women's role in the political arena.
We compare newspaper coverage of men and women candidates in recent elections in Australia, Canada, and the United States. We theorize that the role of political parties and women's electoral successes shapes candidate coverage. We expect to find greater gender differences in the United States, where campaigns are more candidate oriented and there are fewer women in office.
GENDER BIAS IN THE AMERICAN MEDIA
A rich set of studies in the United States describes gender differences in news media coverage of candidates, suggesting that the news media play an important role in shaping women's efforts to achieve elective office. Pioneering research, based on elections in the 1980s, demonstrated that women candidates received less coverage and less prominent coverage relative to male candidates (Kahn Reference Kahn1992, Reference Kahn1994; Kahn and Goldenberg Reference Kahn and Goldenberg1991). In addition, the substance of coverage differed for men and women candidates. The news media tended to emphasize women's lack of viability, focusing more attention on the “horse race” when covering women candidates.
In addition, coverage of the men and women candidates corresponded with common gender stereotypes. For example, the news media tended to focus on “expressive strengths” such as honesty and compassion when describing women candidates, whereas “instrumental” traits, such as experience and leadership, were more commonly used to describe male candidates. Finally, coverage of policy matters corresponded to men's and women's stereotypical strengths. The issues of foreign policy, defense, trade, and the economy were more likely to be discussed for male candidates, whereas women candidates, in their campaign coverage, were more frequently linked to issues of poverty, education, and health care (Carroll and Schreiber Reference Carroll, Schreiber and Norris1997; Kahn Reference Kahn1996).
Research examining more recent electoral cycles suggests that gender differences in news coverage may have become less pronounced (e.g., Bystrom, Robertson, and Banwart Reference Bystrom, Robertson and Banwart2001; Jalalzai Reference Jalalzai2006; Smith Reference Smith1997). Dianne Bystrom, Terry Robertson, and Marie Christine Banwart's (Reference Bystrom, Robertson and Banwart2001) analysis of the 2000 Senate and gubernatorial primaries indicates that women candidates received more coverage than men. However, they continue to find that certain issues are more likely to be linked with male candidates (e.g., taxes) and other issues are more likely to be linked with female candidates (e.g., education). Moreover, women candidates receive more attention regarding their marital and family status when compared to male candidates.
Kevin Smith's (Reference Smith1997) study of 11 campaigns for the U.S. Senate and governorship in 1994 also suggests a general trend toward more equitable coverage for men and women candidates and less gender stereotyping. However, Smith continued to find more emphasis on the horse-race aspects of campaigns for women candidates. Similarly, James Devitt (Reference Devitt1999) examined news coverage in six statewide races and found that male and female gubernatorial candidates received about equal amounts of news attention. However, the news media were more likely to focus on the women candidates' personal life, appearance, and personality, while male candidates received more news attention for their policy positions and policy priorities. Finally, Farida Jalalzai (Reference Jalalzai2006), examining news coverage for senatorial and gubernatorial candidates between 1992 and 2000, finds only modest gender differences in news coverage.
Gender differences in press treatment appear to be more dramatic for presidential candidates. In particular, recent studies of Elizabeth Dole's run for the 2000 Republican nomination for president find that Dole received less coverage than some of her male opponents and more gendered coverage (Aday and Devitt Reference Aday and Devitt2001; Bystrom Reference Bystrom, Carroll and Fox2006; Heldman, Carroll, and Olson Reference Heldman, Carroll and Olson2005). Both television and print media coverage focused more on the appearance, sex, and viability of Dole, compared to her male counterparts. She received even less coverage and less positive coverage than those men who were consistently behind her in the polls.
Finally, women candidates believe that the news media treat their campaigns less favorably than their male counterparts. For example, Richard Fox's (Reference Fox1997) study of women congressional candidates and their campaign managers finds that women believe that media coverage of women candidates' campaigns reinforces gender stereotypes.
In sum, research suggests that some differences in press treatment for men and women candidates may have become less dramatic in recent election cycles (Bystrom, Robertson, and Banwart Reference Bystrom, Robertson and Banwart2001; Jalalzai Reference Jalalzai2006). However, the press continues to distinguish between male and female candidates in terms of their policy priorities and their personal qualities, including their appearance and marital status (Bystrom, Robertson, and Banwart Reference Bystrom, Robertson and Banwart2001; Devitt Reference Devitt1999; Heldman, Carroll, and Olson Reference Heldman, Carroll and Olson2005; Smith Reference Smith1997). Finally, the magnitude of gender differences in press treatment appears to be affected by the electoral office (e.g., governor, senator, president), the political context (e.g., the health of the economy, the presence of war), the status of the candidate (e.g., incumbent, challenger), and the gender of one's opponent (e.g., all male races, all female races, female versus male races) (see Heldman, Carroll, and Olson Reference Heldman, Carroll and Olson2005; Kahn Reference Kahn1996; Woodall and Fridkin Reference Woodall, Fridkin, Han and Heldman2007).
Press coverage may diverge from the candidate's own messages during the campaign. Early research examining the campaign messages of men and women candidates found significant gender differences in issue priorities, with women focusing on “female” issues like education and health care and men more likely to focus on “male” issues like the economy and foreign policy (Dabelko and Herrnson Reference Dabelko and Herrnson1997; Iyengar et al. Reference Iyengar, Valentino, Ansolabehere, Simon and Norris1997; Kahn Reference Kahn1992, Reference Kahn1992; Kahn and Goldenberg Reference Kahn and Goldenberg1991; Kahn and Gordon Reference Kahn, Gordon and Pippa1996; Witt, Paget and Matthews Reference Witt and Karen1994). More recent research has documented few gender differences in policy emphasis (Bystrom Reference Bystrom, Carroll and Fox2006; Dolan Reference Dolan2005; Sapiro and Walsh Reference Sapiro and Walsh2002).
While gender differences in policy emphasis during campaigns may be disappearing, men and women candidates continue to highlight different personality traits in their campaigns for elective office. Research examining campaigns in the 1980s and 1990s showed that women tended to focus on “male” traits like experience and leadership, while men were more likely to stress “female” traits like empathy and integrity in their controlled communications (Iyengar et al. Reference Iyengar, Valentino, Ansolabehere, Simon and Norris1997; Kahn and Goldenberg Reference Kahn and Goldenberg1991; Kahn Reference Kahn1992, 1994a). Studies examining more recent campaigns show that men and women continue to focus on different traits in their campaign messages. In particular, Bystrom (Reference Bystrom, Carroll and Fox2006) and Virginia Sapiro and Katherine Cramer Walsh (Reference Sapiro and Walsh2002) find that women candidates focus on their toughness more often than do their male counterparts.
COMPARATIVE STUDIES OF MEDIA, WOMEN, AND ELECTIONS
A rich set of country studies has examined gender and the media in Australian and Canadian elections. In both countries, women politicians are portrayed more often in terms of feminine stereotypes (Acker Reference van2003; Robinson and Saint-Jean Reference Robinson, Saint-Jean and Gingras1995). Based on television news coverage of the 1993, 1997, and 2000 Canadian elections, female party leaders' messages were less likely to receive neutral coverage than those of male party leaders (Gidengil and Everitt Reference Gidengil, Everitt, Tremblay and Trimble2003). In the 2000 election, newspaper headlines employed more aggressive language for male party leaders, and more passive discourse for Alexa McDonough, the only female party leader (Sampert and Trimble Reference Sampert, Trimble, Tremblay and Trimble2003). Further, the Canadian media tend to emphasize women's physical attributes and backgrounds (Gingras Reference Gingras and Gingras1995; Everitt Reference Everitt2003). Coverage of Pauline Hanson, founder of Australia's One Nation, speaks to this trend. Hanson was often covered on the basis of her physical appearance (Deutchman and Ellison Reference Deutchman and Ellison2004).
In a cross-national investigation of gender differences in media coverage of heads of state, Pippa Norris (Reference Norris1997) finds that women receive less coverage than their do male counterparts. Furthermore, news stories on female leaders commonly mention gender-related themes, employing the “first woman” frame. The insights of her study are illuminating, but the paucity of female presidents and prime ministers severely limits the scope, and the news sources are solely American.
On the basis of interviews in Britain, Australia, and South Africa, Karen Ross (Reference Ross2002) finds that women politicians view the media as a barrier to office. Similarly, the Inter-Parliamentary Union (1997, 169) surveyed women politicians worldwide and reported that 70% of respondents “believed men and women were viewed differently by the media.” Many believe the media question women more frequently on social issues and gender equality. While these studies offer insights into women's perceptions of the news media, it is important to examine the actual news coverage to document gender differences in media treatment.
CONTEXTUAL INFLUENCES ON COVERAGE OF CANDIDATES
We theorize that the broader context in which elections are held may condition press coverage of candidates. We focus on two categories of explanation—the candidate-centered nature of campaigns and women's representation in the political arena.
To begin, the degree to which a political system focuses on candidates over parties will influence patterns of press coverage. The more the campaign spotlight shines on individual candidates, the more the campaign showcases candidates' personal achievements, mistakes, and characteristics. Thus, coverage of candidates' stances and traits are maximized relative to the party message. By contrast, where parties play a stronger role, campaigns are more centrally coordinated, and campaign themes and issues are more uniform, leaving less room for variation among candidates. Thus, we hypothesize that more highly candidate-centered systems will magnify gender differences in candidate news coverage, while party dominated systems will minimize differences.
The extent to which a political system focuses on candidates, relative to parties, is shaped by its political institutions. Overall, countries with political institutions that encourage more cohesive parties tend to hold election campaigns with greater attention to party messages. The executive and legislative branches are interdependent in parliamentary systems, and members of parliament (MPs) find incentives to toe the party line. Thus, parliamentary systems are generally associated with more cohesive parties than are presidential systems.
Furthermore, the control over the electoral ballot also influences the incentive for candidate-centered voting (Carey and Shugart Reference Carey and Shugart1995). In particular, where party leaders exercise more control over candidate nominations, the party's hand is strengthened, and parties tend to be more cohesive. Candidate-centered campaigns, in contrast, are enhanced by party primaries, in which candidates from the same party compete publicly, and voters play an integral role in party nominations.
Finally, the number of votes a citizen may cast for a given seat also influences candidate-centered voting (Carey and Shugart Reference Carey and Shugart1995). When a single vote is cast for a single seat, a politician's personal reputation is important because candidates compete in head-to-head competition.
Legislative campaigns in the United States will be the most candidate-centered, potentially creating the greatest gender differences in press coverage. The American presidential system and use of candidate primaries leaves very little control for party elites in the nomination process, encouraging candidates to build their own constituencies, limiting reliance on their party. In the Australian and Canadian parliamentary systems, we expect more focus on party messages and, therefore, fewer gender differences in news treatment.
The second contextual factor influencing patterns of campaign coverage is women's presence in politics. We expect that countries with higher percentages of women in office will display fewer gender differences in press treatment. Elizabeth Van Acker (Reference van2003, 132) suggests that “not until the number of women increase, particularly in the higher ranks, will they generate less curiosity for the media.”
Research on stereotypes has identified several conditions leading to the revision of stereotypes (Schneider Reference Schneider2004, 146–50). All things being equal, individuals are more likely to rely on stereotypes until contradictory information arises. Stereotypes will be weakened when more pieces of contradictory information are available and if the contradictory information is spread across different individuals.
Gender stereotypes follow these patterns (Schneider Reference Schneider2004). Therefore, as journalists and voters gain exposure to women politicians acting in ways that do not conform to common gender stereotypes, they will revise their stereotypes about typical women politicians. As women achieve greater numbers in office, their activities will supply the electorate with more pieces of information that allow for more balanced and less stereotypical assessments.
Table 1 profiles women's presence in politics in each nation. With 26.7% women in the House, Australia boasts the greatest proportion of women in the lower chamber. The United States trails considerably with only 16.8% women in the House.
Table 1. Women in elected office, by country
Sources: Inter-Parliamentary Union (http://www.ipu.org [30 March 2007]); Still Counting: Women and Politics Across Canada (http://stillcounting.athabascau.ca [5 November 2007]); Center for American Women and Politics, Rutgers University, Election Watch.
An examination of women in leadership positions shows that the United States continues to lag behind Canada and Australia. Nancy Pelosi's election to Speaker of the House after the 2006 election represented a first for women in U.S. politics. In this way, Pelosi is a novelty in American politics. By contrast, Canada had already seen a woman as a major party leader (Progressive Conservative Party) and prime minster: Kim Campbell in 1993. Moreover, the New Democratic Party (NDP) elected Alexa McDonough in 1997 and 2000 and Audrey McLaughlin in 1993 as its leader (Trimble and Arscott Reference Trimble and Arscott2003). And the Australian Democrats have had nine female leaders (Grey and Sawer Reference Grey, Sawer, Galligan and Tremblay2005, 175).
Overall, women elected officials are least common in the United States. In contrast, women in politics are less of a novelty in Canada and Australia. On the basis of both the historical experience and institutional patterns discussed here, we may expect the greatest gender differences in press coverage in the United States. In contrast, men and women candidates may enjoy more equitable treatment in Canada and Australia.
Although these three nations offer significant differences in each of our theoretically important dimensions, they also share some common features that may underpin similarity in campaign coverage. First, the case studies cited in the previous section suggest the presence of gender differences in the press coverage of candidates. Second, despite differences in electoral rules, each country allows for candidate voting. Third, these three countries also house similar media systems. The United States and Canada share Daniel Hallin and Paolo Mancini's (Reference Hallin and Mancini2004) “North Atlantic or Liberal Model.” Although Australia is not part of Hallin and Mancini's study and demonstrates an extremely high degree of concentration in media ownership, it appears compatible with the Liberal Model on the most relevant dimensions for this research: professionalism, focus on information, partisan neutrality, and autonomy from government.
Finally, the three nations also share similar attitudes towards women in politics. In a measure of gender equality attitudes, developed by Ronald Inglehart and Pippa Norris (Reference Inglehart and Norris2003) from the 2000 World Values Study, Canada, Australia, and the United States score 83, 78, and 79—all significantly above the average score of 68. Gender stereotypes are similarly pervasive. In a cross-cultural study, Williams and Best (Reference Williams and Best1990) find a high degree of similarity in the attributes ascribed to women and men across 25 countries, suggesting a pattern of “pancultural” gender stereotypes. Because gender stereotypes are so pervasive, journalists are likely to adhere to the same stereotypes as the general population.
EXAMINING GENDER DIFFERENCES IN MEDIA COVERAGE OF CANDIDATES
Taken together, research on press treatment of women candidates has shown that women candidates sometimes receive less coverage and less prominent coverage than do their male counterparts. And the coverage that women receive often focuses less on their substantive concerns and more on their viability, personality, and family. Finally, the press often discusses different issues and different personality traits when covering male and female candidates (e.g., Bystrom et al. Reference Bystrom, Banwart, Kaid and Robertson2004; Devitt Reference Devitt1999; Heldman, Carroll, and Olson Reference Heldman, Carroll and Olson2005; Kahn Reference Kahn1996). Given these findings, in our content analysis we look at the differences in the quantity and prominence of coverage for men and women candidates, the press attention devoted to the candidates' families and to the candidates' viability, and the substance of issue and trait coverage. The code sheet employed in the content analysis is presented in Appendix A.
Sample of Elections and News Sources
We analyze the content of newspaper coverage of candidates (men and women) in the 2006 Canadian, 2004 Australian, and 2006 U.S. elections. Although the campaign period differs for each nation, we cover four weeks leading up to the election in each case.Footnote 1 To assemble articles for content analysis, we utilized Access World News. Two research assistants retrieved all of the stories from the national and/or regional news sections of the newspaper for each day within our time frame, and selected and saved every story that mentioned an election candidate for House or Senate.
Newspaper articles are ideal for tracing media depictions of candidates because of their impact on other news sources. Larry Bartels (Reference Bartels1996) shows that national, high-circulation newspapers such as the New York Times have a stronger impact on agenda setting in politics than does television news and that these newspapers shape content among local newspapers. In addition, local television news dedicates virtually no attention to races for the House and Senate in a given broadcast. A typical 30-minute local news broadcast gives only 36 seconds to election campaigns (Newslab 2008).
We selected three comparable broadsheet dailies: the Toronto Star, Sydney Morning Herald, and New York Times.Footnote 2 In order to select papers that have the greatest impact on voters' perceptions of candidates, we used high-circulation numbers as one criterion: The Toronto Star and Sydney Morning Herald are the highest-circulating nontabloid dailies in their respective countries. The top-circulating American newspaper, USA Today, is not available on Access World News. Equally important, the newspapers included in our sample are nationally recognizable and offer regional coverage from major metropolitan centers.
Our goal was to gather articles on a wide variety of races, not just stories on party leaders. The regional coverage maximizes the likelihood of obtaining stories on district-level races. At the same time, the national recognition of each paper encourages coverage of races from different regions within the country. The Sydney Morning Herald yielded stories on electorates in every state and territory in Australia, from New South Wales and Queensland to Tasmania. The Toronto Star led us to stories on a variety of constituencies from every Canadian region. Similarly, the New York Times offered several races from New York and also covered races from a variety of American states.
This strategy allowed us to capture a wide variety of candidates. The number of different candidates mentioned is high for each country: in Canada, 106 men and 42 women; in Australia, 74 men and 30 women; and in the United States, 146 men and 41 women.
Analysis of Canadian, Australian, and U.S. Election Coverage
Our content analysis includes a total of 354 articles: 149 from Canada, 129 from Australia, and 157 from the United States. Intercoder reliability ranges from 85% to 100% across all of the content categories. Cohen's Kappa, a far more conservative measure, ranges from .47 to 1.00, with most categories falling in the “substantial” agreement range (Landis and Koch Reference Landis and Koch1977). Each candidate mentioned in each article is a separate observation in our data set: 417 in Canada, 269 in Australia, and 426 in the United States.Footnote 3
Unlike some of the prior research examining media treatment of women candidates (e.g., Devitt Reference Devitt1999; Heldman, Carroll, and Olson Reference Heldman, Carroll and Olson2005; Kahn Reference Kahn1996; Smith Reference Smith1997), ours fails to find any significant gender differences in 1) the amount and prominence of candidate coverage, 2) the amount of attention given to the candidates' viability, 3) the focus on the candidates' family background, and 4) the tone of coverage. In each of the three countries, men and women candidates are treated equitably in terms of amount and prominence of press attention, and the news media do not focus more attention on the viability or family of female candidates running for office.Footnote 4
Turning to the substance of issue and trait coverage, we rely on a rich literature showing that people rely on stereotypes to draw inferences about the trait characteristics and issue competencies for men and women candidates. Research on stereotypes suggest that men, generally, are perceived as possessing agentic traits, such as being bold, rational, and unemotional, whereas women, generally, are perceived as possessing communal traits, such as sensitivity, empathy, and passivity (Banaji, Hardin and Rothman Reference Banaji, Hardin and Rothman1993; Diekman and Eagly Reference Diekman and Eagly2000; Sczesny et al. Reference Sczesny, Bosak, Neff and Schyns2004).
Political science research has shown that people apply these gender stereotypes to the political arena. Furthermore, these stereotypes lead people to view men and women candidates as having distinct areas of policy expertise. In particular, women candidates are viewed as being more competent at handling “compassion” issues, such as poverty, education, the environment, child care, and health-care policy, whereas men are seen as more competent at dealing with “male” issues, such as the economy, foreign policy, and other defense issues (Leeper Reference Leeper1991; Rosenwasser and Seale Reference Rosenwasser and Seale1988; Sapiro Reference Sapiro1981–Reference Sapiro82).Footnote 5
These issue and trait stereotypes are robust and have been identified by researchers using a variety of methods, including surveys (Alexander and Andersen Reference Alexander and Andersen1993; Cook and Wilcox Reference Cook, Wilcox and Weisberg1995; Dolan Reference Dolan2004; Fox Reference Fox1997; Koch Reference Koch1999, Reference Koch2000, Reference Koch2002; McDermott Reference McDermott1997; Niven Reference Niven1998; Sanbonmatsu Reference Sanbonmatsu2003) and experiments (Dolan Reference Dolan1997; Huddy and Terkildsen Reference Huddy and Terkildsen1993; Kahn Reference Kahn1996; King and Matland Reference King and Matland2003; Leeper Reference Leeper1991; Riggle, Shields, and Johnson Reference Riggle, Miller, Shields and Johnson1997; Rosenwasser and Dean Reference Rosenwasser and Dean1989; Rosenwasser and Seale Reference Rosenwasser and Seale1988; Sapiro 1981–2).
To examine gender differences in issue coverage, we rely on gender stereotyping and define “male” issues as the economy, business, taxes, energy/oil, trade, employment/jobs, defense, international organizations, nuclear arms control, treaties, and foreign affairs, and “female” issues as women's rights, gender quotas, abortion, HIV/AIDS, violence against women, gay rights, women in elected office, education, health, welfare, environment, care for the elderly, child care, parental leave, and pensions.Footnote 6 For instance, in coverage of the 2006 Canadian election, ideologically diverse female candidates, from the NDP's Olivia Chow (Trinity-Spadina) to the Liberal's Deborah Coyne (Toronto Danforth) and Bonnie Brown (Oakville) to the Conservative's Arnjeet Kaur Sangha (Mississauga-Brampton South), are all linked with child-care issues (Toronto Star, 21 December 2005, 22 December 2005, 11 January 2006).
Our data can only reveal differences in press coverage and do not reflect the candidates' own campaign strategies. In other words, it is possible that the news media in these countries focus more attention on female issues for women candidates, compared to male candidates, because women candidates are more likely to emphasize these issues.
In Table 2, we look at the amount of “male” and “female” issue coverage given to men and women candidates. The results are striking in their consistency. Across each of the countries, male issues receive significantly more press attention for male candidates than female candidates. In particular, male issues are covered 54% of the time for male candidates in Canada, compared to only 40% of the time for female Canadian candidates. In Australia, male issues are covered less frequently, overall. However, the emphasis on male issues is almost twice as likely for male candidates than female candidates (i.e., 42% versus 24%). Finally, in the United States where male issues are most common, coverage of male issues continues to be significantly more common for male candidates than for female candidates (e.g., 80% versus 70%).Footnote 7
Table 2. Gender and type of issue coverage, by country
Note: “Male” issues are composed of references to economic and foreign policy issues, including the economy, business, taxes, energy/oil, trade, employment/jobs, defense, international organizations, nuclear arms control, treaties, foreign affairs, peace. “Female” issues are social and social policy issues, including women's rights, gender quotas, abortion, HIV/AIDS, violence against women, gay rights, women in elected office; “Social Policy” issues include education, health, welfare, environment, care for the elderly, childcare, parental leave, and pensions.
Entries are percentages of male and female issues linked to a given candidate, based on the total number of male and female issues. The p value is based on a difference in proportions test. Calculations include prime ministerial candidates in Canada and Australia.
**p < .01
*p < .05
These results demonstrate a consistent and dramatic difference in the substance of issue coverage given to men and women candidates. “Male” issues receive more attention in the coverage of men, while “female” issues receive more attention for women.Footnote 8
Just as we expect coverage of men and women candidates to focus on different types of issues, we expect the press to focus on different personality traits. We categorize traits as “male” and “female” traits based on a wealth of stereotype literature (Banaji, Hardin, and Rothman Reference Banaji, Hardin and Rothman1993; Diekman and Eagly Reference Diekman and Eagly2000; Fox and Oxley Reference Fox and Oxley2003; Huddy and Terkildsen Reference Huddy and Terkildsen1993; Rosenwasser and Seale Reference Rosenwasser and Seale1988; Sczesny et al. Reference Sczesny, Bosak, Neff and Schyns2004). Female traits include: gentle, honest, weak, weak leader, attractive, passive, emotional, uninformed, unintelligent, compassionate, noncompetitive, and dependent. Male traits include hardworking, untrustworthy, strong leader, strong, vital, competitive, effective, tough, intelligent, aggressive, knowledgeable, independent, and ambitious.
We expect coverage to follow common gender stereotypes, with female traits covered more extensively for women candidates, compared to male candidates. For instance, one article describes Nancy Pelosi as gentle and attractive and compares her to a grandmother at Christmas (New York Times, 30 October 2006). In coverage of the Australian election, one article dubs Jackie Kelly, MP for Lindsay, as Prime Minister John Howard's “pet,” denoting her dependent status (Sydney Morning Herald, 20 September 2004).
In Table 3, we look at the proportion of “male” and “female” trait coverage given to men and women candidates. We find consistent and strong support for our hypotheses. Male traits dominate coverage of male candidates in each of the three countries, with more than 80% of trait coverage focusing on male traits for male candidates. In contrast, male traits are discussed significantly less often for female candidates.
Table 3. Gender and candidate traits, by country
Note: “Female” traits include the following: gentle, honest, weak, weak leader, attractive, passive, emotional, uninformed, unintelligent, compassionate, noncompetitive, and dependent. “Male” traits include hardworking, untrustworthy, strong leader, strong, vital, competitive, effective, tough, intelligent, aggressive, knowledgeable, independent, and ambitious.
Entries represent the percentage of male and female traits linked to a specific candidate, based on the number of male and female trait mentions. The p value is based on the difference in percentages test. Calculations include prime ministerial candidates in Canada and Australia.
**p < .01
*p < .05
More specifically, 40% of the traits discussing women candidates in Canada focus on female traits, while just 16% of the trait discussion given to male Canadian candidates discusses female traits. Similarly, in Australia, 34% of women's trait coverage focuses on female traits, compared with only 17% of male candidates.Footnote 9
In the United States, the gender differences in trait coverage are dramatic. In the coverage of male candidates, male traits are dominant, accounting for almost 90% of all trait coverage. In contrast, female traits are more common than male traits in the coverage of women candidates, accounting for 56% of all trait coverage. The emphasis on male traits is more than twice as great for male candidates than for female candidates in the United States (88% versus 43%). Conversely, the focus on female traits in the United States is more than four times greater for women than for men candidates (56% versus 12%).Footnote 10 Overall, our findings show that gender differences in issues and traits extend across national boundaries.Footnote 11
CONCLUSIONS
Despite differences in women's representation and electoral institutions, candidates are often portrayed in terms of long-standing gender stereotypes, and these stereotypes are common to all three democracies. Thus, it appears unlikely that gender biases in coverage can be attributed to a particular election or to country-specific forces. Indeed, gender is tightly woven into contemporary campaigns. The parliamentary systems of Canada and Australia, with their more cohesive parties and more unified party messages, still perpetuate gender stereotypes that link female candidates more often with stereotypically “female issues.”
If we were to have found that greater shares of women in Parliament were linked to fewer gender stereotypes in candidate coverage, the prevalence of these stereotypes would have been less troubling for the democratic process. One might assume that over time, as women gained more seats, candidate coverage would become more equitable. However, our findings suggest that higher numbers of women in office will not automatically translate into more gender-neutral coverage. Despite Australia's and Canada's higher proportions of women in their national legislatures, relative to the United States, all three display similar patterns of gender stereotypes linked to candidates' traits and issues. Thus, gender differences are likely to characterize election campaigns in future elections and continue to pose challenges for male and female candidates alike.
“Male” traits tend to overlap with “leadership” traits (e.g., leadership, strength, intelligence, and toughness), and these traits are often valued more highly by voters when they are evaluating competing candidates for electoral office (Jamieson Reference Jamieson1995; Markus Reference Markus1982). Linked to stereotypically “female” traits, women candidates miss the opportunity to demonstrate their leadership, strength, and knowledge. If citizens are learning more about the leadership traits of male candidates than female candidates, and if citizens believe that these traits are more important, then gender differences in trait coverage may hurt women candidates at the polls.
Similarly, due to the relatively stronger focus on social issues when covering women candidates, women do not have the luxury of demonstrating their competence to deal with “male” issues. Given common gender stereotypes, voters and party leaders may be less convinced that women candidates can handle these male issues, compared to their male counterparts. Moreover, when issues like the economy, terrorism, or foreign policy dominate the political landscape, women candidates may be particularly disadvantaged.
Not only may gender differences in news treatment influence women candidates in contemporary campaigns, but these coverage patterns may also affect the future of women in politics. Gendered images of political figures may influence the decisions of political or party elites when they nominate or solicit candidates. Furthermore, these gender differences may also shape the ways in which men and women in the electorate look at political life. Margaret Scammell and Holli Semetko (Reference Scammell and Semetko2000) argue that a central part of the media's role in a democracy is to accurately represent all social groups. If particular groups, such as women, receive unfair treatment, then the democratic process is ill-served.
This research represents a first step in comparing gender in campaigns. Future research should include a wider variety of democratic systems. It is also important to measure local coverage of political candidates in their districts, in addition to national coverage. Scholars should also look at the connection between the candidates' own messages and media coverage of these messages. By measuring the political context, the media messages, and candidate messages in a comparative perspective, we will improve our understanding of how gender shapes contemporary election campaigns.
APPENDIX A. CODESHEET FOR NEWSPAPER CONTENT ANALYSIS
Newspaper Code, Date (Month/Day/Year), Sex of Author: 1 = Male, 2 = Female
Sex of Candidate: 1 = Male, 2 = Female
Amount and Prominence of Coverage
Length of Article (Number of Paragraphs)
Candidate Mentioned in Headline: 1 = Yes, 0 = No
Number of Paragraphs about Candidate
Is There a Picture of Candidate: 1 = Yes, 0 = No
Tone of Coverage
Tone of Headline (1 = Positive, 2 = Negative, 3 = Mix 4 = Neutral; 0 = Not about Candidate)
Tone of Article (1 = Positive, 2 = Negative, 3 = Mix 4 = Neutral; 0 = Not about Candidate)
Number of Criticisms about Candidate
Viability Coverage
“Horse race” Paragraphs
“Horse race” Content (7 = likely winner, 6 = likely winner, but losing ground, 5 = competitive, but gaining ground, 4 = competitive, 3 = competitive, but losing ground, 2 = noncompetitive but gaining ground, 1 = noncompetitive, sure loser, 0 = no mention horse race)
Campaign Resources (1 = Positive, 2 = Negative, 3 = Mix, 4 = Neutral, 0 = Not about Candidate)
Qualifications (1 = Mention of Prior Elective Office, 2 = Mention of Prior Appointive Office, 3 = Mention of Other Qualifications, 4 = Mention lack of Qualifications, 0 = No Mention)
Gender Coverage
Marital Status (1 = Married 2 = Never Married, 3 = Divorced, 4 = Widow, 0 = No Mention)
Number of Paragraphs about Candidate's Spouse
Are Children of Candidate Mentioned? 1 = Yes, 0 = No
Number of Paragraphs about Candidate's Appearance
Number of Paragraphs about Candidate's Gender (first woman, etc.)
Issue Coverage
Defense Issues Nuclear Arms Control Foreign Affairs Child Care
International Organizations Treaties Welfare Parental Leave
Economy Taxes Education Employment/Jobs
Pensions Energy/Oil Trade Women in Politics
Environment Abortion Drugs Business
Crime HIV/AIDS Health Gay Rights
Women's Rights Violence Against Women
Gender Quotas Government Spending
Trait Coverage
Honest/Trustworthy Attractive Compassionate Strong Leader
Gentle Expressive Moral Objective
Analytical Effective Knowledgeable Consistent
Hardworking Tough Strong Vital
Intelligent Independent
Weak Passive Noncompetitive Untrustworthy
Biased Emotional Dependent Immoral
Unattractive Uninformed Erratic Aggressive
Weak Leader Unintelligent Ineffective Insensitive
Competitive Unexpressive Ambitious/Power-Hungry
