mixed signals? gender and the media's coverage of the …€™s coverage of the 2008 vice...

polp_350 221..257

Mixed Signals? Gender and theMedia’s Coverage of the 2008 Vice

Presidential Candidates

LETICIA BODEUniversity of Wisconsin-Madison

VALERIE M. HENNINGSIowa State University

The 2008 presidential election was one of the most watched campaignsin American history, and prominently featured the vice presidentialcandidates, Governor Sarah Palin and Senator Joseph Biden. Thiselection contest presents an exciting opportunity to expand and test ourcurrent understandings of the relationship between gender and mediacoverage. We examine this relationship using computer-assisted contentanalysis of major newspapers, television news broadcasts, and politicalblogs. These three media are analyzed in terms of both quantity andsubstance of coverage received by Palin and Biden. Using a multiplemedium perspective in our investigation of the contest between Palinand Biden, this article finds differences in volume of coverage (GovernorPalin receives more), substance of coverage (roughly reflecting gender-based stereotypes), and medium (with differences in volume andsubstance of coverage across newspapers, television, and blogs).Implications for future contests and for American politics in general arediscussed.

Keywords: Gender, Media Coverage, 2008 Election, Stereotypes, U.S. VicePresidential Candidates, Content Analysis, Sarah Palin, Joe Biden.

Related Articles: “Women Candidates and the Media: 1992-2000 Elections.”(2006)http://onlinelibrary.wiley.com/doi/10.1111/j.1747-1346.2006.00030.x/abstract

Acknowledgements: The authors would like to thank Politics & Policy’s anonymous reviewers fortheir helpful comments on this manuscript. They would also like to thank the participants of theirpanel at the 2009 Midwest Political Association’s Annual Conference, and the Women andPolitics Research Group at the University of Wisconsin for their suggestions on previous drafts ofthis project.

bs_bs_banner

Politics & Policy, Volume 40, No. 2 (2012): 221-257. 10.1111/j.1747-1346.2012.00350.xPublished by Wiley Periodicals, Inc.© The Policy Studies Organization. All rights reserved.

http://onlinelibrary.wiley.com/doi/10.1111/j.1747-1346.2006.00030.x/abstract


“The Influence of Female Candidates’ Campaign Issues on PoliticalProselytizing.” (2008)http://onlinelibrary.wiley.com/doi/10.1111/j.1747-1346.2008.00093.x/abstract“Incumbent Responsiveness to Female Challengers.” (2010)http://onlinelibrary.wiley.com/doi/10.1111/j.1747-1346.2010.00271.x/full

Related Media: Blogs at http://themoderatevoice.com/,http://committedtoromney.com/, and http://www.barackoblogger.com/

La elección presidencial del 2008 fue una de las campañas más vistasen la historia de los Estados Unidos, dando un lugar destacado a loscandidatos para la vicepresidencia, la gobernadora Sarah Palin y elSenador Joseph Biden. Esta contienda electoral presenta una granoportunidad para ampliar y poner a prueba nuestra comprensión actualde la relación entre el género y la cobertura de los medios decomunicación. Este artículo examina esta relación usando un análisis decontenido de los principales periódicos, programas de televisión y blogspolíticos. Estos tres medios son analizados en términos de cantidad ysustancia de la cobertura recibida por Palin y Biden. Usando unaperspectiva de múltiples medios en nuestra investigación de lacompetencia entre Palin y Biden, este estudio encuentra diferencias enel volumen de cobertura (la gobernadora Palin recibió más), sustanciade la cobertura (más o menos reflejando estereotipos de género), y demedio (con diferencias de volumen y sustancia de cobertura entreperiódicos, televisión, y blogs). Las implicaciones para futurascontiendas y la política estadounidense en general son discutidas.

On August 29, 2008, Governor Sarah Palin of Alaska became the firstwoman in the history of the United States to be named as the vice presidentialcandidate for the Republican Party. At a press conference announcing hernomination, she acknowledged this historical moment with the followingreflection.

I think as well today of two other women who came before me in nationalelections. I can’t begin this great effort without honoring the achievementsof Geraldine Ferraro in 1984 and of course Senator Hillary Clinton, whoshowed such determination and grace in her presidential campaign. It wasrightly noted in Denver this week that Hillary left 18 million cracks in thehighest, hardest glass ceiling in America but it turns out the women ofAmerica aren’t finished yet and we can shatter that glass ceiling once andfor all. (Roll Call 2008)

At a time when women hold less than 17 percent of Congressional seats,approximately 23 percent of statewide elective executive offices, and have yet tobe elected to the offices of president and vice president, this moment, as noted by

222 | POLITICS & POLICY / April 2012



http://onlinelibrary.wiley.com/doi/10.1111/j.1747-1346.2010.00271.x/full

http://themoderatevoice.com/

http://committedtoromney.com/

http://www.barackoblogger.com/

Governor Palin, was indeed an important political achievement for women inthe United States (Center for American Women and Politics 2010).

The electoral contest between Governor Palin and the Democratic pick forvice president, Senator Joseph Biden from Delaware, presents an excitingopportunity to expand and test our current understandings of the relationshipbetween gender and media coverage. This article examines this relationshipusing computer-assisted content analysis of major newspapers, television newsbroadcasts, and political blogs. These three media are analyzed in two ways.First, we investigate differences in the quantity of coverage received by Palinand Biden. Second, we analyze the substance of the media coverage in terms ofmultiple forms of gender stereotypes. Using a multiple medium perspectiveallows a unique comparison across conventional media and new forms ofpolitical communication, thus providing a useful understanding of the portrayalof this most recent political achievement for women in the United States.

The importance of this research is threefold. First, the electoral contestbetween Palin and Biden was one of the most-watched and most-covered vicepresidential campaigns in history. Discussion of McCain’s choice of Palin alonerepresented a full 7 percent of the television election coverage of the entire 2008presidential campaign (Center for Media and Public Affairs 2009), andGovernor Palin—fifth—and Senator Biden—thirteenth—were both among thetop 20 news makers in 2008, outranking the likes of Governor RodBlagojevich—sixth—and the Pope—sixteenth—(Pew Project for Excellence inJournalism 2009). The debate between Palin and Biden was the most-watchedvice presidential debate in United States history, drawing 69.9 million viewers(Gorman 2008). This represents 33 percent more viewers than the first debatebetween Obama and McCain, and 61 percent more viewers than the 2004 vicepresidential debate between Cheney and Edwards. It is essential that weunderstand the media coverage of this vice presidential contest simply becausethe public and the media decided how important a contest it was.

Second, the nomination of Sarah Palin as the first female vice presidentialrunning mate for the Republican Party marks a new political achievement forU.S. women. As a new level of political success, this achievement suggests apossible shift in the role of women in the political system, which warrants athorough examination of how political factors, such as media coverage ofwomen candidates, may shift alongside the changing roles of women. It alsoallows an important point of comparison to 1984—the only other vicepresidential contest featuring a female candidate. Without an examination ofvice presidential contests, our understanding of the intersection of politicalcommunication and the experience of women as political candidates isincomplete.

Finally, this research examines the Palin–Biden case across three differentmedia. While there is a growing collection of research examining the roles ofonline media in politics (Herrnson, Stokes-Brown, and Hindman 2007; Howard2005), there is a lack of research that takes into consideration how campaign

Bode / Hennings / GENDER AND MEDIA COVERAGE | 223

coverage by political blogs compares with that of other media, such asnewspapers and television. Even less attention is paid to the relationship ofgender and media coverage across this divide of “newer” (blogs) and “older”(newspapers, television) media. Whereas most extant research addresses only asingle medium, we examine both the volume and substance of coverage Palinand Biden received across newspapers, television, and political blogs.

Gender and Media Coverage

The relationship between gender and media coverage has served as thestarting point for a growing body of research. The results of these studies reveala great deal regarding the amount and substance of coverage men and womenrunning for office receive. Although attention to this relationship continues togrow, questions remain concerning whether differences exist in the amount andtype of coverage that male and female candidates receive.

Leading work examining gender and media coverage shows that femalecandidates tend to receive less coverage than their male counterparts (Kahn1992, 1994a, 1994b; Kahn and Goldenberg 1991). For example, in an analysis ofpresidential campaigns, Falk (2008, 101) finds that newspapers printed “fewerstories and fewer words per story about women than they did about men whohad similar credentials and polled about the same.” This lack of parity inquantity of media coverage has also been detected in the coverage of womenofficeholders (Carroll and Schreiber 1997).

However, the question of whether or not differences exist in the amount ofcoverage between men and women candidates is far from settled. Whilehistorically women running for president tended to receive less coverage (Falk2008), recent scholarship suggests this may not hold true for contemporary racesfor U.S. president (Lawrence and Rose 2010). This finding has been suggested inresearch examining other office levels as well. The discussion of gender andcandidate communication by Bystrom and others (2004) highlights a possibleshift in how women candidates are covered by the media. These authors foundthat in the 1998 senatorial and gubernatorial races, women received lessnewspaper coverage than their male colleagues, but in the 2000 and 2002 races,a shift in quantity of coverage occurred, with candidates of both gendersreceiving similar amounts of coverage (Bystrom et al. 2004). Results suggestingsimilar amounts of media coverage between men and women candidates havebeen found in studies of senatorial and gubernatorial races (Jalalzai 2006; Smith1997), in mayoral races (Atkeson and Krebs 2008), as well as in comparativework on gender and media coverage (Hinojosa 2010; Kittilson and Fridkin2008; Semetko and Boomgaarden 2007).1

1 In an article of Globe and Mail coverage of the 1976, 1993, and 2004 Canadian Conservativeparty leadership races, women candidates received more news attention than their mencounterparts (Trimble 2007).


As our understanding of how the quantity of coverage received continues toevolve, scholarship is also building upon what we know concerning thesubstance of media coverage of men and women candidates. The work ofJamieson (1995) maps the various “double binds” women face as public,political figures and the preference for masculine traits in political leaders overthose traits typically associated with women. A wealth of scholarship supportsthis instrumental understanding of masculine and feminine traits, and how thepreference for masculine traits over feminine can translate into stereotypeswomen face in the political sphere (Huddy 1994; Huddy and Terkildsen 1993;Koch 2002; Rosenwasser and Dean 1989).

Stereotypes invoking masculine and feminine traits, as well as traditionalnotions of male/female roles, consistently arise in studies analyzing differencesin the substance of coverage of men and women candidates (Everitt andGidengil 2003; Kahn 1996; Lawrence and Rose 2010). More mentions ofemotions, family, and personal background information in relation to femalecandidates are ways the media have reinforced stereotypical notions ofappropriate feminine traits and female roles in coverage of women candidates(Aday and Devitt 2001; Falk 2008; Jamieson 1995; see also Foreit et al. 1980).The media raise questions of campaign viability and experience for almost allpolitical candidates, yet these questions are raised more often in cases of femalecandidates (Kahn 1996; cf. Jalalzai 2006). Highlighting family and electabilitywhen covering women candidates are ways the media perpetuate notions oftraditional male/female roles, which can disadvantage women in the politicalrealm where masculine traits are preferred.

Another way stereotypes emerge in media coverage of women candidates isthe linkage of certain types of issues and policies to women candidates andothers to men candidates. This delineation of issues falls along social definitionsof masculinity and femininity with men being linked to more “masculine” issues,such as the economy, foreign affairs, and defense, while women tend to belinked to more “feminine” issues, such as education, health care, and poverty(Bystrom, Robertson, and Banwart 2001; Carroll and Schreiber 1997; Kahn1996; Kittilson and Fridkin 2008; Murray 2010).

Discussions of physical appearance are yet another way media coverage hasdiffered in covering men and women candidates. In research on women whohave run for president, senator, and governor, scholars have found morephysical appearance descriptions of attire or age in newspaper coverage of thewomen candidates than of their male opponents (Devitt 1999; Falk 2008;Heldman, Carroll, and Olson 2005; Hinojosa 2010; Kahn 1992; 1996; Kahn andGoldenberg 1991).

Research Hypotheses

Previous work analyzing the relationship between gender and mediacoverage tells us to expect differences in the media coverage of female


candidates, in both amount and substance of coverage. Although scholarlyattention to these potential differences has grown, it is unclear whether thesedifferences will apply to the case of vice presidential candidates. Given thecompeting hypotheses regarding the specifics of when and in what formcoverage differences arise between male and female candidates, currentscholarship does not provide a clear projection of what the presence of a femalecandidate in a contest, such as the 2008 Palin–Biden contest, “should” hold.

The importance of testing the Palin–Biden case is clear. This campaign washeavily covered by the media, and we know media coverage is influential inshaping the political landscape (Ansolabehere, Behr, and Iyengar 1993; Bartels1993; Iyengar and Kinder 1987). If differences in media coverage exist, then anuneven field could have negative effects on the campaigns of certain candidates.Research has shown that stereotypes of women can negatively influence boththeir potential for election success and their political careers (Alexander andAndersen 1993; Heflick and Goldenberg 2010; Huddy and Terkildsen 1993;Rosenwasser and Seale 1988; Sapiro 1982). If gender stereotypes in mediacoverage have the ability to negatively affect women candidates, this calls intoquestion the American political system’s ability to produce electedrepresentatives in a fair and democratic manner. Moreover, although vicepresidential candidates receive less media coverage than their presidentialcounterparts, research shows that perceptions of vice presidential candidatesmay have an impact on vote choice, especially when the vice presidentialcandidates are unique in some way (Ulbig 2009; Wattenberg 1995). Given thepotential implications of unbalanced media coverage in the political world, weconsider three ways where coverage might be different across issue and medium.

As highlighted above, male and female candidates in more recent electionsappear to be moving toward parity in the amount of coverage they receive(Bystrom et al. 2004; Smith 1997). We expect this trend toward parity tocontinue in the 2008 contest concerning the vice presidential candidates. Weanticipate the amount of coverage Palin receives in comparison to Biden to begreater due to the fact that she is the first female vice presidential candidate tobe nominated by the Republican Party. The novelty of Palin’s nomination,therefore, makes for an interesting story that could increase the amount ofattention she receives by the media. This “first woman” effect was detected inmedia coverage of the campaigns of other political firsts for women, includingthose of Elizabeth Dole and Geraldine Ferraro (Heith 2003; Heldman, Carroll,and Olson 2005). Although scholarship has suggested women candidatesgenerally receive less media coverage than their male peers, we expect thenovelty of Palin’s candidacy to outweigh any factors that would produce lesscoverage of her candidacy as compared with Biden.

Hypothesis 1: The amount of coverage Sarah Palin received from the mediawas greater than the amount of coverage her opponent Joseph Bidenreceived as a vice presidential candidate.


In addition to differences in volume, it seems clear from the literature thatdifferences in substance also tend to arise when comparing the coverage of maleand female candidates. The literature informs us, however, that not all genderstereotypes are detected in the races of all women candidates. This may be aresult of changes in the political landscape and the accumulation of politicalsuccess by women over time, or may be a function of individual races (Jalalzai2006). Vice presidential contests including a female candidate, such as that of2008, allow for an entirely new case with which we can test conflictingexplanations, adding an important data point to the ongoing attempt to discernwhether and when gender stereotypes persist in media coverage of electoralcandidates.

To determine whether such stereotypes emerge in the substance of thePalin–Biden media coverage, we examined the coverage in terms of four genderstereotypes identified throughout the literature on gender and media coverage,based on the following topics: family, electability, policy issues (of which weconsidered education, foreign policy, health care, social issues, and taxes/economy), and physical appearance. If gender stereotypes were to emerge in thisarticle, we would expect to see Palin’s coverage include more discussion offamily, electability, physical appearance, and the “feminine” policy issues ofeducation, health care, social issues, and physical appearance, whereas Biden’scoverage would be more closely linked with discussions of “masculine issues,”such as foreign policy, taxes, and the economy.

Hypothesis 2: Gender stereotypes will be reflected in the substance of thecoverage Sarah Palin and Joseph Biden received from the media as vicepresidential candidates. These stereotypes are likely to emerge whencomparing the candidates’ coverage in four areas: family, electability,policy issues, and physical appearance.

We began our project with the assertion that the medium matters and couldcertainly matter when it comes to the relationship between gender and mediacoverage.2 This article contributes to the greater discussion of gender andmedia coverage by examining the Palin–Biden coverage across three differentmedia: television, newspapers, and political blogs. We expect to see differencesin the coverage of Palin and Biden across these three media. This approach isgrounded in the understanding that all three media not only influence thepolitical sphere, but also do so by working in different ways for differentaudiences. For instance, citizens gain different types of information fromtelevision than from newspapers (Chaffee and Frank 1996). The different effects

2 This expectation regarding the importance of the medium is grounded in the research ofMcLuhan (1994). His work, summed appropriately with his “the medium is the message” phrase,suggests that the medium used to communicate a message, not just the content of the message, caninfluence how that message is perceived. The notion that differences in media can matter can belinked directly to McLuhan’s seminal work Understanding Media: The Extensions of Man.


on citizens of television as compared with newspapers span a variety of areas,including differences in decision making (Clarke and Fredin 1978), as well asdifferences in political knowledge and behavior (Garramone and Atkin 1986).Research on online political news coverage and blogs shows that users of onlinemedia are also influenced in different ways as compared with traditional medialike newspapers and television (Atkinson 2007; Johnson and Kaye 2004).

Blogs are a relatively new way of communicating, and we have yet to learntheir full effect in conveying political information, or whether and how thateffect may differ from that of traditional forms of media. Some scholarshipexamining differences between traditional and new media coverage of politicalissues suggests that blog coverage is different from traditional media sources:Blogs with partisan leanings judge the newsworthiness of stories using a partisanlens while traditional media, such as wire services, adhere to more conventionaljournalistic standards in determining newsworthiness (Baum and Groeling2008). Other studies suggest that the coverage between blogs and traditionalmedia is more similar than different, both in terms of volume and content(Haynes and Pitts 2009; Lee 2007). These contradictory conclusions emphasizethe need to include blogs as a part of additional investigations into possibledifferences across media. Any differences detected between old and new mediaare particularly important and relevant given the continued growth of newmedia use—blog readership increased from 13 percent in 2006 to 24 percent in2008 (Ketchum and USC Annenberg Strategic Public Relations Center 2009).

Hypothesis 3: The amount and substance of coverage Sarah Palin andJoseph Biden received as vice presidential candidates will differ acrossmedia.

Research Design

We assembled a universe of media coverage of Governor Palin and then-Senator Biden by making use of the LexisNexis database, which provides accessto five billion searchable documents from more than 40,000 legal, news andbusiness sources (LexisNexis 2009). Through the use of LexisNexis, we gathereda distinctive collection of coverage that spanned three separate media:newspapers, television, and blogs. We chose to examine the time period ofAugust 16-November 11, which roughly consists of one week before theDemocratic National Convention until one week post election. Because Palinwas added to the McCain ticket so late, this minimizes the amount of time(about two weeks) when only Biden was a vice presidential candidate, but stillincludes the lead-up to the Democratic National Convention which we felt wasimportant for parallel coverage of the two candidates. Likewise, the time periodextends just past the election to capture any post-election discussion that mayhave taken place with regard to either of the vice presidential candidates. Thistime frame results in about three months of coverage of the vice presidentialcandidates across newspapers, television news, and blogs.


To establish our universe of newspaper coverage, we selected three of thelargest, most widely read national newspapers in the United States: The NewYork Times (daily circulation: 1,000,665), The Washington Post (dailycirculation: 673,180), and USA Today (daily circulation: 2,284,219).3 This iscertainly not a complete look at all newspaper coverage of the vice presidentialcandidates, but it offers a view of what many people seeking national news wereseeing during the 2008 election. We performed three searches—for those articlesthat mentioned both Palin and Biden, only Palin and only Biden—which yieldeda total of 1,676 articles.4 While this is a somewhat crude search technique, andsurely includes articles not focused primarily on the vice presidential candidates,it ensures that every potential article covering Palin and Biden is included in oursample.

Similar procedures were conducted for selection of television news mediacoverage. We included LexisNexis transcripts of the three major networks’evening news broadcasts (NBC Nightly News with Brian Williams, ABC WorldNews with Charles Gibson, and CBS Evening News with Katie Couric), as wellas CNN and Fox News equivalents (CNN Situation Room and Fox SpecialReport with Brit Hume),5 which together encompass the vast majority ofnational television evening news available to the average media consumer.Again, three searches were performed for transcripts mentioning bothcandidates, only Palin and only Biden, for a total of 697 stories.

Finally, we turned to blogs. Selection for this medium was more complex,since the potential universe of blogs is essentially infinite and always changing—thus it is impossible to even sample from a known universe, let alone ensure arandom or representative sample. Instead, we chose a purposive sample ofblogs, representing those easily available to the average online citizen. To keepparallel structure with the other media samples, we looked within the realm ofblogs archived by LexisNexis. We then took a sample from that population tofocus on our subject matter of interest, by looking only at Newstex Government& Politics Blogs.6 This is a subset of blogs collected by Newstex LLC via theirservice, “Blogs on Demand,” which pertains to government and politics topics,and consists of a total of 391 separate blogs, varying in topic, length, format,ideology, and agenda. Of those 391 blogs, 55 mentioned Palin, Biden, or bothcandidates over our collection time period. To ensure a relatively unbiasedsample, those 55 blogs were hand-coded by two coders to determine their

3 This selection also has the benefit of capturing both “elite” (New York Times and WashingtonPost) and more mass-based (USA Today) newspapers.4 Exact search strings available from the authors upon request at [email protected] MSNBC transcripts would also have been included but were unavailable from LexisNexis.6 LexisNexis archives only a selection of blogs. Newstex is a content archival service that allowsbloggers to opt into having their content syndicated through Newstex, and by extension, throughLexisNexis. When that content is viewed through Newstex, bloggers receive small royalties. Thereis no reason to believe that bloggers would opt into this service in any kind of systematicallybiased way. For more information, see Newstex (2009).


mailto:[email protected]

ideological slant, if any. This revealed a remarkably balanced sample, with 22liberal blogs, 20 conservative blogs, ten neutral blogs, and three that could notbe determined (94.5 percent coder agreement, see Appendix B for moreinformation).7 Again, we are not looking at the entire “blogosphere,” but thisassures us that our sample allows us a glimpse of how political blogs generallycovered the 2008 vice presidential candidates. Searching as above, we generateda universe of LexisNexis blog coverage, divided by those blog entries thatmentioned both candidates, those mentioning only Palin and those mentioningonly Biden, for a total of 1,187 blog entries.

Content Analysis

To understand how each candidate was covered by various media during the2008 election cycle, we needed to analyze the content of our collected material.Because we are fortunate enough to have a large volume of coverage to workwith, which spans three separate media, hand coding of our sample was notpractical. Instead, we used a computer-assisted content analysis program thatallowed us to systematically examine the volume of coverage we collected toinvestigate the media’s treatment of Palin and Biden. We employed Yoshikoder,a program for performing computer-aided content analysis which allows usersto generate their own dictionaries, allowing for a great deal of flexibility in whatone can analyze (Lowe 2006). It offers frequencies, proportions, and statisticalcomparisons between two subsets of text. Use of computer-assisted contentanalysis allows us to examine a much larger sample of news coverage thantraditional hand coding, and also facilitates replication to a much greater extent.

We generated our own dictionary of over 200 words based on informationfrom previous literature and building off of other content analysis dictionaries,including the Linguistic Inquiry and Word Count (LIWC) dictionary, andLaver and Garry’s policy positions dictionary.8 For each area of interest—family, electability, policy issues (including education, foreign policy, healthcare, social issues, and taxes and the economy), and physical attributes—adictionary “pattern” (or category) was created, containing a series of wordsassociated with that concept. For each category we created, we performed avalidity check. Sixty-three respondents were shown word lists from eachcategory and asked to match them to a concept. The agreement was extremelyhigh, ranging from 89 percent to 100 percent, and averaging 95 percent, givingus a great deal of confidence that we are, in fact, capturing the concepts we

7 Examples of these blogs include The Moderate Voice at http://themoderatevoice.com/ (neutral),ComMITTed to Romney at http://committedtoromney.com/ (conservative), and BarackOblogger at http://www.barackoblogger.com/ (liberal).8 These dictionaries have been extensively analyzed and validated. Regarding LIWC, see Kahnand others (2007) and Pennebaker, Mehl, and Niederhoffer (2003). See Laver, Benoit, and Garry(2003) and Laver and Garry (2000) for more information on the Laver and Garry policy positionsdictionary.


http://themoderatevoice.com/

http://committedtoromney.com/

http://www.barackoblogger.com/

describe.9 We then compared the “Palin-only” material with the “Biden-only”material, and each to the coverage of both candidates, within each medium. Thefinal dictionary employed can be found in Appendix A. There are trade-offs forusing any technique, and computer-assisted content analysis is no exception.While we are able to consider a large volume of media coverage, the analysis isnot as fine-grained as it would be with traditional hand coding. However, thelaw of large numbers allows us confidence that any trends we find are not due tochance, but rather systematic differences between types of coverage.

We report differences between the candidates in two ways. First, we use riskratios, which are a measure reflecting the probabilistic differences betweenproportions of words used in the first sample (Palin-only coverage) and thesecond sample (Biden-only coverage). In essence, the first is divided by thesecond, producing a ratio. If the ratio is greater than one, it meansthe proportion is greater in the first sample, thus reflecting greater likelihood ofthe content occurring in Palin coverage rather than Biden coverage.10 The ratioof proportions has the benefit of being independent of disparities in volumecandidates receive, allowing us to compare across different sample sizes withease. A 95 percent confidence interval is generated, and if that interval does notcross one, the difference is statistically significant (Lowe 2006). The further theratio is from one, the greater the difference between the two samples (i.e.,between Palin coverage and Biden coverage).

For ease of interpretation, percent change is also reported. This simplyindicates how much more likely content is to occur in one candidate’s coverageversus the other—positive percent changes reflect a greater likelihood of thecontent occurring in Palin content. The risk ratios and percent changes ofdifferences that are not significant are less clear to interpret. While they mayindicate parity in coverage, insignificant risk ratios may also be an artifact of asample too small. Risk ratios, reported with 95 percent confidence intervals andpercent difference between samples, subdivided by medium, are found inTables 1a and 1b.

Results

Amount of CoverageOur first hypothesis predicted that Palin would receive more news coverage

due to the novelty of her candidacy. Was this the case? The answer is aresounding yes. In overall coverage, in every medium, in every submedium, andin every individual source we examined, Palin received more media coverage

9 Full results available from the authors upon request at [email protected] When a single candidate’s coverage is being compared with coverage of both candidates, thecandidate is in the first group and the coverage of both is in the second group. Thus a ratio greaterthan one in this instance suggests that the proportion of words is greater in the single candidatecontent than in the coverage of both candidates.



Tab

le1a

.R

isk

Rat

ioan

dP

erce

ntC

hang

esof

Diff

eren

ces

ofP

alin

-onl

yan

dC

ombi

ned

Med

iaC

over

age

New

spap

erC

over

age

Tel

evis

ion

Cov

erag

eB

log

Cov

erag

e

%C

hang

eR

isk

Rat

io%

Cha

nge

Ris

kR

atio

%C

hang

eR

isk

Rat

io

Fam

ily36

.64

1.37

[1.2

7,1.

47]*

28.5

21.

29[1

.21,

1.37

]*2.

471.

16[.9

9,1.

35]

Ele

ctab

ility

-30.

09.7

7[.7

2,.8

2]*

-2.6

1.9

7[.9

3,1.

02]

-43.

25.6

82[.6

0,.7

7]*

Phy

sica

lapp

eara

nce

15.1

11.

15[1

.06,

1.25

]*-3

.20

.97

[.91,

1.03

]8.

251.

11[.9

6,1.

2]E

duca

tion

11.8

41.

12[1

.02,

1.23

]*-2

.34

.98

[.90,

1.06

]-1

9.07

.67

[.56,

.79]

*F

orei

gnpo

licy

-25.

49.8

0[.7

4,.8

6]*

.84

1.01

[.95,

1.07

]-6

4.45

.46

[.41,

.53]

*H

ealt

hca

re-5

.47

.95

[.83,

1.08

]4.

035

1.04

[.92,

1.17

]-2

2.48

.49

[.38,

.63]

*So

cial

issu

es10

.17

1.10

[1.0

1,1.

20]*

79.8

91.

80[1

.66,

1.96

]*47

.20

1.44

[1.3

0,1.

61]*

Tax

es/e

cono

my

-3.3

5.9

7[.9

0,1.

04]

-16.

80.8

6[.8

2,.8

9]*

-55.

25.5

2[.4

6,.6

0]*

Not

es:

Rep

orte

dva

lues

are

risk

rati

os,

orth

epr

opor

tion

ofw

ords

inth

atca

tego

ryin

Pal

inco

vera

gedi

vide

dby

the

prop

orti

onof

wor

dsin

that

cate

gory

inco

mbi

ned

cove

rage

(men

tion

sbo

thP

alin

and

Bid

en).

Rat

ios

>1,

thus

,re

flect

agr

eate

rte

nden

cyof

thos

ew

ords

toap

pear

inP

alin

-onl

yco

vera

ge,

whe

reas

rati

os<

1re

flect

agr

eate

rte

nden

cyof

thos

ew

ords

toap

pear

inth

eco

mbi

ned

cove

rage

.The

furt

her

the

rati

ois

from

1,th

egr

eate

rth

edi

ffer

ence

betw

een

the

two.

*p

<.0

5


Tab

le1b

.R

isk

Rat

ioan

dP

erce

ntC

hang

esof

Diff

eren

ces

ofB

iden

-onl

yan

dC

ombi

ned

Med

iaC

over

age

New

spap

erC

over

age

Tel

evis

ion

Cov

erag

eB

log

Cov

erag

e

%C

hang

eR

isk

Rat

io%

Cha

nge

Ris

kR

atio

%C

hang

eR

isk

Rat

io

Fam

ily78

.57

1.79

[1.5

7,2.

03]*

18.8

61.

19[1

.10,

1.28

]*54

.63

1.55

[1.2

8,1.

87]*

Ele

ctab

ility

-43.

50.7

0[.6

4,.7

6]*

7.44

1.07

[1.0

2,1.

13]*

-74.

57.5

7[.4

7,.6

9]*

Phy

sica

lapp

eara

nce

-1.3

7.9

9[.8

7,1.

12]

-.65

.99

[.92,

1.07

]-2

4.92

.80

[.65,

.99]

*E

duca

tion

-5.9

1.9

4[.8

1,1.

09]

-26.

67.7

9[.7

1,.8

8]*

-2.8

5.9

7[.7

7,1.

23]

For

eign

polic

y51

.04

1.51

[1.3

6,1.

68]*

49.5

31.

50[1

.41,

1.59

]*-7

.87

.93

[.79,

1.09

]H

ealt

hca

re-4

7.96

.68

[.57,

.81]

*-3

2.06

.76

[.65,

.89]

*-1

.98

.98

[.72,

1.34

]So

cial

issu

es27

.91

1.28

[1.1

0,1.

49]*

34.5

41.

35[1

.21,

1.50

]*-2

.28

.98

[.84,

1.14

]T

axes

/eco

nom

y-2

2.22

.82

[.73,

.91]

*-1

83.0

8.3

5[.3

3,.3

8]*

-49.

20.6

7[.5

5,.8

0]*

Not

es:R

epor

ted

valu

esar

eri

skra

tios

,or

the

prop

orti

onof

wor

dsin

that

cate

gory

inB

iden

cove

rage

divi

ded

byth

epr

opor

tion

ofw

ords

inth

atca

tego

ryin

com

bine

dco

vera

ge(m

enti

ons

both

Pal

inan

dB

iden

).R

atio

s>

1,th

us,

refle

cta

grea

ter

tend

ency

ofth

ose

wor

dsto

appe

arin

Bid

en-o

nly

cove

rage

,w

here

asra

tios

<1

refle

cta

grea

ter

tend

ency

ofth

ose

wor

dsto

appe

arin

the

com

bine

dco

vera

ge.T

hefu

rthe

rth

era

tio

isfr

om1,

the

grea

ter

the

diff

eren

cebe

twee

nth

etw

o.*

p<

.05.


than did Biden. This was true across time and at almost every point in time afterPalin was announced as the Republican vice presidential candidate (seeFigure 1). Although the trends for both candidates tended to run together, Palincoverage was more than 100 percent greater than Biden coverage—the averagenumber of stories per day in overall coverage, for instance, was 14.9 for Bidenand 36 for Palin.11 Keeping in mind that Biden had a two-week head start incoverage due to the fact that the Democrats announced their vice presidentialcandidate before the Republicans, this is even more impactful. This trendcontinued across all media. Palin outpaced Biden in blogs (14.2-3.2 stories perday), newspapers (15-8.2 stories per day), and television (8.8-4.6 stories per day).

Natural spikes in coverage of both candidates occur across all media at themain events of the campaign. Biden’s only comparative coverage advantageoccurs during the Democratic National Convention (August 25-28, 2008),before Palin was announced as the Republican vice presidential nominee(August 29, 2008). Palin coverage undergoes an enormous uptick at the time ofher announcement, and again during the Republican National Convention(September 1-4, 2008). Both candidates received greater amounts of coverage atimportant points in the campaign and election cycle, especially near the vicepresidential candidate debate (October 2, 2008). While there are somedifferences between media, the clear pattern is that Palin received more coveragethan Biden at every point following McCain’s announcement of her as hisrunning mate (Table 2).

Because the 2008 vice presidential contest is only one race, it is difficult tosay whether or not the differences observed in volume of coverage of Palin andBiden are gender-related, related to some other aspect like party affiliation, orcompletely normal for a vice presidential race. To gain leverage on this question,we made two descriptive comparisons to other similar races.

First, we compared the Palin–Biden coverage with the most similar caseavailable—the most recent vice presidential election without an incumbent, in2000, featuring Dick Cheney and Joe Lieberman.12 While Palin received 2.3times as much coverage as Biden overall, Cheney received slightly lesscoverage than did Lieberman.13 Because both candidates in 2000 receivedsimilar amounts of coverage, it suggests that the Palin–Biden phenomenon issomewhat unusual for a vice presidential contest, and also signifies that thedisparity in coverage between Palin and Biden was not simply due to partisanaffiliation—seeing as Republicans and Democrats were covered roughly equallyin 2000.

11 Articles including both candidates followed similar trends.12 Although Lieberman was the first Jewish vice presidential candidate, he was still a relativelytypical candidate, as a white, experienced senator.13 The 2000 sample uses the same newspapers and television programs as does that of 2008, butdoes not include blogs as their archival does not extend back that far. Search strings were identicalto those for the main sample but replacing names as appropriate.


Fig

ure

1.V

olum

eof

Cov

erag

eby

Day

ofP

alin

and

Bid

enac

ross

Med

ia


Second, perhaps it was just the novelty of Palin as a candidate, rather thanher gender, which drove the differences in coverage. To explore this possibility,we compared the overall coverage of Palin and Biden with that of the 1984 vicepresidential contest, where Geraldine Ferraro became the first womannominated for the vice presidency, and the only other woman ever nominated.If differences in coverage between Palin and Biden were due to Palin as anindividual, rather than as a woman, we would expect the Ferraro–Bushcoverage to more closely resemble the coverage of Cheney and Lieberman in2000 than that of Palin and Biden in 2008.

The coverage of Ferraro outpaced that of Bush in 1984 by about 33 percent,available coverage was limited to the New York Times and the Washington Post.While this difference is not as drastic as that of 2008, it is also quite removedfrom the parity of coverage observed in 2000. This strongly suggests that genderis a driving factor in the differences of coverage of vice presidential candidatesseen in both 1984 and 2008.

Substance of CoverageThe clear disparity in amount of coverage has the potential to exacerbate

any differences in the substance of that content, which brings us to the secondpart of our analysis (Tables 1a and 1b). Our second hypothesis suggested thatgender stereotypes would emerge in specific content and issue areas. Weanticipated that Palin would receive more coverage speaking in terms ofexperience, family, and physical appearance, and dealing with “feminine”issues, such as education, health care, and social issues. The 2008 mediacoverage we examined offers a mixed perspective (see Table 3 for risk ratios andpercent changes, and Table 2 for an overview of trends). The media did tend totalk about family (newspapers and television) and physical appearance(newspapers and blogs) significantly more in coverage of Palin than in coverageof Biden. Additionally, they were more likely to refer to social issues in Palin’s

Table 2. An Overview of the Differences in Media Coverage of Palin and Biden

Newspapers Television Blogs

Family + + -Electability - +Physical + +Education + - -Foreign policy - + -Health care - -Social issues + + +Taxes/economy - -

Notes: + Indicates this is more likely in Palin-oriented articles. - Indicates this is more likely inBiden-oriented articles.


Tab

le3.

Ris

kR

atio

and

Per

cent

Cha

nges

ofD

iffer

ence

sof

Med

iaC

over

age

ofP

alin

and

Bid

en

New

spap

erC

over

age

Tel

evis

ion

Cov

erag

eB

log

Cov

erag

e

%C

hang

eR

isk

Rat

io%

Cha

nge

Ris

kR

atio

%C

hang

eR

isk

Rat

io

Fam

ily44

.86

1.45

[1.3

5,1.

56]*

12.6

41.

13[1

.03,

1.24

]*-3

3.50

.75

[.65,

.86]

*E

lect

abili

ty-2

7.85

.78

[.74,

.83]

*5.

191.

05[.9

8,1.

12]

19.0

91.

19[1

.01,

1.41

]*P

hysi

cala

ppea

ranc

e17

.73

1.18

[1.0

9,1.

28]*

-2.8

6.9

7[.8

9,1.

06]

38.5

71.

39[1

.16,

1.65

]*E

duca

tion

12.2

71.

12[1

.02,

1.23

]*-2

8.47

.78

[.68,

.89]

*-4

5.94

.69

[.57,

.83]

*F

orei

gnpo

licy

-29.

52.7

7[.7

2,.8

3]*

51.0

61.

51[1

.41,

1.62

]*-9

9.53

.50

[.44,

.57]

*H

ealt

hca

re-5

.79

.95

[.83,

1.07

]-4

2.87

.70

[.57,

.86]

*-9

9.47

.50

[.38,

.65]

*So

cial

issu

es22

.51

1.23

[1.1

2,1.

34]*

33.9

71.

34[1

.17,

1.53

]*47

.77

1.48

[1.3

1,1.

67]*

Tax

es/e

cono

my

1.34

1.01

[.95,

1.09

]-1

61.6

9.3

8[.3

5,.4

8]*

-28.

05.7

8[.6

6,.9

3]*

Not

es:

Rep

orte

dva

lues

are

risk

rati

os,

orth

epr

opor

tion

ofw

ords

inth

atca

tego

ryin

Pal

inco

vera

gedi

vide

dby

the

prop

orti

onof

wor

dsin

that

cate

gory

inB

iden

cove

rage

.Rat

ios

>1,

thus

,re

flect

agr

eate

rte

nden

cyof

thos

ew

ords

toap

pear

inP

alin

-onl

yco

vera

ge,

whe

reas

rati

os<

1re

flect

agr

eate

rte

nden

cyof

thos

ew

ords

toap

pear

inB

iden

-onl

yco

vera

ge.

The

furt

her

the

rati

ois

from

1,th

egr

eate

rth

edi

ffer

ence

betw

een

the

two.

*p

<.0

5.


coverage than in that of Biden. We also anticipated Biden would receive greatercoverage referencing “masculine” issues, including foreign policy and theeconomy. Our results strongly confirm these expectations, in that Biden receivedsignificantly greater coverage dealing with foreign policy (newspapers andblogs) and the economy (television and blogs). Each of these trends isconfirmation of our second hypothesis.

Our hypothesis was disconfirmed in certain areas as well. While weanticipated Palin would receive more coverage dealing with education and healthcare, typically considered women’s issues, Biden actually received significantlymore coverage in each of those realms (in television and blogs for each; seeTables 2 and 3).

One element we examined was how the issue of family emerged inthe substance of candidate coverage. We used terms spanning variousfamilial relations, such as “mom,” “husband,” “parent,” and “wife,” toanalyze the content of media coverage in relation to family (see Appendix Afor full dictionary of terms). Women are traditionally associated withincreased coverage of family and motherhood, and between Palin’s new babyand her oldest daughter’s pregnancy, we expected to find a great deal ofcoverage associating Palin with family. As illustrated in Tables 2 and 3,these expectations were confirmed with Palin-oriented coverage in newspapersand television being more associated with family-related language (riskratio = 1.45 in newspapers and 1.13 in television). However, Biden was morelikely than Palin to be associated with family in the realm of blogs (riskratio = .75).

Perhaps the most interesting component of this finding confirms theimportance of examining gender and media coverage across media. Movingaway from associating women more with family and reversing this genderstereotype could be a result of how blogs as a medium work. Politicalblog readers, for instance, tend to be more politically interested than theconsumers of traditional news, so perhaps bloggers cater to that interest byproviding more issue coverage and less soft news content concerning personalbackground (Veenstra 2007). Moreover, political blogs often try to coverinformation not being reported by the mainstream media (Ekdale et al.2010). Because Palin’s background and family were covered well in traditionalmedia outlets, perhaps bloggers took it upon themselves to provide similarcoverage of her Democratic counterpart. One such example can be found inbloggers’ discussions of a family tragedy involving Biden’s first wife andchildren.

Five weeks after Biden’s improbable victory, school-teacher wife Neilia andtheir children, Beau, 3, Hunter, 2, and Naomi, 1, were crossing LimestoneRoad after Christmas shopping. Their car was T-boned, killing wife anddaughter and critically injuring the boys. All of a sudden a political careerseemed unimportant to the bereft Biden, who told his family and


Democratic leaders that he would have to relinquish his Senate seat inwaiting. (Mullen 2008, emphasis added)

The explanation for this finding could be any of these or a combination of themall. Either way, this result lends support to calls of further cross-mediumexaminations.

The literature is particularly strong in leading us to expect significantdifferences between coverage of male and female candidates with regard tophysical appearance. This section of our analysis dictionary included a varietyof terms, including different types of clothing, such as “dress,” “shirt,” “shoe,”and “tie” as well as adjectives like “attractive,” “handsome,” and “pretty.”Depictions of outward appearance, which are typically covereddisproportionately with female candidates, were covered fairly consistentlyacross the various media sources we examined. In both newspapers and blogs,Palin was covered more heavily with regard to physical appearance than wasBiden—see Table 2: “+” indicates terms more likely in Palin-oriented coverageand “-” indicates more likely in Biden-oriented coverage. This finding confirmsour expectations of seeing Palin linked to coverage regarding physicalappearance and suggests some stereotypes are alive and well—risk ratios varyfrom 1.18 to 1.39; television failed to reach significance. Interestingly, thisdifference only surfaces in newspapers and on blogs, suggesting yet again thatthere exist important differences in coverage across media.

The question of electability arose throughout the course of the 2008campaign and was often couched in terms of “experience” or “viability.” Thiswas particularly evident in public opinion polls asking about the two vicepresidential candidates. When asked to describe the vice presidential candidatesin one word, “experienced” was the word used the most to describe Biden, while“inexperienced” was at the top of the list of words for Palin (Pew ResearchCenter 2008). This attention to experience and electability in general features inour findings as well.

References to electability were significantly associated with the vicepresidential candidates in the coverage they received across all three media, butthe associations varied by medium. Mentions of experience were more likely inBiden-oriented coverage in newspapers, whereas Palin-oriented blog coveragewas more likely to include such references. We expected Palin’s coverage to bemore closely associated with discussion of this type; thus these results provideonly partial confirmation of our expectations. When comparing individualcoverage to combined coverage, articles discussing both candidates were morelikely to include discussions of electability as compared with either Palin-only orBiden-only coverage (see Tables 1a and 1b). This makes sense given the natureof the topic of experience—discussions of electability are more likely thanother issues to be comparative in nature, and thus would be more abundant incoverage that mentions both candidates as opposed to single-candidatecoverage.


While we cannot be certain about the associated tone of discussions ofexperience, further analysis of the actual texts reveal some trends. In blogs,where they discussed experience in Palin coverage more often than in Bidencoverage, a typical example of an experience discussion might look like this onefrom Matthew Kilburn:

[C]an anyone name me a single benefit provided by Palin? She failed tobring Hillary supporters to the ticket, she added no additional experienceor expertise to the ticket, and was zero help in any state. In fact, poll afterpoll has shown that, not only did the vast majority of voters believe Palinto be unqualified, but in some areas, she actually had a NEGATIVEapproval rating. (ComMITTed to Romney 2008)

A typical discussion of experience in Biden coverage looked more like this:

Biden was direct, not verbose, and his answers came crisply in contrast toObama’s more studied and sometimes pausing style of speaking. That heknew his brief was less surprising, given his experience, but he avoidedspeaking in the kind of senatorial vernacular that often hampers someonewho has been in the capital as long as he has. (Balz 2008)

In each article (of any medium) we randomly selected to examine, Biden wasnever negatively associated with experience, whereas Palin was consistentlydiscussed in terms of a lack of experience.

While our results do not indicate definitively whether or not experience andelectability were discussed more negatively or positively with either candidate,they do show that electability was significant in the political discussion of thesetwo candidates as presented by the media. The differences across media canperhaps be seen as indirect confirmation that the stereotype of women lackingexperience and experience as candidates was not a driving force in all coverage ofSarah Palin. If it were, we would expect mentions of electability to be more likelyin all Palin-oriented coverage than in Biden’s coverage. The differences betweennewspapers and blogs with regard to experience indicate another reason toexamine multiple media in investigations of gender and media coverage.

In terms of the coverage differences examined with substantive policy issues,several trends emerge. For some issues, things are as we might expect. Palin isstrongly associated with “feminine” topics, such as social issues (in everymedium), and Biden is strongly and consistently covered with respect to foreignpolicy (newspapers and blogs) and the economy (television and blogs).Moreover, one of these relationships—social issues—is the most consistent wesee across all content areas (see Table 3). Social issues were 23-48 percent morelikely to occur in Palin-only coverage, depending on the medium, the low isnewspapers, and the high is blogs. Palin, who many suspected was chosen toenergize the Republican base, was extremely strong on social issues, so this is


likely an effect of the candidate as much as the gender (Spangler 2008). Thedifferences in foreign policy coverage are quite pronounced: The foreign policypattern was more likely to occur in Biden-only coverage by as much as 100percent (blogs) and about 30 percent (newspapers). Again, this is not surprisinggiven Biden’s well-acknowledged wealth of foreign policy experience, especiallywhen compared with his Republican counterpart (Kranish 2008).

As the economy was reported on so extensively in 2008, it is particularlyinteresting that differences between the candidates’ coverage emerged at all,because everyone had to speak about the economy to even have a chance at theelection (Spangler 2008). Even so, Biden-only television and blog coverage wasmore likely to be associated with language linked to taxes and to the economythan Palin-only coverage. This finding, in combination with the solid foreignpolicy results, suggests that gender stereotypes regarding male candidates arestrongly at work in coverage of the 2008 vice presidential candidates.

In other policy areas, our expectations were disconfirmed. We expectedPalin’s coverage to be linked more closely with discussions of health care andeducation due to the notion that these issues are stereotypically considered more“feminine.” Although Palin’s newspaper coverage was more likely to be linked toeducation, Biden’s television and blog coverage was more closely tied to bothhealth care and education issues. This may be explained by a second dimension ofissue coverage, in that health care and education, although “feminine” issues, arealso generally associated with Democrats, which might indicate that Biden’sconnection to these issues is based on partisan ties as opposed to genderedassociations (Petrocik 1996). As the number of potential cases to study increaseswith future vice presidential contests, research should consider the interplay ofgender and partisan identification in the distribution of media coverage betweencandidates.

Overall, we found evidence of gender stereotypes in much of the substanceof coverage of the 2008 vice presidential candidates, partially confirming oursecond hypothesis. The coverage of Palin was more likely to include referencesto family, physical appearance, and social issues. However, not all ourexpectations of likely instances of gender stereotypes were realized. We did notsee Palin covered more in terms of electability or “women’s issues,” such aseducation and health care. This may be a result of partisan influences, orartifacts of this particular case and these specific candidates. However, theseunrealized expectations may also be a sign that gender stereotypes in relation towomen candidates are not static and are not ever-present.

To attempt to tease out the driving forces behind these trends, and to whatextent they are motivated by gender as opposed to partisan identification, weperform an additional descriptive comparison, again to the 1984 contest betweenGeraldine Ferraro and George H. W. Bush. This is a convenient case, as it againoffers a contest between a female and a male candidate, but in the 1984 contest,the partisanship of the candidates is reversed from that of 2008. Unfortunately,only newspaper coverage is available going back to 1984 (and only two of our


three newspapers, as USA Today did not yet exist), but we feel the comparison isstill worthwhile. Analysis of the coverage of Ferraro and Bush by issue area isavailable in Table 4.

Four issue areas are consistent across electoral context in achievingdisparate coverage by candidate gender, including family, education, and socialissues, for which the female candidate received greater coverage, as well asforeign policy, for which the male candidate received greater attention. Thissuggests that these issues, at least, are not driven by partisanship, but rather bythe candidate’s gender. The two topics for which we coded that are inconsistentbetween 1984 and 2008 are quite interesting—electability and physicalappearance. These are strongly suggested by the literature to be topics coveredwith regard to female candidates, but we find inconsistent results depending onthe election. Whereas Palin was covered by newspapers more than Biden interms of physical appearance, Bush surprisingly outpaced Ferraro in this realm.And while Biden received the bulk of newspaper coverage related to experience,Ferraro actually received more experience-related coverage than Bush. Thissuggests that the expectations of the literature may not apply as easily to the vicepresidential context, at least in the case of coverage of physical appearance andexperience.

These results, although preliminary, could also point to a shift where genderstereotypes of women as candidates are no longer necessarily pervasive in mediacoverage. Moreover, this shift may have begun as early as the 1980s—far earlierthan most scholarship has suggested in the past. Both of these possibleexplanations serve as starting points for further investigation.

Table 4. Risk Ratio and Percent Changes of Differences of Media Coverage ofFerraro and Bush

Newspaper Coverage

% Change Risk Ratio

Family 42.64 1.43 [1.36, 1.49]*Electability 42.43 1.42 [1.33, 1.52]*Physical appearance -8.48 .92 [.88, .96]*Education 6.19 1.06 [1.00, 1.12]*Foreign policy -52.90 .65 [.62, .69]*Health care -11.90 .89 [.81, .99]*Social issues 35.35 1.35 [1.29, 1.42]*Taxes/economy -25.27 .80 [.77, .83]*

Notes: Reported values are risk ratios, or the proportion of words in that category in Ferrarocoverage divided by the proportion of words in that category in Bush coverage. Ratios > 1, thus,reflect a greater tendency of those words to appear in Ferraro-only coverage, whereas ratios < 1reflect a greater tendency of those words to appear in Bush-only coverage. The further the ratiois from 1, the greater the difference between the two. * p < .05


An additional avenue to consider when evaluating the use of genderstereotypes in political candidates’ media coverage is to note how genderstereotypes, and any changes in their usage, are applied to both women andmen candidates. If the gender stereotypes used in the coverage of women areshifting, or perhaps even fading, what is happening to the stereotypicalnotions typically associated with male candidates? Are the changes in howwomen political candidates are covered by the media reflected in any way bycorresponding shifts in the coverage of men? Although not addressed directlyby our article, these questions are worthy of further discussion and academicattention.14

Differences among Media

Our third hypothesis dealt specifically with the three media sources weexamined. As a result of different norms, motivations, and standards, weexpected variation of candidate coverage by medium. Again, the picture ismixed, but we see partial support for our third hypothesis. We see significantdifferences between media in about half the areas we examined, includingcoverage of family, electability, education, and foreign policy. In other areas,however, there are no significant differences between the medium consideredand the tilt of the coverage toward one candidate or another. This is true formost policy areas we examined, including the issues of health care, socialissues, and the economy, as well as for coverage dealing with physicalappearance.

When we made the decision to examine media coverage from a number ofsources, we did so because we expected to see differences in how various mediaformats covered the 2008 vice presidential election. Overall, we saw moreagreement between sources than we expected to find. Slightly more than half ofthe areas we examined saw a clear trend toward coverage of one candidate oranother in at least two of the three types of media we examined. Yet differencesin volume of coverage of the candidates and the proportional attention paid toeach varied widely between media outlets. Moreover, as indicated above, manyof these differences are notable.

While each media outlet covered Governor Palin by herself much more thanthey covered only Senator Biden, the extent to which such differences emergedvaried quite a bit between newspaper, television, and blog coverage. Newspapers

14 Scholarship on the intersection of celebrity politics and gender considers such questions. Thisintersection has been examined by van Zoonen (2006) in her analysis of media coverage of twofemale, European heads of state (Finland’s Tarja Holonen and Germany’s Angela Merkel). Forexample, while van Zoonen (2006, 287) argues that the “increasing presence of popular culture inpolitics presents a complex and often unfavorable arena to women,” she notes that the media’semphasis on appearance and style may no longer be a gendered phenomenon, but a focus putupon both men and women politicians due to a general shift in the role of celebrity and fame inmedia coverage of politics (297-8).


in our sample had a total of 924 Palin stories, or 2.3 times more coverage than thatreferring to Senator Biden. For television, the differences are similar, with 324Palin-only stories, 3.3 times the amount of Biden-only coverage. For blogs, thesedifferences are even more pronounced, with Palin’s 950 blog posts representinga surprising 80 percent of total blog coverage of the 2008 vice presidentialcandidates. This is more than seven times the amount of content blogs devoted toBiden during our time period. While each of these differences indicates a strongbias toward coverage of Palin, they also show to what extent volume disparitiesemerge by medium. Just in terms of volume, for instance, a blog reader wouldhave seen a very different picture of the 2008 vice presidential candidates thanmight a newspaper reader. When considered in combination with the differencesin content, this becomes exceedingly important.

Volume aside, we found significant differences in the substance of coveragebetween media in several areas, including family references, education, andforeign policy. It is not immediately clear why differences would emerge onthese specific issues. They are a mix of policy issues, including both “feminine”and “masculine” issues, as well as coverage of personal matters andbackground. Where there were no differences between media, there is similarlylittle obvious pattern, including several policy areas as well as coverage ofphysical appearance

The differences between media are not consistent across issue areas. For someissues, such as education, newspapers appear to be the exception to the rule. Forcoverage based on foreign policy, television news seems to cover candidatesdifferently than other media outlets. And for coverage of the candidate’s family,blogs are the exception. There is no clear hard news versus soft news, or old versusnew media divide. For several of these issues, including coverage of foreign policyand family, the results are somewhat strange. Biden was specifically chosen for hisforeign policy experience, so it is odd that television news would cover foreignpolicy and Governor Palin more often than they would cover Biden and foreignpolicy. Similarly, Palin’s family was a hot topic of conversation during theelection season, and yet our sample of blogs was less likely to cover her inconjunction with language concerning family than with Biden. Our data do notallow us to completely understand these anomalies, but future research shouldboth replicate our methods and push further on the question to understand whenand how differences in coverage emerge between media.

Again, while these differences are not necessarily intuitive, we think themost important thing to note is that they exist at all. Our results indicate thatcoverage of candidates varies not only by issue area, but also by medium, andfuture research should consider both factors when studying how vicepresidential candidates are covered by the media. Failing to do so will result inan incomplete picture of electoral media coverage. As media options proliferate,this becomes even more important to consider, as the choices people makeconcerning where to gain information may have strong implications for whatkind of candidate coverage they receive.


Limitations

While this article represents an important expansion of the literature ongender differences in media coverage, it is limited by several factors. First, thereis the issue of the dictionary we created for computer-assisted content analysis.Obviously, not every potential word in each category was included, and thiscould take away some degree of certainty from our results. Additionally, ourexamination does not capture the tone of the coverage analyzed, whichprohibits us from saying conclusively exactly how coverage treated eachcandidate. However, the risk ratios, by comparing proportions of words ratherthan simple counts, allow us to assess the relative frequency of our categories ofinterest, which should help to mitigate the problem of missing words andsentiment. Moreover, our reliance on other dictionaries in building our own andour validity checks of dictionary categories should help remove any systematicbias in pattern creation.

A second limitation is that of our sample. While our television sampleis relatively complete with regard to national television news, we analyzed thetranscripts only and not the visual components of these broadcasts, andincluded national but not local broadcasts. Also, we examined only threenewspapers, and a systematic sample of blogs. The three newspapers areall national in scope, thus removing any local news coverage from ouruniverse of media coverage. We think this makes sense, giventhat this article is one of the few to consider a truly national contest, butthere are likely to be systematic differences in coverage between national andlocal newspapers, and this is something for which we cannot account (Rozell1991). However, given local news’ limited resources, it is likely that nationalnews drives local coverage, especially through the provision of wire stories,which may make differences between national and local coverage lesspronounced. We also chose a convenient sample of blogs from a largeruniverse of all blogs. It is possible that this sample is systematically biased insome way, and thus fails to present a true picture of blog coverage.Nevertheless, the number of blogs included, as well as their obvious diversity(conservative and liberal, local and national, Democrat and Republican, andissue-oriented) assures us that this is a suitable sample of blog coverage(see Appendix B).

Additionally, we consider only two election contests, for a single office, andone with limited available coverage. While this provides for an interesting andimportant test case, it offers an N of only two, which decreases our ability togeneralize from our results. We also cannot say to what extent our results aredriven by associations between vice presidential and presidential candidates.However, these races have the potential to affect perceptions of candidates,parties, and genders, and are important as coverage and perceptions of both raceswill affect which candidates are chosen in the future, and how they are perceivedand covered.


Future Research

We think future research should similarly focus on differences betweenmedia when considering other levels of office seeking, including statewide andlocal races. Differences between “old” and “new” media are particularly ripefor continued observation, and these considerations may be expanded beyondthe realm of blogs to consider other online news sites, and soft news coverageon the web. Concomitantly, future research should consider drawingconnections between media coverage and other content. For instance, it wouldbe of use to consider how media coverage compares to campaign-generatedcontent, such as speeches and press releases, to better understand thedifferences in media coverage. Comparing volume and substance of mediacoverage over time to polling data on the vice presidential candidates mightalso reveal interesting patterns, and addressing classic questions of agendasetting: Do polls lead the media or does the agenda setting power of the mediainfluence the public and drive poll results? Comparing coverage over time topublic perceptions of candidates through opinion polls, and public perceptionof the most important problem facing the country might help reveal theimpact of media coverage of the vice presidential candidates in 2008. Finally,differences in media coverage uncovered by this article might be paired withexperiments to determine what the effects of such differences might be on thevoting public.

Conclusion

Our article represents an important step forward in research on genderdifferences in media coverage. Most notably, we add an important case—that ofthe vice presidential contest—to consider. The Palin–Biden case is particularlyrelevant given the attention it received by the media, and offers a uniquecomparison to the legendary 1984 contest. Perhaps even more important, ourarticle includes three separate media and uncovers noteworthy differences incoverage between and within them. The importance of these findings becomeseven more transparent when we consider the implications of the differences wediscovered in media coverage of the vice presidential candidates (see Table 5).

What are the substantive effects of these differences? Depending on one’spreference for or access to different media, the information to be gainedconcerning the 2008 vice presidential candidates differs markedly. Three keydifferences should be noted. First, across the three media the substance ofcoverage differs for each candidate. Second, across the three media one candidatereceives more coverage than the other. Between Palin and Biden, the amount ofcoverage Palin received outweighed the amount received by Biden. Finally, notonly did one candidate receive more coverage than the other, but the differencesare quite large.


Each of these differences could have had important influences on publicopinion formation and the public’s voting decisions in this particular race, andour findings are generally replicated with the case of 1984, although with limiteddata. If future research finds similar differences in coverage in other races, theimplications could significantly affect the election process, particularly for femalecandidates. The key to revealing the extent of such differences, and the resultingimplications for candidates of any gender is continued analysis of politicalcoverage across all media.

Table 5. Amount and Substance of Palin and Biden Media Coverage by Medium

Newspapers Television Blogs

Palin Amount Amount Amount924 articles 324 stories 950 entries

Substance Substance SubstanceMentions of family Mentions of family Mentions of

electabilityMentions of physicalappearance

Issue coverage:foreign policy, socialissues

Mentions of physicalappearanceIssue coverage: socialissues

Issue coverage:education, socialissues

Biden Amount Amount Amount402 articles 97 stories 123 entries

Substance Substance SubstanceMentions ofelectability

Issue coverage:education, healthcare, taxes/economy

Mentions of family

Issue coverage:foreign policy

Issue coverage:education, foreignpolicy, health care,taxes/economy


Appendix A: Content Analysis Dictionary

FAMILYaunt*brother*cousin*dad*daughter*exfamilies*familyfather*folksgrandchil*granddad*granddau*grandf*grandm*grandpa*grandson*great-*husband*kinmommomma*mommy*momsmom’smother*mummummy*mumsmum’snephew*niece*parent*relativessister*sonsons

son’sspouse*step-*stepfat*stepmot*uncleunclesuncle’swife*wive*teen*pregnant

PHYSICALAPPEARANCEhairmakeupdress*pant*skirt*tie*wardrobesuit*shoe*heel*boot*jacketcoatscarfglassescloth*attractiveunattractiveprettybeautifulhandsomeuglysmile

eye*leg*arm*hand*feetjewelrywatchcufflinksjeansblouse*shirt*

TAXES/ECONOMYglobaltax*economyspendingdeficitrecessiondepressionjob*stimulusfiscalbanks*bailoutrecoveryretir*saving*

EDUCATIONeducat*school*learn*collegeuniversityPellK-12

pre-schoolpreschooltest*teach*

HEALTH CAREhealth*doctor*hospital*medicinemedicalinsurancemedicaidmedicareprescriptiondrugnurse

SOCIALISSUESabort*punishmentmarijuanadrug*pray*gayhomosexualcivilrightspro-lifepro-choiceimmigrantimmigrat*religio*godjesuslordchristchurch


FOREIGNPOLICYforeignmilitarywarIraqAfghanistan

soldier*troop*enemycombatantcombatterror*security

border

ELECTABILITYexperience*viabilityviablegovernor

senatormayorskillknow-howwisdomreadyreadiness

Appendix B: Source Information

Newspapers Included:New York TimesUSA TodayWashington Post

Television Transcripts Included:ABCCBSCNNFOXNBC

Blogs Included: Newstex Government & Politics Blogs*List of Blogs Covering Vice Presidential Candidates, August 16-November 11, 2008

Liberal Conservative Neutral Could Not Be Determined

All Spin Zone 7.62mm Justice 2008Central.net PoliticsBarack Oblogger Adam Graham DC Download Politics, Sports, AnythingCapitol Annex BatesLine Face The State What I KnowD-Day ComMITTed to Romney Inside PoliticsDallas South Conservatives with

AttitudeMideast Youth

DonklephantDaniel W. Drezner

Nation-BuildingLiberal Values

Fausta’s BlogPolitical Fix

Manchester SquareFITSNews For Now

Pop and PoliticsNye-Gateway to Nevada’s

Rurals GayPatriotThe Moderate Voice

Saudi JeansGirl on the Right

Trail Blazers BlogScholars and Rogues

Joust the FactsSocratic GadflyJust Average AmericanSpin CycleLibertarian LeaningsTappedLiberty MavenThe Democratic DailyNew England RepublicanThe GistNH InsiderThe Mo’Kelly ReportNOVA TownHall BlogThe South ChicagoanRandom JottingsThe World Around YouReality HammerVernaSmith.comThe Jawa ReportWonkette

WriteChic Press22 blogs 20 blogs 10 blogs 3 blogs

* Please contact the authors at [email protected] for a list of the 391 blogs archived by LexisNexis in theNewstex Government & Politics Blogs section.



About the Authors

Leticia Bode is a PhD candidate in the Department of Political Science at theUniversity of Wisconsin-Madison. Her research interests span the fields ofpolitical science and communication, and include political communication,political behavior, and new media and politics. Her current major researchproject focuses on the flow of political information between mass and elites viasocial media.

Valerie M. Hennings is the scholar-in-residence at the Carrie Chapman CattCenter for Women and Politics and an adjunct assistant professor of politicalscience at Iowa State University. Her research interests include gender andpolitics, political participation, political communication, and state and localgovernment and policy. Her current research project focuses on how candidatetraining programs influence women’s decisions to run for office.

References

Aday, Sean, and James Devitt. 2001. “Style Over Substance: NewspaperCoverage of Elizabeth Dole’s Presidential Bid.” Harvard International Journalof Press/Politics 6 (2): 61-73. Accessed on January 9, 2012. Available onlineat http://www.deepdyve.com/lp/sage/style-over-substance-newspaper-coverage-of-elizabeth-dole-s-InT0Yan4qO

Alexander, Deborah, and Kristi Andersen. 1993. “Gender as a Factor in theAttribution of Leadership Traits.” Political Research Quarterly 46 (3): 527-545.Accessed on January 9, 2012. Available online at http://prq.sagepub.com/content/46/3/527.abstract

Ansolabehere, Steven, Roy Behr, and Shanto Iyengar. 1993. The MediaGame: American Politics in the Television Age. New York: Macmillan.

Atkeson, Lonna Rae, and Timothy B. Krebs. 2008. “Press Coverage ofMayoral Candidates: The Role of Gender in News Reporting and CampaignIssue Speech.” Political Research Quarterly 61 (2): 239-252. Accessed onJanuary 9, 2012. Available online at http://prq.sagepub.com/content/61/2.toc

Atkinson, Lucy. 2007. “Online and Ticked Off? An Exploration of OnlinePolitical News Coverage and Hostile Media Phenomenon.” Paper presented atthe Annual Meeting of the International Communication Association, SanFrancisco. May 23-28.

Balz, Dan. 2008. “Palin Delivers, But Doubts Linger.” Washington Post.Accessed on December 20, 2011. Available online at http://www.washingtonpost.com/wp-dyn/content/article/2008/10/03/AR2008100300074.html


http://prq.sagepub.com/content/46/3/527.abstract

http://prq.sagepub.com/content/46/3/527.abstract

http://prq.sagepub.com/content/61/2.toc

http://www.washingtonpost.com/wp-dyn/content/article/2008/10/03/AR2008100300074.html

http://www.washingtonpost.com/wp-dyn/content/article/2008/10/03/AR2008100300074.html

Bartels, Larry M. 1993. “Messages Received: The Political Impact of MediaExposure.” American Political Science Review 87 (2): 267-285. Accessed onJanuary 9, 2012. Available online at http://www.jstor.org/pss/2939040

Baum, Matthew A., and Tim Groeling. 2008. “New Media and thePolarization of American Political Discourse.” Political Communication 25 (4):345-365. Accessed on January 9, 2012. Available online at http://www.tandfonline.com/doi/abs/10.1080/10584600802426965

Bystrom, Dianne G., Mary Christine Banwart, Lynda Lee Kaid, andTerry A. Robertson. 2004. Gender and Candidate Communication: VideoStyle,WebStyle, NewsStyle. New York: Routledge.

Bystrom, Dianne G., Terry A. Robertson, and Mary Christine Banwart.2001. “Framing the Fight: An Analysis of Media Coverage of Female and MaleCandidates in Primary Races for Governor and U.S. Senate in 2000.” AmericanBehavioral Scientist 44 (12): 1999-2013. Accessed on January 9, 2012. Availableonline at http://www.deepdyve.com/lp/sage/framing-the-fight-an-analysis-of-media-coverage-of-female-and-male-EM4Tc9NtDy

Carroll, Susan J., and Ronnee Schreiber. 1997. “Media Coverage of Womenin the 103rd Congress.” In Women, Media, and Politics, edited by Pippa Norris.New York: Oxford University Press. 131-148.

Center for American Women and Politics. 2010. “Facts on WomenOfficeholders, Candidates and Voters.” Eagleton Institute of Politics. Accessedon December 20, 2011. Available online at http://www.cawp.rutgers.edu/fast_facts/index.php

Center for Media and Public Affairs. 2009. “Election Watch: Campaign2008 Final: How TV News Covered the General Election Campaign.” Centerfor Media and Public Affairs. Accessed on December 20, 2011. Available onlineat http://www.cmpa.com/pdf/media_monitor_jan_2009.pdf

Chaffee, Steven, and Stacey Frank. 1996. “How Americans Get PoliticalInformation: Print versus Broadcast News.” Annals of the American Academy ofPolitical and Social Science 546: 48-58.

Clarke, Peter, and Eric Fredin. 1978. “Newspapers, Television and PoliticalReasoning.” Public Opinion Quarterly 42 (2): 143-160. Accessed on January 9,2012. Available online at http://poq.oxfordjournals.org/content/42/2/143.abstract

ComMITTed to Romney. 2008. “Autopsy on the McCain Campaign.”ComMITTed to Romney. Accessed on December 20, 2011. Available online athttp://committedtoromney.com/2008/11/05/autopsy-on-the-mccain-campaign/

Devitt, James. 1999. Framing Gender on the Campaign Trail: Women’s ExecutiveLeadership and the Press. Washington, DC: Women’s Leadership Fund.


http://www.jstor.org/pss/2939040

http://www.tandfonline.com/doi/abs/10.1080/10584600802426965


http://www.cawp.rutgers.edu/fast_facts/index.php

http://www.cawp.rutgers.edu/fast_facts/index.php

http://www.cmpa.com/pdf/media_monitor_jan_2009.pdf

http://poq.oxfordjournals.org/content/42/2/143.abstract


http://committedtoromney.com/2008/11/05/autopsy-on-the-mccain-campaign/

Ekdale, Brian, Kang Namkoong, Timothy Fung, and David Perlmutter.2010. “Why Blog? (Then and Now): Exploring the Motivations for Blogging byPopular American Political Bloggers.” New Media and Society 12 (2): 217-234.Accessed on January 9, 2012. Available online at http://nms.sagepub.com/content/early/2010/01/19/1461444809341440.abstract

Everitt, Joanna, and Elisabeth Gidengil. 2003. “Tough Talk: HowTelevision News Covers Male and Female Leaders of Canadian PoliticalParties.” In Women and Electoral Politics in Canada, edited by ManonTremblay and Linda Trimble. Toronto: Oxford University Press. 194-210.

Falk, Erika. 2008. Women for President: Media Bias in Eight Campaigns.Chicago, IL: University of Illinois Press.

Foreit, Karen G., Terna Agor, Johnny Byers, John Larue, Helen Lokey,Michael Palazzini, Michele Patterson, and Lillian Smith. 1980. “Sex Biasin the Newspaper Treatment of Male-Centered and Female-Centered NewsStories.” Sex Roles 6 (3): 475-480. Accessed on January 9, 2012. Availableonline at http://www.springerlink.com/content/p7463j1589480132

Garramone, Gina M., and Charles K. Atkin. 1986. “Mass Communicationand Political Socialization: Specifying the Effects.” Public Opinion Quarterly 50(1): 76-86. Accessed on January 9, 2012. Available online at http://poq.oxfordjournals.org/content/50/1/76.abstract

Gorman, Steve. 2008. “Palin-Biden Debate Sets TV Ratings Record.” Reuters.Accessed on December 20, 2011. Available online at http://www.reuters.com/article/2008/10/03/us-debate-record-idUSTRE4928VT20081003

Haynes, Audrey A., and Brian Pitts. 2009. “Making an Impression: NewMedia in the 2008 Presidential Nomination Campaigns.” PS: Political Scienceand Politics 42 (1): 53-58. Accessed on January 9, 2012. Available online athttp://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=3260564

Heflick, Nathan A., and Jamie Lynn Goldenberg. 2010. “Sarah Palin, aNation Object(ifie)s: The Role of Appearance Focus in the 2008 U.S.Presidential Election.” Sex Roles 65 (3): 149-155. Accessed on January 9, 2012.Available online at http://www.springerlink.com/content/qu45n74354472186

Heith, Diane J. 2003. “The Lipstick Watch: Media Coverage, Gender,and Presidential Campaigns.” In Anticipating Madam President, edited byRobert P. Watson and Ann Gordon. Boulder, CO: Lynne Rienner Publishers.123-130.


http://nms.sagepub.com/content/early/2010/01/19/1461444809341440.abstract

http://nms.sagepub.com/content/early/2010/01/19/1461444809341440.abstract

http://www.springerlink.com/content/p7463j1589480132



http://www.reuters.com/article/2008/10/03/us-debate-record-idUSTRE4928VT20081003

http://www.reuters.com/article/2008/10/03/us-debate-record-idUSTRE4928VT20081003

http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=3260564


http://www.springerlink.com/content/qu45n74354472186

Heldman, Caroline, Susan J. Carroll, and Stephanie Olson. 2005. “SheBrought Only a Skirt: Print Media Coverage of Elizabeth Dole’s Bid for theRepublican Presidential Nomination.” Political Communication 22 (3): 315-335.Accessed on January 9, 2012. Available online at http://www.tandfonline.com/doi/abs/10.1080/10584600591006564

Herrnson, Paul S., Atiya Kai Stokes-Brown, and Matthew Hindman.2007. “Campaign Politics and the Digital Divide: Constituency Characteristics,Strategic Considerations, and Candidate Internet Use in State LegislativeElections.” Political Research Quarterly 60 (1): 31-42. Accessed on January 9,2012. Available online at http://prq.sagepub.com/content/60/1.toc

Hinojosa, Magda. 2010. “‘She’s Not My Type of Blonde’: Media Coverage ofIrena Sáez’s Presidential Bid.” In Cracking the Highest Glass Ceiling: A GlobalComparison of Women’s Campaigns for Executive Office, edited by RainbowMurray and Pippa Norris. Boulder, CO: Praeger. 31-48.

Howard, Phillip N. 2005. “Deep Democracy, Thin Citizenship: The Impact ofDigital Media in Political Campaign Strategy.” Annals of the AmericanAcademy of Political and Social Science 597: 153-170.

Huddy, Leonie. 1994. “The Political Significance of Gender Stereotypes.” InResearch in Micropolitics: New Directions in Political Psychology, edited byMichael X. Delli-Carpini, Leonie Huddy, and Robert Y. Shapiro. Greenwich,CT: JAI Press. 169-193.

Huddy, Leonie, and Nayda Terkildsen. 1993. “Gender Stereotypes and thePerception of Male and Female Candidates.” American Journal of PoliticalScience 37 (1): 119-147. Accessed on January 9, 2012. Available online athttp://www.jstor.org/pss/2111526

Iyengar, Shanto, and Donald R. Kinder. 1987. News That Matters: TV andAmerican Opinion. Chicago, IL: University of Chicago Press.

Jalalzai, Farida. 2006. “Women Candidates and the Media: 1992-2000Elections.” Politics and Policy 34 (3): 606-633. Accessed on January 9, 2012.Available online at http://onlinelibrary.wiley.com/doi/10.1111/j.1747-1346.2006.00030.x/abstract

Jamieson, Kathleen Hall. 1995. Beyond the Double Bind: Woman andLeadership. Oxford: Oxford University Press.

Johnson, Thomas, and Barbara Kaye. 2004. “Wag the Blog: How Reliance onTraditional Media and the Internet Influence Credibility Perceptions ofWeblogs among Blog Users.” Journalism and Mass Communication Quarterly81 (3): 622-642.




http://prq.sagepub.com/content/60/1.toc




Kahn, Jeffrey H., Renee M. Tobin, Audra E. Massey, and Jennifer A.Anderson. 2007. “Measuring Emotional Expression with the Linguistic Inquiryand Word Count.” American Journal of Psychology 120 (2): 263-286. Accessedon January 9, 2012. Available online at http://www.jstor.org/pss/20445398

Kahn, Kim Fridkin. 1992. “Does Being Male Help: An Investigation of Genderand Media Effects in U.S. Senate Races.” Journal of Politics 54 (2): 497-517.Accessed on January 9, 2012. Available online at http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=6276960

___. 1994a. “The Distorted Mirror: Press Coverage of Women Candidates forStatewide Office.” Journal of Politics 56 (1): 154-173. Accessed on January 9,2012. Available online at http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=6285576

___. 1994b. “Does Gender Make a Difference? An Experimental Examinationof Sex Stereotypes and Press Patterns in Statewide Campaigns.” AmericanJournal of Political Science 38 (1): 162-195. Accessed on January 9, 2012.Available online at http://www.jstor.org/pss/2111340

___. 1996. The Political Consequences of Being a Woman. New York, NY:Columbia University Press.

Kahn, Kim Fridkin, and Edie Goldenberg. 1991. “Women Candidates in theNews: An Examination of Gender Differences in U.S. Senate Campaigns.”Public Opinion Quarterly 55 (2): 180-199.

Ketchum and USC Annenberg Strategic Public Relations Center. 2009.“Media Myths and Realities: Media Usage Survey.” Ketchum Perspectives.Accessed on December 20, 2011. Available online at http://ketchumperspectives.com/archives/2009_i1/index.php

Kittilson, Miki Caul, and Kim Fridkin. 2008. “Gender, Candidate Portrayalsand Election Campaigns: A Comparative Perspective.” Politics and Gender 4 (3):371-392. Accessed on January 9, 2012. Available online at http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=2183560

Koch, Jeffrey. 2002. “Gender Stereotypes and Citizens’ Impressions of HouseCandidates’ Ideological Orientations.” American Journal of Political Science 46(2): 453-462. Accessed on January 9, 2012. Available online at http://www.jstor.org/pss/3088388

Kranish, Michael. 2008. “Foreign Policy Got Biden the VP Nod.” The BostonGlobe. Accessed on December 20, 2011. Available online at http://www.boston.com/news/nation/articles/2008/08/24/foreign_policy_got_biden_the_vp_nod








http://ketchumperspectives.com/archives/2009_i1/index.php

http://ketchumperspectives.com/archives/2009_i1/index.php





http://www.boston.com/news/nation/articles/2008/08/24/foreign_policy_got_biden_the_vp_nod



Laver, Michael, Kenneth Benoit, and John Garry. 2003. “Extracting PolicyPositions from Political Texts Using Words as Data.” American PoliticalScience Review 97 (2): 311-331. Accessed on January 9, 2012. Available online athttp://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=152187

Laver, Michael, and John Garry. 2000. “Estimating Policy Positions fromPolitical Texts.” American Journal of Political Science 44 (3): 619-634. Accessedon January 9, 2012. Available online at http://unoacademia.ch/webdav/site/mia/users/Imene_Ajala/public/laver-estimatingpol.pdf

Lawrence, Regina G., and Melody Rose. 2010. Hillary Clinton’s Race for theWhite House: Gender Politics and the Media on the Campaign Trail. Boulder,CO: Lynne Rienner Publishers.

Lee, Jae Kook. 2007. “The Effect of the Internet on Homogeneity of the MediaAgenda: A Test of the Fragmentation Thesis.” Journalism and MassCommunication 84 (4): 745-760.

LexisNexis. 2009. LexisNexis Academic. Albany, NY: Reed Elsevier, Inc.Articles accessed Spring 2009.

Lowe, William. 2006. “Yoshikoder: An Open Source Multilingual ContentAnalysis Tool for Social Scientists.” Yoshikoder. Accessed on December 20,2011. Available online at http://www.yoshikoder.org/courses/apsa2006/apsa-yk.pdf

McLuhan, Marshall. 1994. Understanding Media: The Extensions of Man.Cambridge, MA: MIT Press.

Mullen, Shaun. 2008. “The Joe Biden That I Know & Admire.” The ModerateVoice. Accessed on December 20, 2011. Available online at http://themoderatevoice.com/21887/the-joe-biden-that-i-know-admire/

Murray, Rainbow. 2010. “Madonna and Four Children: Ségolène Royal.” InCracking the Highest Glass Ceiling: A Global Comparison of Women’s Campaignsfor Executive Office, edited by Rainbow Murray. Denver, CO: Praeger. 49-68.

Newstex. 2009. “Authoritative Blog Contents (Publishers).” Newstex. Accessedon December 20, 2011. Available online at http://newstex.com/publishers/blogs-on-demand

Pennebaker, James W., Matthias R. Mehl, and Kate G. Niederhoffer. 2003.“Psychological Aspects of Natural Language Use: Our Words, Our Selves.”Annual Review of Psychology 54: 547-577. Accessed on January 9, 2012.Available online at http://www.annualreviews.org/doi/abs/10.1146/annurev.psych.54.101601.145041?journalCode=psych




http://unoacademia.ch/webdav/site/mia/users/Imene_Ajala/public/laver-estimatingpol.pdf

http://unoacademia.ch/webdav/site/mia/users/Imene_Ajala/public/laver-estimatingpol.pdf

http://www.yoshikoder.org/courses/apsa2006/apsa-yk.pdf

http://www.yoshikoder.org/courses/apsa2006/apsa-yk.pdf

http://themoderatevoice.com/21887/the-joe-biden-that-i-know-admire/

http://themoderatevoice.com/21887/the-joe-biden-that-i-know-admire/

http://newstex.com/publishers/blogs-on-demand

http://newstex.com/publishers/blogs-on-demand

http://www.annualreviews.org/doi/abs/10.1146/annurev.psych.54.101601.145041?journalCode=psych

http://www.annualreviews.org/doi/abs/10.1146/annurev.psych.54.101601.145041?journalCode=psych

Petrocik, John R. 1996. “Issue Ownership in Presidential Elections, with a 1980Case Study.” American Journal of Political Science 40 (3): 825-850. Accessed onJanuary 9, 2012. Available online at http://www.jstor.org/pss/2111797

Pew Project for Excellence in Journalism. 2009. “The State of the NewsMedia: An Annual Report on American Journalism 2008.” State of the Media.Accessed on December 20, 2011. Available online at http://stateofthemedia.org/2008

Pew Research Center. 2008. “Presidential Race Remains Even.” PewResearch Center. Accessed on December 20, 2011. Available online at http://people-press.org/http://people-press.org/files/legacy-pdf/450.pdf

Roll Call. 2008. “Sen. McCain and Gov. Palin Deliver Remarks at CampaignEvent, Dayton, Ohio.” Roll Call. Accessed on December 20, 2011. Availableonline at http://www.rollcall.com/news/72097-1.html?zkMobileView=true

Rosenwasser, Shirley M., and Norma Dean. 1989. “Gender Roles andPolitical Office.” Psychology of Women Quarterly 13 (1): 77-85.

Rosenwasser, Shirley M., and Jana Seale. 1988. “Attitudes toward aHypothetical Male or Female Candidate: A Research Note.” PoliticalPsychology 9 (4): 591-599.

Rozell, Mark J. 1991. “Local v. National Press Assessments of Virginia’s 1989Gubernatorial Campaign.” Polity 24 (1): 69-89. Accessed on January 9, 2012.Available online at http://www.jstor.org/pss/3234985

Sapiro, Virginia. 1982. “If U.S. Senator Baker Were a Woman: AnExperimental Study of Candidate.” Political Psychology 9 (1-2): 591-598.

Semetko, Holli A., and Hajo G. Boomgaarden. 2007. “Reporting Germany’s2005 Bundestag Election Campaign: Was Gender an Issue?” InternationalJournal of Press/Politics 12 (4): 154-171. Accessed on January 9, 2012. Availableonline at http://hij.sagepub.com/content/12/4/154.abstract

Smith, Kevin B. 1997. “When All’s Fair: Signs of Parity in Media Coverage ofFemale Candidates.” Political Communication 14 (1): 71-82. Accessed onJanuary 9, 2012. Available online at http://www.tandfonline.com/doi/abs/10.1080/105846097199542

Spangler, Todd. 2008. “McCain, Palin Must Address Economy to Win State,Some Say.” Detroit Free. Accessed on December 20, 2011. Available onlineat http://www.freep.com/article/20080907/NEWS15/809070401/McCain-Palin-must-address-economy-win-state-some-say



http://stateofthemedia.org/2008

http://stateofthemedia.org/2008

http://people-press.org/http://people-press.org/files/legacy-pdf/450.pdf

http://people-press.org/http://people-press.org/files/legacy-pdf/450.pdf

http://www.rollcall.com/news/72097-1.html?zkMobileView=true


http://hij.sagepub.com/content/12/4/154.abstract



Trimble, Linda. 2007. “Gender, Political Leadership and Media Visibility:Globe and Mail Coverage of Conservative Party of Canada LeadershipContests.” Canadian Journal of Political Science 40 (4): 969-993. Accessed onJanuary 9, 2012. Available online at http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=1438752

Ulbig, Stacy. 2009. “Senator Scranton and Governor Hockey Mom: TheImportance of Vice-Presidential Candidates in Presidential Vote ChoiceYesterday and Today.” Paper presented at the Annual Meeting of the MidwestPolitical Science Association, Chicago, Illinois. April 2-5.

van Zoonen, Liesbet. 2006. “The Personal, the Political, and the Popular: AWoman’s Guide to Celebrity Politics.” European Journal of Cultural Studies 9(3): 287-301. Accessed on January 9, 2012. Available online at http://ecs.sagepub.com/content/9/3.author-index

Veenstra, Aaron S. 2007. “Blogger/Reader Interaction: How MotivationsImpact Pathways to Political Interest.” Paper presented at the annual meetingof the Midwest Association for Public Opinion Research, Chicago, Illinois.November 16-17.

Wattenberg, Martin P. 1995. “The Role of Vice-Presidential CandidateRatings in Presidential Voting Behavior.” American Politics Quarterly 23 (4):504-514.




http://ecs.sagepub.com/content/9/3.author-index

http://ecs.sagepub.com/content/9/3.author-index

mixed signals? gender and the media's coverage of the …€™s coverage of the 2008 vice...

Documents