Anda di halaman 1dari 21

Southern Political Science Association

The Distorted Mirror: Press Coverage of Women Candidates for Statewide Office Author(s): Kim Fridkin Kahn Source: The Journal of Politics, Vol. 56, No. 1 (Feb., 1994), pp. 154-173 Published by: Cambridge University Press on behalf of the Southern Political Science Association Stable URL: http://www.jstor.org/stable/2132350 . Accessed: 08/09/2011 03:21
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Cambridge University Press and Southern Political Science Association are collaborating with JSTOR to digitize, preserve and extend access to The Journal of Politics.

http://www.jstor.org

TheDistorted Mirror: PressCoverage Women of Candidates for StatewideOffice

Kim FridkinKahn
ArizonaState University
Voters see the politicallandscapelargelythroughthe eyes of the news media. In racesfor statewide office, where direct contact with politiciansis rare,citizens receive most of their news about the campaign from state newspapers.Voters' dependenceon the press for politicalinformationmay be problematic for women running for office. A content analysis of newspaper coverage in 47 statewide campaignsbetween 1982 and 1988 shows that the press differentiatebetween male and female candidates in their campaigncoverage.These differencesare more dramaticin U.S. Senate races, but the contests. In senatorialraces, women receiveless campaign differencesare still evident in gubernatorial coveragethan their male counterpartsand the coveragethey receive is more negative-emphasizing their unlikely chances of victory. In both senatorialand gubernatorial races, women receive consistently less issue attentionthan their male counterparts.Furthermore,the news media seem more responsive to the messages sent by male candidates.The media's agenda more closely resembles the agendaissued by male candidatesin their televised politicaladvertisements.These systematicdifferences in press treatmentof male and femalecandidatesmay hinder women as they strive for statewide electiveoffice.

Most people do not experiencepolitics directly. Instead, their perceptionsof Yet the politicalworld are shapedlargelyby the news media'srepresentations. becausemany significantevents takeplace dailyand news organizations cannotcover all of these events, newspeoplemust be selective.As a resultof this selectivity,the news media shape, ratherthan mirror,the politicallandscape.The news media's ability to shape political reality is especially potent during electoral campaigns where citizens rely almostexclusivelyon the media for their politicalinformation. In this paper, I examinehow accuratelythe news media representthe candidates they cover by lookingat whetherthe news media treatmale and femalecandidates differently.If the media differentiatebetweenmale and femalecandidatesin their coverageof campaigns,this differencemay influencethe decisionsof voters.
I would like to thankJohn Geer, Rick Herrera,Pat Kenney, and WarrenMiller for their comments on this manuscript.I would also like to thankJulian Cantorand the Political CommercialArchive at the University of Oklahomafor assistanceand Pat Crittendenfor her editorialassistance.An earlier version of this paper was delivered at annual meeting of the AmericanPolitical Science Association, Washington,DC. THEJOURNAL POLITICS, 56, No. 1, February1994,Pp. 154-73 OF Vol. C 1994by the Universityof Texas Press, P.O. Box 7819, Austin, TX 78713-7819

The DistortedMirror

155

The media may cover male and female candidatesdifferentlyfor a number of reasons.First, gender differencesin news coveragemay reflectstereotypesnewspeople hold about male and female candidates.So, for instance, reportersmay comcompassion,while emphasizinga male candidate's stressa femalecandidate's petence because reporters,like other voters, may believe that certain personality of traitsare more characteristic women, while other traits are more characteristic of men (see Gallup 1984; National Women's Political Caucus 1987; Sapiro 1982, for evidenceof sex stereotypingamongvoters). Second, differences in news coverage of male and female candidatesmay be Like other orgaincentivesactive in news organizations. driven by organizational nizations, news organizationsuse standard operating procedures and rules to maketheir reportingmore efficient (e.g., Epstein 1973;Sigal 1973). For instance, of when selecting amongvarious newspeoplerely on "standards newsworthiness" potentialnews items, and one of these standardsis novelty. Since women candidates are relativelyrare, reportersand editors may view them as especiallynewsworthy, and thereforewomen may receive more news attention than their male counterparts. Finally, genderdifferencesin news coverageof male and femalecandidatesmay reflectdifferencesin the campaignstrategiesof men and women. Men and women may conduct their campaignsdifferently,and genderdifferencesin press coverage may simply mirrorthese differences.In this paper, I test whether gender differences in press patterns reflect differencesin the campaignsof men and women candidatesby comparingthe content of news coverage with the content of the own campaigncommunications. candidates' Relyingon theoriesof voting, we can speculateabouthow genderdifferencesin news presentationsinfluence people's voting decisions. First, an importantcriteria for vote choice in electoral campaignsis recognitionof the opposing candidates. Although candidaterecognitionis not an issue at the presidentiallevel, in are it can be consequential subpresidential contestswherecandidates less widely known(see Goldenberg Traugott1984;MannandWolfinger,1980).Since variand ationsin the quantityof news coveragecan influencerecognitionrates(Goldenberg and Traugott 1984; Goldenberg and Traugott 1987), gender differences in the amountof pressattentioncan haveimportantelectoralconsequences.For example, if womenreceiveless coveragethanmen, then votersmaybe less likelyto recognize womencandidatesand votersmay thereforebe less likelyto supportwomencandidatesat the polls. After recognition has been achieved, a candidate still needs to be positively evaluated.Evaluations candidatescan be influencedby these four factors:(1) the of voter's party identification(e.g., Campbell, Converse, Miller, and Stokes 1960; Markus and Converse 1979); (2) evaluations of the candidates' issue stands (e.g., Markus and Converse 1979; Page and Jones 1979); (3) evaluationsof the candidates'personality (e.g., Markus 1982); and (4) assessments of the candidates' viability (e.g., Bartels 1987; Brady and Johnston 1987). Media coverageof

156

Kim FridkinKahn

campaignscan influenceeach of these dimensionsof candidateevaluation.Therefore, if the news media differentiatebetween male and female candidatesin their campaigncoverageof issues, party,traits,and the horserace,these differencesmay influencepeople's impressionsof male and femalecandidates.For instance,if the press focus more extensively on policy matters for male candidates,voters may come to believe that male candidatesare more knowledgeableabout issues than Similarly,if the coverageof the horseraceis more pertheir female counterparts. vasive and more negative for women candidates,then people may develop more negativeimpressionsof womencandidates. in Genderdifferencesin news coverageare unlikelyto be equallyconsequential all electoralcontests. Instead,becausethe influenceof the news media grows with the size of the constituency (Goldenberg and Traugott 1984; Goldenberg and Traugott 1987), gender differencesin news coverageare potentiallymore important in national and statewide elections than in local contests. In particular,in statewideraces for governorand U.S. Senate becauseof the good "mediamarket fit" (Goldenbergand Traugott 1987) and the competitivenessof these contests (Abramowitz1980;Jacobson 1987; Piereson 1977; Sabato 1983; Westlye 1983), media coverageis more prevalentand genderdifferencesin news attentionmay be important. and senatorialraces are likely to generatemore Although both gubernatorial press coveragethan local contests, the two offices differ in a numberof important ways that may influencepatternsof news coverage.First, press coveragemay differ becausethe relevantissue domainsfor the two offices differ. While U.S. senators deal with foreignpolicy and nationalsecurityissues along with other issues of consequence(Sabato 1983;Stein 1990), governorsdeal nationaland international more extensively with statewide concerns such as education and health (Sabato 1983; Seroka 1980; Tidmarch, Hyman, and Sorkin 1984). These two alternative issue domainscorrespondto the stereotypicalstrengthsof male and femalecandidates. Female candidatesare viewed as more competentat dealingwith education, health, and environmentalissues, while men are considered better able to deal with foreign policy and defense issues (e.g., Gallup 1984;National Women's Political Caucus 1987; Sapiro 1982). Given these alternativeissue domainsand corresponding stereotypes, news coverage patterns may be different for these two offices. Similarly,differentialratesof success for women candidatesin U.S. Senate and campaignsmay influencenews coveragepatterns.Women havebeen gubernatorial than winning seats in the about four times as successfulin winning governorships U.S. Senate, and this significantdifferencein success rates for these two offices may influence press coverage of these statewide campaigns. Since women can stress their stereotypicalstrengths in races for governor, and because they are news coverageof female gubernatorial more successfulat winning governorships, than news coverageof femalesenatecandidates. candidatesmay be more favorable

The DistortedMirror

157

Given the differencesbetween these two statewideoffices, I will examine gender differencesin campaigncoverageof both senatorial gubernatorial and elections.
DATA AND METHODS

To investigatewhether the news media differentiatebetween male and female candidatesin their campaigncoverage-and whetherthe magnitudeof these differencesvariesin gubernatorial senatorial and races-I examinecampaigncoverage in 47 statewideraces. If the news media do distinguishbetween male and female candidatesin their coverage, these differences may reflect alternativecampaign strategiesof male and female candidates.To examine this possibility, I compare press coverageof the campaignswith the candidates' own politicaladvertisements. Sampleof Races The data reportedhere are drawnfrom a content analysisof newspapercoverage of 26 U.S. Senate races and 21 gubernatorial races between 1982 and 1988.1 The choice of these years maximizesthe number of women candidatesselected and avoids extendingback so many years that coveragepatternsno longer represent current practice.2To draw the sample of races, the populations of U.S. Senate and gubernatorial races were stratifiedby the type of race and the competitiveness of the race.3Each type of race was divided into one of seven types: (1) Male Incumbent v. Female Challenger,(2) Male Incumbent v. Male Challenger, (3) Female Incumbentv. Male Challenger,(4) Female Incumbentv. Female Challenger,(5) Male v. Male in an Open Race, (6) Male v. Female in an Open Race, and (7) Female v. Female in an Open Race. The competitivenessof the race was measuredby final vote return;competitive races were those where winners won less than 65% of the vote, and noncompetitiveraces were those where the winner'sshare of the vote was at least 65%. The stratification races of by these two variablesyielded 14 possible categoriesof races. For each of these categories,three raceswere randomlyselectedfor analysis,if available.4
' Only generalelection racesbetween two majorpartycandidateshave been included in the sample. The news media may also be influentialin nominationcampaignsbecause voters have little information about the competingcandidates,and they cannotrely on partyas a cue when voting. Yet primary campaignsreceive significantlyless press attention than general election campaigns.For example, Goldenbergand Traugott (1987), in their study of senate elections, find that generalelection campaignsreceivetwice as much coverageas primarycampaigns. 2To increasethe numberof competitiveraceswith femalecandidates,three additionalraceswere included in the sample:the 1982 U.S. Senate races in Missouri and New Jerseyand the 1983 Kentucky racefor governor. 3The senatorial gubernatorial and racesweresampledseparately. 4For some categoriesof races,three racesdid not exist. In these instances,all available raceswere included.For example,thereis only one competitiveracewitha femaleincumbentfacinga malechallenger (Type 3) in the population U.S. Senateraces(1982-1988), so this singleraceis includedin the sample. of

158

Kim FridkinKahn

The newspaperschosen for analysis were whenever possible those with the largestcirculationin the state.5These majornewspaperswere chosen becauseof their potentialimpact on largenumbersof people in the state. Yet the largestpapers, becausethey are more professional,may be less likely than smallerpapersto differentiatebetween male and female candidatesin their coverage,and therefore the present samplemay underrepresent gender differencesin news patterns.For the content analysisof news coverage,every day from September 1 through the day of the election was analyzed.6 Any item in the newspaperthat mentionedeither candidatewas considered, including news articles, columns, editorials,and "newsanalysis" articles. Choice NewsMedium of Newspapers were chosen for analysis for both substantiveand practicalreasons. On the substantive side, there is considerableevidence that newspapers carry more informationabout state-level campaigns than local television news (Goldenbergand Traugott 1987;Westlye 1991), and that people receivemore informationaboutstatewideracesfrom newspapersthan from television(Clarkeand Fredin 1978). Furthermore,Westlye (1991) explains that, comparedwith local broadcastnews, "newspapers present an amountof informationthat more closely what campaignsare issuing"(45). On the practicalside, newspapers approximates are routinely saved on microfilm,making them easily accessible for analysis. In contrast, tapes of local television news are seldom availableafter a campaign, which makesthe examination televisionnews much more difficult. of Development Coding of Scheme Voting theoriesfrompoliticalscience as well as psychologicaltheoriesregarding sex stereotypingguided the developmentof the content analysis codesheet. To known to be significantfor electoralsuccess were anabegin, news characteristics lyzed. As an example,we know that people's evaluationsof a candidate'sviability can influencetheir overallevaluationsof the candidateand, ultimately,their vote decisions(Bartels1987;BradyandJohnston 1987). Becausenewspeoplemay view female candidatesas less viable than male candidates,the discussion of a female candidate'sviabilitymay be more extensive (and more negative).To explore this possible gender difference in news coverage, the content analysis examined the
'The largestcirculatingnewspapersin Texas and Coloradocould not be obtainedso the state papers with the second largest circulation(The HoustonPost and The DenverPost) were analyzed instead. A complete listing of the sample of races and the correspondingnewspapersis availableupon request. 6In those cases where the state primarywas held after September 1, coding began the day after the primary.

The DistortedMirror

159

coverageeach candidatereceivedas well as the press'sassessamountof horserace viability.7 ment of the candidate's Second, givenpastresearchon sex stereotyping,we expectnews coverageof male to and femalecandidates varyin systematicways.For instance,we knowthatpeople believe that male and female candidatesare competent in alternativeissue areas. Male candidates,for example,areconsideredbetterableto dealwith foreignpolicy, while femalesare consideredbetter equippedto deal with issues relatedto education (e.g., Gallup 1984; Sapiro 1982). Since reportersand editors may hold these stereotypesand may reflectthese views in their coverageof candidates,the content of examinedthe substance issuecoverageformaleandfemalecandidates.8 analysis If gender differencesin news coverageexist, they may be driven by reality,sex stereotypes,or by the standardoperatingproceduresemployedby news organizaof tions. Yet regardless the cause,genderdifferencesin news attentioncaninfluence and people'sperceptionsof maleand femalecandidates eventualvote choice.9 of PoliticalAdvertisements Candidates In addition to the data on press coverage,I have also analyzedthe candidates' own campaignadvertisements.'0 Because they are completely controlled by the candidate and the candidate'scampaign, political advertisementsrepresent the candidate'sown "presentationof self" (Kaid and Davidson 1986). By examinwe ing the spot advertisements, can look at what the candidateschoose to emphaIn size in their campaigns."I addition, by comparing the candidates'television
7The coding for the newspapercontent analysiswas done by the authorand two researchassistants (who were unawareof the purposeof the study). Intercoderreliabilitywas assessedat randomintervals during the coding process. Intercoderagreementaveraged92% with agreementnever falling below 85% and reachingas high as 100%for some content categories. 8A copy of the contentanalysiscodesheetis availableupon request. 9Althoughsurvey workhas failed to find a consistentgenderbias in voting (e.g., Darcy, Welch, and Clark1987), this work has focused almost exclusivelyon local racesfor the U.S. House and state legislativeoffices.Yet the genderof the candidatemay play a more powerfulrole in statewideraceswhere are characteristics more influentialand other factors,like inthe news media and candidate-centered cumbency,are less important(e.g., Abramowitz1988;Goldenbergand Traugott 1987). advertisements for 10Ihave selected televisionpoliticaladvertisements analysisinsteadof newspaper becausetelevision ads are consideredsignificantlymore effective in swayingvoters' opinionsand they are used much more frequentlyduring statewideand nationalcampaigns(Goldenbergand Traugott 1984;Jacobson1987;Lutz 1988).Goldenbergand Traugott(1987) report,for example,that more than half of all campaignexpendituresare devoted to the productionand placementof televisionadvertisements in senatecampaigns. "IThe author coded all commercialsin the sample. To ensure the reliabilityof the coding, two checkswere performed.First, I coded a sampleof articlestwice-once at the startof the coding process and once near the end. This reliabilitycheck revealedthat coding remainedstable;reliabilitycoefficients were greaterthan 98% on most measures,never fallingbelow 94%. Second, intercederreliabilwith the objectivesof the study analyzea randomsample ity was checkedby havinga coder unfamiliar of politicalads. Intercoderagreementfor this sampleof ads averaged96%.

160

Kim FridkinKahn

commercialswith the press coverageof the campaign,we can look at the correspondence between the candidate's message and the news media's message." Advertisementswere analyzed for 72% of the candidatesincluded in the newspaperanalysis;the sampleincludes693 advertisements 68 candidates.'3 for
RESULTS

Quantityof Campaign News Media treatmentof male and female candidatesmay differ in a numberof important ways, but one potentially importantdifference is the sheer quantity of news coverage.First, as discussed earlier,standardsof newsworthiness(Epstein 1973; Graber 1989) may lead reportersand editors to consider female candidates as especiallynewsworthy.Female candidatesmay thereforereceivemore press attention than their male counterparts. the other hand, if newspeoplehold cerOn tain stereotypesthat lead them to view women as less viable, women candidates may actuallyreceiveless news coverage. The resultsof the content analysisof news coveragesuggest that women candidates do receiveless press attention,but only in racesfor the U.S. Senate.'4While more than 95 paragraphs about male U.S. Senate candidatesare published each a week, fewer than 79 paragraphs week are writtenaboutthe senatorial campaigns of women candidates.'5 This gender difference in the amount of news coverage occurs in both competitiveand noncompetitiveraces.'6Similarly,the gender differencein news attentionis not a mere reflectionof statusdifferencesfor men and women;femaleSenatecandidatescontinueto receiveless coveragethan their male counterparts,regardlessof their status as incumbents,challengers,or candidates in open races. In contrast,women do not consistentlyreceiveless coveragein gubernatorial racesthan their male counterparts. a Overall, 113 paragraphs week are writtenaboutboth male and femalegubernatorial candidates.
'2In drawing these comparisons,we must remain cautious since we are comparingtwo different mediums, each with its own unique characteristics constraints(Epstein 1973;Sigal 1973). For exand ample, the natureof the television medium places a premiumon simple messageswith interestingvisuals while longer,morecomplicatedmessagescan be presentedmore effectivelyin print. '3The correspondence between the sample of races included in the news analysisand the sampleof racesincludedin the politicalad analysisis not perfectbecause(1) some of the candidatesexaminedin the newspaperstudy did not use televisedspot ads and (2) televisedspot ads could not be obtainedfor all the candidatesin the newspaperstudy. The political ads for this study were provided by the PoliticalCommercial Archiveat the Universityof Oklahoma. '4The candidate-as opposedto the race-is the unit of analysisfor all of the analysesin this article. '5This difference is not statisticallysignificant(p < .10). Although the sample of candidatesexamined here does not representa randomsample from a largerpopulation,statisticaltests have been providedsince the candidatescan be conceptualized a samplerepresenting largertheoreticalpopuas a lationof candidates. '6Fordataanalysis,racesare divided into two levels of competitivenessbased on preelectionratings published in the Congressional Quarterly special election issue. Preelectionratingsare used instead of

The DistortedMirror

161

These resultssuggest that votersmay havea more difficulttime acquiringinforcanmationabout femalesenatorialcandidatesas opposed to femalegubernatorial didates.This lack of availableinformationmay result in an electoraldisadvantage for womencandidates runningfor the U.S. Senate. Since the quantityof news coverage is positively related to voters' recognitionof candidates(Goldenbergand Traugott 1984),and since votersare unlikelyto vote for a candidatewhom they do not recognize(Jacobson1987),votersmay be less willing to vote for womencandidatesin senatecampaigns. News Substance Campaign of News coverageof campaignsnot only affects candidaterecognitionrates, but also determinesthe qualityof informationavailableto potentialvoters. By examining the substanceof news coveragein statewideelections, we can see what people are likely to learn during campaigns.If the substanceof news coveragevaries with the gender of the candidate,then the voting calculusused by citizens is also likelyto varyfor male and femalecandidates. in Horserace Coverage the News. Since assessmentsof a candidate'sissue positions, traits, and viabilityall influencevoting decisions (e.g., Bartels 1987;Brady and Johnston 1987; Markus 1982; Markus and Converse 1979; Page and Jones 1979), news informationabout these evaluationdimensionsare likely to influence vote choice. Turning first to viability concerns, reportersand editors may hold certain stereotypesregardingthe electabilityof women candidates.In particular, been less viablethan their malecounbecausefemalecandidateshave traditionally issues (e.g., who is ahead terparts,the press may focus more intenselyon horserace or behindin the polls, who has the strongercampaignorganization) when covering women for statewideoffice. Althoughwomen are less likely than men to win their senatorialand gubernatorial contests, women Senate candidateshave been far less candidates.'7 Given the poorerperformance successfulthan women gubernatorial of femalesenatorial emphasismaybe candidates,the genderdifferencein horserace contests. moreimpressivein senatorial conteststhanin gubernatorial The results of the content analysis shows that in U.S. Senate campaigns,the press does focus more extensively on the horseracewhen covering female candidates. Twenty-seven percentof all articleswrittenabout femaleU.S. Senate candidatesdiscuss the horserace,while only 21% of the articlesaboutmale candidates
the final vote tally because these preelectionratings are likely to representthe electoralenvironment that reportersare representingin their coverageof the campaigns.The final vote tally, on the other hand, may not correspondto the actualcompetitivenessof the campaign.Preelectionratingsand the betweenthe finalvote and CQprefinalvote tally are highly correlated: there is a 90% correspondence election ratingsfor this sampleof races. '7Between1970 and 1988, only 11%of all femalecandidatesfor U.S. senatorwere elected while alcandidateswere successful. most four times as many (40%) femalegubernatorial

162
TABLE 1

Kim FridkinKahn

PERCENTAGE OF ARTICLES DEVOTED TO THE HORSERACE SenateRaces All Candidates Male Candidates Female Candidates Competitive Male Candidates Female Candidates Noncompetitive Male Candidates Female Candidates Incumbents Male Female Challengers Male Female Candidates Open-Race Male Female 21% 24% (2,148;14) (522;4) 13% 16% (1,684;15) (340;5) 21%** 30% 19%** 28% (1,954;15) (254;2) (1,606;11) (602;6) 16%** 10% 17% 16% (1,682;9) (342;2) (1,853;8) (879;3) 17%** 27% (1,437;14) (531;6) 13% 13% (2,308;14) (134;2) 22%# 27% (4,271;26) (847;6) 17% 15% (2,911;18) (1,427;8) 21%##a 27% (5,708;40) (1,378;12) 15% 15% (5,219;32) (1,561;10) Gubernatorial Races

Note: Horseracecoverageincludes any discussion of a candidate'schances of winning, including of discussionof poll results,and discussionof the candidates' comparisons the candidates' organizations, in performance campaigndebates.Entries are the percentageof articlesdevoted to the horseracewith in the numberof articlescoded for eachcandidate type and the numberof candidates parentheses. aThe p-value is based on the differencein proportionstest: "indicates p < .01, indicatesp < .05.

mention horseraceissues (see table 1). While the magnitudeof this genderdifference is not great, these differences are remarkably consistent and occur for all types of senatorial candidates,women in competitiveand noncompetitiveracesreceive more horseracecoveragethan their male counterparts, do female incumas bents, challengers,and femalesrunningin open races. As expected, gender differencesin horseracecoveragedo not occur in gubernatorialcontests. The print media do not differentiatebetween male and female candidateswhen discussing horseraceissues in gubernatorial campaigns.Fifteen percent of all articlesabout gubernatorial candidates,regardlessof the sex of the candidate,focus on the horserace. Since voters are exposed to more horseraceinformationfor female U.S. Senate candidates,voters may weigh viability concerns more heavily when developing overall evaluationsof these candidates(Iyengarand Kinder 1987). But the

The DistortedMirror
TABLE 2 PRESS ASSESSMENTS OF THE CANDIDATE'S VIABILITY SenateRaces Mean (S.D)a All Candidates Male Candidates FemaleCandidates Incumbents Male Female Challengers Male Female Candidates Open-Race Male Female 3.1 (0.64)* 2.5 (0.88) 3.5 (0.42)* 2.7 (0.67) 2.4 (0.61)* 2.1(0.61) 2.9 (0.45) 3.0 (0.46)
Nb

163

Gubernatorial Races Mean (S.D) 2.9 (0.61)* 2.7 (0.46) 3.6 (0.61) 3.7 (0.46) 1.9 (0.78) 2.1 (0.90) 2.9 (0.76)* 2.4 (0.63) N 695;32 184;10 232;9 25;2 191;8 49;3 272;15 110;5

971;40 289;12 351;15 49;2 252;11 28;6 368;14 112;4

Note: Press viabilityassessmentswere ratedon a four-pointscale where4 denotes "likelywinner,"3 denotes "competitive," denotes "somewhat 2 competitive,"and 1 denotes "sureloser." aEntiesare mean viabilityassessmentsby the press with the standarddeviationin parentheses. bEntriesare the numberof articlesthat mentionthe candidates'viability,followedby the numberof candidates. cThe p-value is basedon the t-statistic: indicates p < .01.

actual content of the horseraceinformationis also consequential.If male and female candidatesreceive the same amount of horseracecoverage (as they do in gubernatorial races), and if the content of this informationis more negative for female candidates,then subsequentevaluationsof these candidatesmay also be 18 negative. If reportersare informed by the historical experience of women candidates, then horseracecoveragemay be more negative for women candidatesand especially for women candidatesfor U.S. Senate.'9The datapresentedin table 2 show that women do receive more negativeviabilityassessmentsthan their male counterparts,but these differences are both substantivelyand statisticallymore impressivein senatorialelections. For example, in senatorialraces women are more
'8The importanceof voters' viabilityassessmentsfor overallevaluationsof candidatesand eventual vote choice has been documentedfor presidentialprimaryvoting (Bartels 1987;Brady and Johnston 1987).Becauseinformation levels are low in statewidegeneralelections,like presidentialprimaries,viabilityassessmentsmay be an importantcomponentof overallcandidateevaluationsin these statewide contests. '9To test this proposition,every press assessmentof a candidate'sviabilitywas ratedon a scale from 1 to 4 with 1 indicating"sureloser,"2 indicating"somewhat competitive,"3 indicating"competitive," and 4 indicating"likelywinner."

164

Kim FridkinKahn

likely to be describedas "somewhatcompetitive"(2.5) in their bids for election (3. while men areusuallyconsidered"competitive" 1).20 Although women candidates in U.S. Senate races generally receive less favorable press ratings, this pattern does not hold for open races. In these nonincumbentraces, male and female U.S. Senate candidatesreceive approximately the same viability ratings, while women candidates in gubernatorialraces are viewed as somewhatless viable than their male counterparts.Since most women candidatesfor governorrun in open races, and since women senatorialcandidates in rarelycompete in these contests, the disadvantage press assessmentsin gubernatorialcampaignsmay be electorallysignificant.2' betweenmen and women in their reportingof the The news mediadifferentiate horserace.Perhapsbecausethey are relyingon the historicalexperienceof women candidates,reportersfocus on viability issues when covering female candidates. This is especiallythe case for female senatorialcandidates,and the characterization of the horseracediscussion is usually more negative for these female candidates. Given these patterns of findings, voters who look to the news media for informationmay come to believe that women candidatesare less electable, and they may be less likelyto vote for these candidates. in PressCoverage Issues.Citizens think about policy considerations, addition of to the candidates'viability, when evaluatingcompeting candidatesfor statewide office (e.g., Abramowitz1981;Hinckley,Hofstetter,and Kessel 1974;Wrightand the Berkman1986).Becauseof its controlover campaigninformation, news media can influencethe significanceof issues during elections (e.g., Iyengarand Kinder 1987;MacKuen 1981). Attention to issues in the news may vary for male and female candidates,with reportersdiscussingpolicy less frequentlyfor femalecandidatesbecauseviabilityconcernsmay takepriority.We expect this to be the case in U.S. Senate races. If attention to policy issues is different for male and female candidates,then the public'sabilityto becomeinformedmay varywith the gender of the candidate.As the data in table 3 show, journalistsdo distinguishbetween male and female candidatesin their reportingof policy.22In both senatorialand races, women receive less issue coveragethan their male countergubernatorial
20Thesegender differencesin viabilityassessmentsin Senate racesdo not simply reflect "real"differencesin the campaignsof male and femalecandidates.Even when one controlsfor the competitiveness of the campaign, women candidates continue to receive less favorableviability ratings. For example,among noncompetitiveSenate candidates,female candidatesare consideredless competitive (2.44) (p < .01). (1.98) than their malecounterparts for governorran 21 For the populationof racesbetween 1984 and 1988,44% of the femalecandidates in open seats while only 18% of the femalecandidatesfor senatorran in these nonincumbentcontests. issue positionsor issue concerns.For example,if 22PoliCy is definedas any referenceto a candidate's an articledescribesSenatorAlan Cranston'sconcern about the drug problem-even if no mention is made of Cranston'sspecificsuggestionsregardingthe problem-the articlewould be counted as menconcernaboutthe drug issue. tioningCranston's

The Distorted Mirror


TABLE

165 3

AVERAGE NUMBER OF PARAGRAPHS PUBLISHED ABOUT ISSUES EACH WEEK SenateRaces Mean (S.D.)a All Candidates Male Candidates FemaleCandidates Incumbents Male Candidates Female Candidates Challengers Male Candidates FemaleCandidates Open-Race Candidates Male Candidates FemaleCandidates 29.3 (17.7) 28.8 (12.9) 14 4 31.9 (18.3) 42.1 (9.4) 13 5 23.6 (14.5) 16.7 (8.1) 11 6 43.7 (18.0)* 14.7 (4.4) 9 3 26.7 (13.4) 23.7 (24.2) 15 2 47.0 (21.9) 42.7 (8.6) 10 2 26.8 (15.1) 21.9 (12.7) 40 12 39.9 (20.0) 34.1 (15.1) 32 10 Nb Gubernatorial Races Mean (S.D.) N

aEntries are the mean numberof paragraphs publishedaboutissues each week with the standarddeviationin parentheses. bEntriesarethe numberof candidatesincludedfor each candidatetype. cThep-value is basedon the t-statistic: indicatesp< .01.

parts,and in some instancesthe differencesare quite significant,both statistically and substantively.23 The tendencyof the press to play down issue concernsfor women candidatesis remarkably consistent.In fact, the issue coverageof femalecandidateslags behind male candidatesin seven of the eight comparisons displayedin table 3*24 There are at least three possible explanationsfor this pattern of findings, and they correspond to the three reasonsofferedat the startof this paper.First, newspeoplemay believe that women are less competentat dealingwith the majorissues of the day, so their views on these issues are consideredless newsworthy.Second, there is the
23Whenwe comparethe percentageof articleswhich mention issues, we find the same gender differencesin coverage.In both senate and gubernatorial races, the percentageof articlesmentioningissues is significantly smallerfor women candidates(p < .01). 24While womenconsistentlyreceiveless issue coveragethan men, this patternis least strikingamong open seats. In open Senate races,virtuallythe same amountof coverageis given to the issue priorities of male and female candidates.In gubernatorial races, women in open races receive somewhatmore issue coveragethan their male counterparts.Given that many women candidatesfor governorrun in open races,this more equitablecoveragemay help women candidatesin gubernatorial races.In Senate races, on the other hand, the equitable issue coveragein open races may be largely inconsequential since womencandidatesrarelycompetein these types of races.

166

Kim FridkinKahn

practicalconsiderationof how much space to allot for the discussion of issues may preferto emphasizeother sorts of campaignnews when covsince journalists ering women. Finally, coverageof issues for women may be less extensivebecause womenmay talkaboutissues less frequentlyin their campaigns. I explicitly test this third explanationby examiningthe candidates'own camWhen we comparethe amount of issue discussion in the paign communications. with the amount of issue coveragein candidate'sown campaigncommunications the news, we find that gender differencesin issue coverageis not a reflectionof the candidates'own campaignmessages.Men do not spend more time describing In their issue positions in their advertisements. fact, women are more likely than men to talk about issues in their campaign advertisements;65% of all advertisementsby female candidatesmention policy concerns,while only 58% of male do.25This greaterconcernwith issues for female cancandidates'advertisements races.26 and didatesoccursin both senatorial gubernatorial Given the scarcityof issue coveragefor women candidates,it is importantto examine the types of issues that do receive media attention. Do reporters,for example, emphasize the same sorts of issues for all candidates, or are there predictablegender differencesin the substanceof issue coverage?Based on researchon sex stereotypes,I categorizeissues as "male"issues or "female"issues (Gallup 1984;NationalWomen's PoliticalCaucus 1987;Rosenwasseret al. 1987; Sapiro 1982). "Male"issues are those issues wheremen are consideredmore competent (e.g., defense, foreignpolicy), while "female"issues are those issues where women are seen as superior(e.g., health, educationpolicy). The discussionof issues will varywith the genderof the candidateif (1) reportersand editorshold the same stereotypesas their readers,or (2) if male and female candidatesemphasize differentissues in their electoralcampaigns. The discussionof issues by the media does varywith the sex of the candidatein senatorialraces. In particular,"female"issues are discussed more extensivelyfor "Female"issues are mentioned women who are running for the U.S. Senate.27 40% of the time for female candidates,but less than one-thirdof the time (30%) for male candidates(p < .01). Overall,however,"female"issues receive relatively little attention(32%) when comparedwith the coverageof "male"issues (68%). races a differentpatternemerges.Overall,"female"issues reIn gubernatorial ceive a greatdeal more news attention(51%), and these issues are not emphasized
significant(p < .10). 25Thisdifferenceis not statistically 26Asnoted above, advertisingdata is not availablefor all the racesin our newspapersample. Given that the newspapersampleand the advertisingsample are not based on the same races,differencesin between these two mediums reflectdifferencesin the samples.To check for this the correspondence possibility, I have reanalyzedthe newspapersample for those races where ads are available.This reanalysis,based on raceswhere news data and advertisingdata are both available,revealsthe same pattern of findings.Male candidatescontinue to receivemore issue coveragein the news when compared to their femalecounterparts. 27Thispatternholds when we controlfor the statusof the candidates.

The DistortedMirror

167

more for female candidates(45%) than for male candidates(53%).28 Differences in the offices of governorand U.S. senatorare probablyresponsiblefor these differencesin issue emphasis. Such "male"issues as foreign policy and defense are candidates,while these issues are more critisimply not relevantfor gubernatorial cal for potentialU.S. senators. Are the gender differencesin policy discussion a reflectionof the candidates' own campaignmessages?The data in table 4, which comparesthe content of powith the content of campaigncoverage,suggest that the corliticaladvertisements respondencebetween the issues presentedin the news and the issues highlighted in the candidate'sadvertisements greaterfor male candidatesin both senatorial is and gubernatorialraces. For example, in U.S. Senate races male candidates' mention "male"issues more than 70% of the time, and news covadvertisements erageof their candidaciesreflectsthis devotionto "male"issues. For femalesenatorial candidates,on the other hand, there is considerableincongruitybetween whatthe candidatesare sayingand whatthe newspapersarereporting.29 between the news media's agendaand the agenda The greatercorrespondence of male candidatesmay reflecta bias of the news media, or it may reflectreal difOn ferencesin the campaignsof men and women candidates.30 the one hand,if reporters hold stereotypes about men and women candidates,they may consider and may listen and reporttheir rhetoricmore male candidatesmore "legitimate" betweenthe press faithfully.On the other hand, differencesin the correspondence and the candidates'messagesmay reflectreal differencesin the campaignsof men and women candidates.Male candidates,for example,may be more effectivecamwhich may accountfor differencesin mepaignersthan their femalecounterparts, dia coverageof their campaigns. Yet regardlessof the cause, gender differencesin campaigncoveragecan have importantelectoralconsequences.Since voters'prioritiesare flexibleand susceptible to mediainfluence(e.g., Iyengarand Kinder 1987;MacKuen 1981),candidates
28Thispatternholds when we controlfor the statusof the candidates. 29Again, when we limit our examinationof news coverageto those raceswhere ads are available,we between news and politicalads for male candidates-in both continue to find greatercorrespondence senate races and gubernatorial races. In the reduced sample of senate races, for example,71% of the male candidates'issue coverageis devoted to "male"issues, while 72% of their ads talk about "male" issues (n.s.). For women, 53% of their news coverageis devoted to "male"issues, while "male"issues are discussedonly 43% of the time in their ads (p < .10). (Unless otherwisenoted, p-valuesare based on the differencesin proportionstest.) 30The greatercorrespondencebetween the male candidates'agendaand the news media's agenda may partlybe a functionof status differencesbetweenmale and femalecandidates.I explicitlytest this by looking within status categories-although the number of cases in some status categoriesis very small. Yet, when comparisonscan be made, the similaritybetween what the candidatesare sayingand what the media are reporting is greater for male candidates.For example, male governorsdiscuss "male"issues 49% of the time in their ads and 51% of their issue coveragein the news is devoted to "male"issues (n.s.). Female governors,on the other hand, talkabout "male"issues 63% of the time in their ads but only 43% of their news coverageis devotedto "male"issues (p < .05).

TABLE

A COMPARISON OF THE CONTENT

OF NEWSPAPERS AND TELEVISED POLITICAL ADVERTISEMENTS Issues in News Coverage SenateRaces Male Candidate Female Candidate 60% 40%1,5 (2,264;12) Gubernatorial Races Male Candidate 47%
53%2,6

Female Candidate 55%


45%2,7

"Male"Issues "Female"Issues

70%
30%1

(9,327;40)a

(8,010;32)

(2,123;10)

Issues in Televised PoliticalAdvertisements SenateRaces Male Candidate "Male"Issues "Female"Issues 72%


28%3

Gubernatorial Races Male Candidate 55%


45%4,6

Female Candidate 43%


57%3,5

Female Candidate 68%


32%47

(269;29)b

(82;10) Traits in News Coverage

(198;21)

(62;18)

Senate Races Male Candidate "Male"Traits "Female"Traits


62%8

Gubernatorial Races Male Candidate 65% 35% (625;32) Female Candidate


65%12

Female Candidate
55%8,11

38% (1,434;40)

45% (388;12)

35% (236;10)

Traits in Televised PoliticalAdvertisements SenateRaces Male Candidate


"Male" Traits 68%9

Gubernatorial Races Male Candidate


64% 10

Female Candidate
78%9,11

Female Candidate
81%10,12

"Female"Traits

32% (98;29)

22% (23;10)

36% (118;21)

19% (26;8)

Note: Cells-sharingsuperscriptsare significantlydifferentfrom each other at the specifiedp-value. The differencein proportionstest was used to calculatestatisticalsignificance."Male"issues include foreign policy, defense spending, arms control, foreign trade, farm isues, and the economy;"female" issues include day care, helping the poor, education,health care, women's rights, drug abuse, and the environment."Male"traitsinclude:independent,objective,strong leader,insensitive,aggressive,amtraitsinclude passive,sensitive,gentle, weakleader,and compassionate. bitious,tough;"female" aEntriesare the number of "male"and "female"issue (trait) paragraphs coded, followed by the numberof candidates. bEntriesare the numberof "male"and "female"issue (trait)mentionscoded, followedby the number of candidates. < 1,2,3,5,10,11,12p .01; 6,7,8p< .05; 4.9p < .10.

The Distorted Mirror

169

who can dominatethe media's agendawill be more successful in influencingthe public's agenda. Candidatesare likely to emphasizetheir strengthsin their campaignappeals,therefore,candidateswho can controlthe media'sagendamay have an easiertime winningelections. Overall,the patternsof issue coverageshown here suggest that the correspondence betweenthe campaignmessagesof the candidatesand the news coverageof the campaignsis greaterfor male candidates.Although women talk about issues more than their male counterparts,reportersactuallycover issues less frequently for femalecandidates.Similarly,the substanceof the issue discussionin the news more clearlyechoes the campaignmessagessent by male candidates. Personality Traitsin the News.Just as assessmentsof a candidate'sviabilityand issue positions influence citizens' vote choices, so do judgments about a candidate's character(Abelson, Kinder, Peters, and Fiske 1982;Markus 1982). Voters may be more likely to evaluatethe personalityof candidatesif trait information is readily accessibleduring campaigns.Yet in news coverageof statewideraces, we find only scant attention given to the personalitytraits of candidates.In the newspapersanalyzed,only aboutfive paragraphs week were publishedaboutthe a while issues receivedsix times as much atcandidates'personalitycharacteristics, tention (31 paragraphs week). a Discussion of character traitsfor U.S. Senate candidatesis sparsefor both male and femalecandidates; fewer than 15%of all articlesmention the candidate'spersonality traits. In races for governor,the attention given to personalitytraits is somewhatmore common for female candidates(21%) than for male candidates (15%).31This greateremphasison personalityfor femalegubernatorial candidates holds for all types of candidates: incumbents,challengers,and candidatesrunning in open races.32 This focus on personal characteristicsfor female gubernatorial candidatesis echoed in the candidates'campaigncommunications.In their politicaladvertisements, women candidatesfor governor talk about their personalitytraits (e.g., their leadership ability, their integrity) somewhat more often than their male Women discuss their personal traits in 55% of their advertisecounterparts.33 ments, while male candidatestalkabouttraitsin 49% of their advertisements. More generally, informationabout a candidate's personal characteris more common in campaignadvertisementsthan in news coverage.The candidatesare more likely than reportersto stress their own personal strengths-or to attack their opponents'personalweaknesses-in their television advertisements. While fewer than 20% of the articlesabout statewidecampaignsmention traits, 39% of
31 significant(p < .05). This differenceis statistically

32Ineach case, the genderdifferencein traitcoverageis statistically significant(p < .10). 33 When we restrict our examinationof news coverageto races where ads have been analyzed,we continueto find that femalecandidatesfor governorreceivemore traitattentionin the news than their malecounterparts.

170

Kim FridkinKahn

the political advertisements examined discuss the candidates'personalattributes


(p < .01).34

on The emphasis traitsin television maybe a reflection the ads of

medium-messages aboutan individual'scharacter may be easierto convey with a visualmediumlike televisionthan with a printmedium. Even though the personalitytraits of competing candidatesare not a major source of news in statewidecampaigns,the substanceof this trait discussionmay vary with the gender of the candidate.If journalistshold stereotypesabout men and women, the coverage of the candidates'personal qualities may reflect this bias. Or the candidatesthemselvesmay stress differenttraitsin their campaignappeals. Male and femalecandidatesmay stress these alternativepersonalitycharacteristics in their campaignmessages if (1) they share the same sex stereotypesas the journalistsand voters, or (2) if they believe it will be effective strategically to stressalternative traitdimensions. To examine the substanceof trait coverageby the press, we can divide traits into two categoriesbasedon the sex stereotypingliterature(e.g., Ashmoreand Del Boca 1979;Ruble and Ruble 1982). "Male"traitsare those traitsthat are seen as of characteristic men (e.g., strong leader, knowledgeable,intelligent) while "female" traits are those traits which are consistentlyassociatedwith women (e.g., we warm,compassionate, honest). Using this categorization, find no differencesin trait coverage for male and female gubernatorialcandidates. In the senatorial cases, however,there is somewhatmore discussionof "male"traitsfor men (62% versus 55%;p < .05).35 In general,reportersdiscuss "male"traitsmore frequentlythan "female" traits, perhapsbecausethey considerthese traitsmore relevantfor statewideoffice. They may think, for instance,that questionsabouta candidate's leadershiparemore important than questions about a candidate'scompassion.36 Yet by emphasizing these "male"traits and by making them salient to the public (e.g., Iyengarand Kinder 1987), reportersmay encouragevoters to develop more favorableimpressions of male candidates. Women may be able to altervoters'stereotypesby emphasizing"male"traitsin their own campaigncommunications.If, for example, women act "unstereotypitheir leadershipabilityand their strength,then votersmay cally"and demonstrate revise their views of "typical"male and female candidates.According to their campaign advertisements,women do act "unstereotypically," stressing "male" traitsfar more frequentlythan "female"traits(see table4). In fact, in both senatorial and gubernatorial races, women are more likely than their male counterparts
34Asbefore, when we limit the examinationof news coverageto those races where ads are available, we find that the discussionof personalitytraits is more common in advertisements than in news coverage. "This differenceholds when we controlfor the statusof the candidate. 36Reporters, who are predominantlymale, may feel more comfortablediscussing "male" traits. When I examinethis possibilityexplicitly, I fail to find a relationshipbetween the sex of the reporter and the contentof the traitdiscussion.

The Distorted Mirror

171

to stress "male"traits. Yet the data in table 4 also show that news coveragemore faithfullyrepresentsthe campaignmessages presented by male candidates.The correspondencebetween the message presented in the advertisementsand the coverageprovidedin the news is clearlyhigherfor male candidates-in both senaAgain, this may be because male candidatesare torial and gubernatorial races.37 or more effectivecampaigners, it may be that reporterspay more attentionto what male candidatesaresaying.38 By stressingtheir stereotypicalweaknessesand talkingalmostexclusivelyabout "male"traits, women may be trying to dispel voters' preconceptionsabout the "typical"female candidate(e.g., women are weak leaders).Yet this strategycan have only limited success because reportersare less responsive to the messages presentedby womencandidates.
CONCLUSIONS

Voterssee the politicallandscapelargelythroughthe eyes of the news media.In racesfor statewideoffice, where direct contactwith politiciansis rare,citizens receive most of their news about the campaignfrom state newspapers.Voters' dependence on the press for political informationmay be problematicfor women running for office. The results of this study suggest that the news media differentiate between male and female candidatesin their coverageof statewide campaigns.These differencesare more dramaticin U.S. Senate races,but differences contests. are still evidentin gubernatorial In U.S. Senate races, women receive less campaigncoveragethan their male and the coveragethey receive is more negative-emphasizing their counterparts, races, women reunlikelychancesof victory.In both senatorialand gubernatorial ceive consistentlyless issue attentionthan their male counterparts.These differences in media treatmentmay hurt women as they strive for statewide elective are office. The lackof press attentionand the scarcityof policy information potential roadblocks women candidatesbecausecitizens are unlikelyto vote for canfor didates with whom they are unfamiliar.Furthermore,the abundanceof negative
371fwe restrictnews coverageto those raceswhere ads are also available,we continueto find greater correspondence between news and politicalads for male candidates-in both Senate racesand gubernatorialraces. In the reducedsampleof gubernatorial races, for example,65% of the male candidates' trait coverageis devoted to "male"traitswhile 64% of their ads talk about "male"traits(n.s.). For female candidatesin gubernatorial contests "male"traits account for 65% of all trait coveragein the news while "male"traitsare emphasized81% of the time in their campaignads (p < .10). 38Asbefore, we can look within status categoriesto see if the male advantagein media coverageis purely a function of status differences.Becausetrait coverageis scant, the comparisonsfor each catecan goryarebasedon verysmallnumbers.Yet when comparisons be made,we find greatersimilaritybetween what male candidatestalk about in their advertisementsand what journalistsreport in the newspaper.For example,amongSenatechallengers,male candidatesmention"male"traits68% of the time in theirads while reportersdiscuss "male"traits67% of the time in their news coverage(n.s.). Femalechallengers the U.S. Senateemphasize"male" for traits85%of time in theirown campaign communicationswhile news reportsof their candidaciesmention"male"traitsonly 56% of the time (p < .05).

172

Kim FridkinKahn

informationpublished about the viability of female senatorialcandidates may color voters' evaluations,making voters reluctantto endorse these candidatesat the polls. Furthermore, resultsof this study suggest that the news mediaare more rethe sponsive to the messages sent by male candidates.Reportersare more likely to emphasizethe personalitytraits and policy areas discussed by male candidates. This gender differencein attentionto the campaignmessagesof statewidecandidates may dampen the electoral prospects for women candidates.Women may adopt different campaign strategies because these alternativestrategies may be more effective for their unique candidacies.Yet by potentiallymuting the campaign messagesof women candidates,differencesin press treatmentmay hamper women in their quest for statewideoffice. 7 submitted February Manuscript 1992 Final manuscript received December 20 1992
REFERENCES Abelson, Robert P., Donald R. Kinder, Mark D. Peters, and Susan T. Fiske. 1982. "Affectiveand SemanticComponentsin PoliticalPerson Perception." and Journalof Personality Social Psychology 42:619-30. Abramowitz,Alan I. 1980. "A Comparison Voting for U.S. Senatorand Representative." of American PoliticalScienceReview74: 633-40. Alan I. 1981. "Choicesand Echoes in the 1978U.S. SenateElections:A ResearchNote." Abramowitz, American Journalof PoliticalScience25:112-18. Abramowitz,Alan I. 1988. "ExplainingSenate Election Outcomes."American PoliticalScienceReview 82:385-403. Ashmore, R., and F. Del Boca. 1979. "Sex Stereotypesand Implicit PersonalityTheory: Toward a Cognitive-Social PsychologicalConceptualization." Roles5:219-48. Sex Bartels, Larry M. 1987. "CandidateChoice and the Dynamics of the PresidentialNominating Process."American Journalof PoliticalScience31:1-30. Brady, Henry E., and RichardJohnston. 1987. "What's the PrimaryMessage: Horseraceor Issue Journalism."In Media and Momentum, Gary R. Orren and Nelson W. Polsby. Chatham,NJ: ed. ChathamHouse. Campbell,Angus, Philip E. Converse,WarrenE. Miller, and Donald E. Stokes. 1960. TheAmerican Voter. New York:Wiley. Clarke, Peter, and Eric Fredin. 1978. "Newspapers,Television, and Political Reasoning."Public Opinion 42:143-60. Quarterly Darcy, R., Susan Welch, and Janet Clark. 1987. Women, Elections, and Representation. White Plains, NY: Longman. Epstein,Edward New York:VintageBooks. Jay. 1973.NewsfromNowhere. GallupReport.1984. 228:2-14. Goldenberg,Edie N., and Michael W. Traugott. 1984. Campaigning Congress. for Washington,DC: Congressional Press. Quarterly Goldenberg,Edie N., and Michael W. Traugott. 1987. "MassMedia Effects in Recognizingand Ratin ing Candidatesin U.S. Senate Elections."In Campaigns the News: MassMediaand Congressional ed. Elections, Jan Vermeer.New York:GreenwoodPress. Graber,D. A. 1989.MassMediaandAmerican Politics.Washington,DC: Congressional Press. Quarterly

The DistortedMirror

173

Hinckley, Barbara,C. RichardHofstetter, and John H. Kessel. 1974. "Information and the Vote: A ElectionStudy."American PoliticsQuarterly 2:131-58. Comparative Iyengar,Shanto,andDonaldR. Kinder.1987.News ThatMatters.Chicago: Universityof ChicagoPress. Boston:Little, Brown. Jacobson,GaryC. 1987. TheJoliticsof Congressional Elections. Kahn, Kim F., and Edie N. Goldenberg.1991. "WomenCandidatesin the News: An Examination of GenderDifferencesin U.S. Senate CampaignCoverage." 55:180-199. PublicOpinion Quarterly Kaid, L. L., and Davidson, D. K. 1986. "Elementsof Videostyle:CandidatePresentationsthrough on Television Advertising."In New Perspectives PoliticalAdvertising, L. L. Kaid, D. Nimmo, ed. and K. R. Sanders.Carbondale, SouthernIllinoisUniversityPress. IL: Lutz, Frank.1988. Candidates, Consultants, Campaigns. and New York:Basil Blackwell. MacKuen,Michael. 1981. "SocialCommunication the Mass Policy Agenda."In More ThanNews, and ed. MichaelMacKuenand Steven Coombs.BeverlyHills: Sage. Mann, Thomas E., and RaymondE. Wolfinger.1980. "Candidates Partiesin Congressional and Elections."American PoliticalScienceReview74: 617-32. Markus,GregoryB. 1982. "PoliticalAttitudes During an Election Year:A Report on the 1980 NES Panel Study."American PoliticalScienceReview76:538-60. Markus,Gregory B., and Philip E. Converse. 1979. "A Dynamic SimultaneousEquationModel of ElectoralChoice."American PoliticalScienceReview73:1055-70. NationalWomen'sPoliticalCaucusSurvey. 1987.Washington,DC: NWPC. Page, Benjamin I., and Calvin C. Jones. 1979. "ReciprocalEffects of Policy Preferences, Party Loyaltiesand the Vote."American PoliticalScienceReview73:1071-90. Piereson, James E. 1977. "Sources of CandidateSuccess in Gubernatorial Elections, 1910-1970." Journalof Politics39:939-58. Rosenwasser,S. M., R. Rogers, S. Fling, K. Silvers-Pickens,J.Butemeyer. 1987. "Attitudestoward Women and Men in Politics:PerceivedMale and Female CandidateCompetenciesand Participant PoliticalPsychology 191-200. 8: PersonalityCharacteristics." Ruble, D. N., and T. L. Ruble. 1982. "Sex Stereotypes."In In the Eye of the Beholder: Contemporary Issuesin Stereotyping, A. Miller. New York:Praeger. ed. to Charlie:TheAmerican Sabato,Larry.1983. Goodbye Good-time Governorship Transformed. WashingPress. ton, DC: Congressional Quarterly Sapiro,Virginia. 1982. "If U.S. SenatorBakerWere a Woman:An ExperimentalStudy of Candidate 61-83. Images."PoliticalPsychology Seroka,Jim. 1980. "Incumbencyand Reelection: Governors v. U.S. Senators." State Government 53:161-65. and and Sigal, Leon V. 1973. Reporters Officials:The Organization Politicsof Newsmaking. Lexington, MA: D. C. Heath. Stein, Robert M. 1990. "Economic Voting for Governor and U.S. Senator: The Electoral Consequencesof Federalism." Journalof Politics52:29- 53. Tidmarch, Charles, Lisa J. Hyman, and Jill E. Sorkin. 1984. "Press Issue Agendas in the 1982 and ElectionCampaigns." Congressional Gubernatorial Journalof Politics26:1226-42. of Westlye, Mark C. 1983. "Competitiveness Senate Seats and Voting Behaviorin Senate Elections." American Journalof PoliticalScience27: 253-83. and Baltimore: Westlye, MarkC. 1991. SenateElections Campaign Intensity. Johns HopkinsUniversity Press. in Wright, Gerald C., Jr. 1974. ElectoralChoicein American: Image,Party, and Incumbency State and in NationalElections. ChapelHill: InstituteforResearch SocialScience,Universityof North Carolina. Wright, Gerald C., Jr., and Michael B. Berkman. 1986. "Candidatesand Policy in U.S. Senate Elections."American PoliticalScienceReview80: 567-90.

Kim FridkinKahn is assistantprofessorof politicalscience, ArizonaState University,Tempe, AZ 85287-2001.

Anda mungkin juga menyukai