1999 Commonwealth.max 1999 Commonwealth.max Pennsylvania Women's Campaign Communication: A Content Analysis of Brochures from 1996 State Assembly Races1 Stephanie Greco Larson Dickinson College This paper examines campaign communication of women who ran for the Pennsylvania State Assembly in 1996. The statements made in their campaign brochures were content analyzed to see if messages in litis medium were 13imilur to those identified in the literature about female candidates' television advertis­ ing. Specifically the issues, traits, and attacks mentioned in the brochures were examined. Differences between incumbents and challengers, winners and los­ ers, and Democrats, Republicans, and Independents were also examined. As elections have become mor~ candidate-centered, campaign adver­ tising has grown in importance. The messages in advertisements have become highly scrutinized by politicul scientists, journalists, and con­ sultants. Specifically, attention to the similarities and differences between how female and male candidates promote themselves has contributed to addressing the question of why there continues to be a relatively small number of women in elected office. Attention to women's political advertising hus focused primurily on the television advertisements of those running for Congress and Gover­ nor (Debelko & Herrnson, 1997; Johnston & White, 1994; Kahn, 1993; Kalm, 1994; Proctor et aL, 1994). However, the most impressive gains made by women in politics have been at the lower levels of governmenti female membership in state legislatures tripled from 1974 to 1994 (Rule, 1996). Since the lack of women "in the pipeline" is a key explanation for women's underrepresentation (Duerst-Lahti, 1998), it is important to examine advertising in campaigns for an important political entry point the state legislature (Ford & Dolan, 1996). Focusing on state legislative races requires examining forms of politi­ cal advertising appropriate to this level of campaign. Theodore Sheckels (1994, 324) reminds political communication scholars not to be "mes­ menzed by televIsion, especially when dealing with non-national cam­ paigns." However, despite brochures being a common campaign com- o munkanon dpV1cf> in 10('al races (Grey, 1994), they have generally gone ignored. Only Paul Raymond (1987) has systematically examined bro­ chures, and his analysis does not consider the candidates' gender. There­ fore, there is much to be learned about the messages in female candi­ dates' political brochures. 56 1999 Commonwealth.max There are reasons to suspect that women's communication for lower level offices might differ from their advertisements for higher offices. Leonard Williams (1998) warns scholars not to assume that women's advertising will be the same regardless of the office sought. After all, voters' willingness to support female candidates varies by office (Dolan, 1997) and their expectations for candidates differ by government posi­ tion (Huddy & Terkildsen, 1993). In an effort to give the voters what they want, women seeking higher offices might emphasize different issues and qualifications than those running for lower level offices. To discuss differences between the content of state-level brochures and that of tele­ vision advertising for national offices and governorships, we need to re­ view the findings of research on women's campaign communication. Studies repeatedly reveal that women's television campaign adver­ tisements discuss issues more than candidate traits Oohnston & White, 1994; Kahn, 1994; Kahn & Goldenberg, 1991; Kahn & Gordon, 1997; Raymond, 1987). They also discuss "female issues" more than other is­ sues (Kahn, 1993; Kahn & Go:pdon, 1997; Witt et al., 1994), although Kahn (1994) shows gubernatorial candidates as an exception to this. "Female issues" are those issues for wh.i.dl "wol1.1.en are seen as superior" (Kahn, 1994,166) and are operationally defined as education, environment, health care, family issues, and social welfare. Attacks on opponents are infrequent in women's television campaign advertising Oohnston & White, 1994; Kahn & Gordon, 1997) perhaps due to the reluctance to appear "unladylike" (Witt et al., 1994). However, Leonard Williams (1994) found that about one-third of the female senato­ rial candidates whose ads he studied attacked their opponents. When women do explicitly critique their opponents, the focus is more likely to be on issues rather than candidate qualifications or traits (Benze & Declercq, 1985b; Johnston & White, 1994; Kahn, 1993; Kalul. & Gordon, 1997). In terms of the character traits that women candidates promote, the literature is mixed. Early research of women campaigning for the House, Senate, and statewide offices indicates that women conform to feminine traits stereotypically by promoting warmth and compassion more than toughness (Benze & Declercq, 1985a). Williams (1994, 1998) confirmed that female Senate candidates stressed feminine traits (especially empa­ thy) more than other h'aits. However, the extensive work of Kim Fridkin Kahn (1993; 1994; Kahn & Gordon, 1997) clearly points in the other direc­ tion. This work illustrates that women seeking seats in the U.S. Senate and governorships spend more time challenging feminine stereotypes than reinforcing them. They emphasize traits like leadership and experi: ence rather than honesty, compassion, warmth, trustworthiness, and empathy. The level of office could make a difference since "typical femi­ nine traits are considered more suitable for lower or non-elective levels 57 1999 Commonwealth.max of office" (Huddy & Terkildsen, 1993,504). This article examines whether these characteristics found in televi­ sion advertising of women running for high-level offices, such as gover­ nor and Senator, were also found in brochure advertiSing of women seek­ ing state-level congressional seats. It does so by examining brochures used by women running for the Pennsylvania State Assembly in 1996. Pennsylvania is an interesting, but not necessarily representative, place to start exploring this question since it has a reputation for being "inhos­ pitable toward women candidates" (Witt et a1., 1994,5) despite the num­ ber of women in the Pennsylvania State Assembly doubling (to 12%) in the past two decades (Rule, 1996). If the finding~ from the research on women candidates' television advertising hold true for brochures at the State Assembly level, then the brochures should: 1) contain more issue appeals than candidate traits; 2) emphasize certain issues (those identified as "female issues") over nthP! issnE's; 3) infrequently attack opponents; 4) emphasize feminine candidate traits more than masculine traits. Data and Coding Letters asking for copies of campaign brochures were sent to 55 can­ didates running for representative in the General Assembly of Pennsyl­ vania in 1996 who had female names.2 A follow-up letter was sent to those who did not send brochures. Of the 55 candidates contacted, 29 (53%) responded to the letter, with 21 (38%) sending brochures. The other eight indicated that they had not used brochures.3 Of the 21 who sent brochures: eleven won their elections, ten lost them; 12 were Democrats, six were Republicans, and three were Independents; eleven were chal­ lengers and ten were incumbents! Of those who senl brochures, all of Table 1 Relationship between Party and Election Outcome Party Loser Winner Democrat 50% 50% (6) (6) Republican 17% 83% (1) (5) Independent 100% 0% (3) (0) NOTE: Cell sizes are too small for reliable statistical calculations. 58 1999 Commonwealth.max the incumbents won their elections and only one challenger, a Demo­ crat, won hers.s Table 1 illustrates the relationship between party and election outcome. Although brochures varied in length, style, professionalism, and breadth, they were standard introductory brochures, which focused on general appeals introducing the candidate to the general voters, rather than targeting messages for specific audiences, like partisans or the eld­ erly. The brochures seemed designed in part to increase name recognition since even the shortest brochures mentioned the candidates! name more than once. The minimum number of times the candidates! name appeared was three and the maximum was 18 with t1;te average being eight. One way that the brochures varied was in their use of photographs. One candidate did not use any photographs and another used 14. The average number of pictures used on a brochure was 2.3 and the mode was 1 (10 brochures had one picture). The most common use of photog­ Idphy was to include a sin.gle photo of the candidate (typically a formal face shot). Incumbents incfuded photographs of themselves with chil­ dren and senior citizens who were not members of their own families, whereas no challengers did. Six candidates included photographs of their family; but for only one candidate was this her sole photograph. Thirteen of the candidates included some reference to themselves as a wife, a mother, and/ or a daughter. Of these 13, ten emphasized this ref­ erence by either including a photograph of her family, making the refer­ ence in bold type or in color, or putting it at the top of her list of qualifi­ cations. The references to being a wife or mother were often explicitly linked to governing skills. Por example, one candidate claimed in her list of "Reasons to Vote for __ If that "her husband was confined in a wheel­ chair for eight years, she understands the needs of the disabled." More Democrats than Republicans included references to their families. In­ cumbents and challengers were equally likely to do so. The analysis that follows uses a content analysis of the brochures treat­ ing statements, sometimes referred to here as references or claims, as the unit of analysis. Statements were all sentences or sentence fragments that appeared on the brochure. All statements were coded by the author for "Topic:" Issues, candidate traits (characteristics of the candidate, such as honesty or strength), opponent attacks (any criticisms of the opponent whether by name or not),6 group references (such as the candidate's identi­ fication with a party or interest group ), candidate qualifications (credentials of the candidate, such as public service or motherhood), and other (a re­ sidual category for statements of the "Vote for Candidate __ If nature); In addition, all issue references were coded for the specific subject discussed and recoded into "female issues," "male issues," and "other" using the definition of these which is common in the literature.7 All refer- 59 1999 Commonwealth.max ences to candidate traits were coded for the specific lrail menlioneu. TIlese traits were recoded into" feminine traits," "masculine traits," and" other" generally using Williams' (1994) coding scheme. The background char­ acteristics of the candidate whose brochures contained these statements were also coded. These characteristics were: party (Democrat, Republi­ can, or Independent), election outcome (won OI lo::;l), anu cduuiudte ::;ta­ tUB (incumbent or challenger). Findings The 21 brochures yielded 625 statements. The number for individual candidates varied greatly hom six to 68. The average number uf stdte­ ments in a brochure was 12. The overall distribution of appeals is shown in the total column of Table 2. Clearly this medium allows ample room for candidates to offer multiple messages. Similar to television advertising of other female candidates, there were more references to issues than other types of appeals. Issue claims (234) made up 37% of all statements. Qualifications were the second most fre- Table 2 Type of Appeal by Incumbency Status Type Challenger Incumbent Total Issues 29% 42% 234 (85) (142) Qualification 17% 21% 208 (85) (69) Traits 16% 23% 124 (85) (78) Attack 2% 0% 7 (85) Groups 4% 10% 15 (85) (32) Other 1% 1% 7 (85) (4) Total 293 332 625 Chi Square 66.8; Cramer's V = .32; Significant at .001 level Whpn t],,, sm"l1 ('P11", "re eliminated by omitting 1/ Attack" and "Other:" Chi Square 54.3; Cramer's V "" .30; Significant at .001 level 60 1999 Commonwealth.max quent type of reference with 208 (33% of the totnl). There were 124 refer­ ences to character traits (20% of the total) and 45 group references (7% of the total). Also similar to television advertising was the lack of attacks on opponents. These were very rare (only seven, or 1 %). The average num­ ber of appeals per candidate was: 11.1 for issues, 9.9 for qualifications, 5.9 [or traits, 2.1 for groups, and .3 for attacks. Table 2 also demonstrates the mix of statement types used by chal­ lengers and incumbents. These two groups differed significantly in the types of appeals they used. Incumbents were more likely to use issue appeals, character traits, and group appeals than challengers were. Chal­ ltmgel's were more likely to discuss their qualjfications and to attack their opponents than incumbents were. This could be because challengers lack the legitimacy that is inherent in the incumbency status of their oppo­ nents and therefore need to explicitly assert it. Since incumbency and winning strongly co-varied (with one exception, a Democratic challenger / Table 3 Type of Appeal by Party Type Democrat Republican Independent Total I~~u!:!t> 29% 53% 45% 234 (lOS) (96) (30) Qualification 40% 15% 43% 208 (152) (27) (29) Trails 19% 27% 8% 124 (70) (49) (5) Attack 2% 0% 0% 7 (7) Groups 9% 5% 5% 45 (33) (9) (3) Olher 2% 0% 0% 7 (7) Tolal 377 181 67 625 Chi Square 7S.2; Cramer's V = .25; Significant at .001 level 'Wll!:!ll lhe small cells are eliminated by omitting" Attack" and "Other:" Chi Square 5S.9; Cramer's V = .22; Significant at .001 level When the small cells are eliminated by omitting" Attack/' "Other/' and Independent:" Chi Square 50.1; Cramer's V = .30; Significant at .001 level 61 1999 Commonwealth.max who won), the observed differences between incumbents a.1lJ challellg­ ers were also true when winners were compared to losers.8 Republicans were more likely to make references to issues and character traits, while Democrats were more likely to discuss their qualifications, attack oppo­ nents, and make group references (see Table 3). Issues-Table 4 illustrates the distribution uf l~::me put:;itiont; taken in the brochures.9 The most frequent was education with 38 mentions (18% of the total number of issue references). The second most frequent was taxes (29 mentions, 14%), then health care (19 mentions, 9%), and crime (18 mentions, 9%). These specific issues were grouped according to whether or not they fit the definition of "felmtle ,it;t;ues" broadly defined in the literature as issues that have a nurturing aspect to them (including health care, education, welfare, children, environment, etc.) or "male is­ sues" (including jobs, economy, taxes, agriculture, crime, and foreign af- Table 4 I Type of Issue Statements Number "Female" Issues Education 38 Health Care 19 Welfare 16 Environment 8 Children's issues 6 Women's rights 3 Other Ii Total 95 "Male" Issues Taxes 29 Crime 18 Jobs 12 Economics 11 Transportation 8 Farming 4 Other ,2 Total 87 Othel' Government Organization 20 Worker compo / min. wage 8 Other 1 Total 29 62 Percentage 18% 9% 8% 4% 3% 1% 2% 45% 14% 9% 6% 5% 4% 2% 2% 42% 10% 4% 1% 15% 1999 Commonwealth.max fairs). There were slightly more references to "female issues" (95 or 45% ) than "male" (87 or 42%). Although the difference between the two was not large (48% "male" compared to 52% "female" when "other issues" were omitted), it is in the direction identified by the television advertis­ ing literature (Kahn, 1993; Kahn & Gordon, 1997). There were no significant differences in terms of the types of issues used ("women's" or "men's") when challengers were compared to incum­ bents, winners were compared to losers, and Democrats, Republicans, and Independents were compared to each other.'° The only significant differ­ ence appears between Democratic winners and losers. Democratic win­ ners were more likely to discuss "female issues" than "male issues" (see Table 5). This finding might demonstrate the benefit that Democratic women have of playing to an agenda both in keeping with their gender stereotypes and, to some extent, assumptions about their party. Opponent Critiques-The infrequency of opponent critiques (7, only 1 % of all references) conform to the literature, which indicates that women / Table 5 Relationship between Election Income and Type of Issues Discussed among Democrats Type of Issue "Female" "Male" Cramer's V=.22 Significant at .05 level Election Outcome Loser Winner 39% (14) 61% (22) 61% (30) 39% (19) do not frequently attack Gohnston & White, 1994; Kahn & Gordon, 1997). Even when the brochures, rather than claims within them, were used as the unit of analysis, the frequency of attacks (present in ::I out of 21 bro­ chures or one-seventh) falls short of Williams' (1994, 1998) one-third. All of the attacks coded here were made by Democratic challengers who lost. Two of these candidates did not name their opponents. One of these two simply implied that there was something wrong with the in­ cumbent by saying that people should vote for someone who is "Caring ... for a change." The other noted that she was the" only candidate in~the __ district with a deep business background." A third candidate re­ printed a local newspaper article in her brochure, which contained five attacks on her opponent. On the surface, using a newspaper article to 63 1999 Commonwealth.max introduce opponent criticism is in keeping with advice that candidates can avoid backlash by having others launch attacks. A closer look at this brochure indicates that the advice was not taken since the article repro­ duced includes comments made by the candidate criticizing her oppo­ nent (for policy positions and his support for his own pay raise). Over­ alt the minimal number of references to opponents was predictable given the television advertising research, but the nature of the criticisms were not since they went beyond issue critiques. The small number of candi­ dates who made these critiques and the unsuccessfulness of their cam­ paigns should curtail any sweeping interpretation of what this says about office-level differences in negative campaigning. Traits-Character traits were coded into 14 different categories. ll The most common of these categories was "hardworking," which included 27 references (22%). The second most frequent category was "caring" (20, 16%). When incumbents and challengers were compared, differences of five percent or greater existed for four traits. A larger percentage of chal­ lengers' claims were about their strength (11 % vs. 4% for incumbents) and expertise (11 % vs. 6% for incumbents), while incumbents more frequently mentioned their activeness (9% vs. 4% for challengers) and caring (19% VS. 11% for challengers). The gap for mentioning caring was even larger for winners and losers (20% of winners' traits fell into the "caring" cat­ egory compared to 10% for losers). No losers included claims of indepen­ dence whereas 6% of the winners' traits claimed fell in this category. When the traits promoted by candidates of different parties were com­ pared, Independents differed from members of the major parties. Only three of the 14 traits were claimed by Independents: expertise (60%), honesty (20%), and hard work (20%). Independents discussed their ex­ pertise much more than others since these claims made up only 6% of those offered by Democrats and Republicans. Republicans and Demo­ crats differed by more than 5% in the following categories: leadership (with Democrats mentioning this more often, 17% to 8%), activity (with Republicans mentioning this more often, 14% to 3%), caring (comprising 20% for Republicans and only 14% for Democrats), and hard work (with Democrats claiming this quality in 29% of their traits mentioned com­ pared to 12% for Republicans). To some extent these messages seem to be contradicting negative pa.rty stereotypes (such as Republicans being uncaring). Traits were recoded into categories used by Williams (1994) in his study of female Senate candidates' television advertising. These categories in­ cluded: compassion, empathy, integrity, activity, strength, knowledge, and a residual group. Williams argued that compassion and empathy were the feminine traits; strength, activity, and knowledge were the mas­ culine ones; and integrity was neither. Since all of the traits categorized 64 1999 Commonwealth.max Table 6 Type of Trait Appeal Number Percentage "Feminine Traits" Compassion 20 16% Includes: caring Empalhy 6 5% Iudwl~~; l:uup~ri:ttl ve, open to learn, outsider Integrity 12 10% Includes: honesty 38 31% "Masculine Traits" Activity 40 32% / Includes: hard work, effectiveness, excitement Strength 29 23% Includes: independent, strong, leader Knowledge 10 8% Includes: expertness 79 / 63% "Other" Other 7 6% includes: thritty, hopeful, proud here as integrity were about honesty, a characteristic identified by Kahn (1993; 1994) as fitting feminine candidate stereotypes, it will be coded as such here. Table 6 demonstrates the prevalence of each type of trait. Even counting the integrity category as feminine, masculine traits appeared more than twice as often as feminine traits (64% mmparf'd to::l1 %). Thf'Sf' results contrast sharply with Williams' results that showed 62% of the traits claimed by women Senators in their television advertisements were feminine compared to 27% that were masculine. These findings are much more consistent with the findings of Kahn (1994, 108) which were that 81 % of the traits claimed by gubernatorial candidates and 78% of those claimed by Senate candidates were masculine.12 The results fail to con­ firm the expectation that women competing for lower level offices would be more likely to emphaSize feminine traits and instead reinforce the idea that women advertise against type in order to conform to masculine stereotypes of leadership. This conformity appears "acroflfl-the-board," as there were no significant differences in the frequency of claiming "fe­ male" versus "male" traits when candidates from different parties, with different election outcomes, and different status - incumbents or chal­ lengers - were comparedY Conclusions Campaign brochures of women running for the Pennsylvania State 65 1999 Commonwealth.max Assembly in 1996 conveyed messages that in many ways were similar to those found in television campaign advertising for higher offices. They promoted issues more than other appeals, rarely critiqued their oppo­ nents, and focused slightly more on "female issues." I Iowever, IllOre traits claimed were masculine lending support to Kahn's work (1993; 1994) rather than the work of others (Benze & Declercq, 1985a; Williams, 1994). The great range of appeals used in the brochures demonstrates how this form of campaign communication gives candidates an opportunity to take positions on many issues and dern.onslrate both masculine and feminine traits. Candidates do not seem to treat campaigning through brochures as a zero-sum game, thus forcing them to choose between ap­ peals. In this way, brochures are a very different form of communication than television ads, which are much more expensive and brief. The scar­ city of television ad time provides less flexibility for making multiple appeals. In these brochures, rather than choose between appeals (femi­ nine or masculine, this kind of issue or that kind, issues or not), candi­ dates make variety of claims. Nevedheless, for the most part the empha­ sis in women's ads follows that of television advertising: emphasis on issues, few opponent attacks, and more masculine traits than feminine. Notes 1. The author would like to thank Martha Bailey, Kathy Dolan, James Hoefler, Stephanie Slocum-Shaffer, and the reviewers for their sugges­ tions. She would also like to express appreciation to Vickie Kuhn for her invaluable help acquiring the brochures and to the women candidates who provided them. An earlier version of this article was presented at the Southwestern Political Science Association Meeting in March 1998. 2. The names and addresses were obtained from the "Official Results for the 1996 Geneml Dlection" issued by the Perms y 1 v anLa Bureau of Com­ missions, Elections, and Legislation. Only individuals with clearly iden­ tifiable female names were sent letters. Those with androgynous names (such as Pat, Chris, and Tracy) were not contacted, even though women were among this group. 3. It is likely that some of the WOInen who did noL reply uiu nut issue brochures due to a lack of resources or a lack of competition. Therefore, it would be inappropriate to consider the response rate 38%. 4. Of the 55 candidates with female names who were running, 25 won. Thirty were Democrats, 21 were Republicans, and foUl' were Independents; therefore, Republican candidates were underrepreseuLeu in the :;ample. 5. This means that 100% of the references made by "losers" were made by challengers and 92% of the comments made by "winners" were made by incumbents. 66 1999 Commonwealth.max 6. These included references for voters to "make a change." 7. It is important to note that the theoretical justification for this di­ chotomy is underdeveloped despite its frequent usage. In fact, the con­ tinued assumption that traits and issues are gendered might reinforce unfounded stereotypes. This article does not address or rectify this po­ tential problcm since thc focus on comparing State Assembly brochures to the literature on television advertising necessitates use of the same concepts and definitions. 8. The relationship between appeal types and winner / loser is Chi square 53.3; Cramer's V= .29; Significant at .001 level. When "attack" and "other" were omitted, the relationship ~s Chi Square 41; Cramer's V = .26; Significant at .001 level. 9. Of the 234 issue statements, 23 could not be coded because they were too vague. 10. Even when the "neither" category was excluded from the analy­ sis, these differences remained insignificant. When Democrats and Re publicans were compared artd Independents were excluded, there was still no significant difference. 11. These were: independent, strong (or aggressive, fighter, tough, courageous, assertive), leader (or experienced, proven, respected), knowl­ edgeable (or intelligent, expert, know how, informed, thoughtful, well traveled), active (or involved, outspoken, tireless, feisty, energetic), car­ ing (or empathetic, listens, concerned, accessible, works for you, respon­ sive, or can be counted on), outsider / new, honest (or trustworthy, reli­ able, keeps promises, integrity, principled, ethical, or responSible), coop­ erative/coalition builder, learns (or open-minded, nonidcological), ex­ cited (or eager, spirited, enthusiastic), effective / gets results, hardworking (or industrious, committed, studies), and other (hopeful, thrifty, orga­ nized, proud, or ready). 12. We need to recognize that Kahn did not have a neutral category. When references coded as "other" are omitted from the brochures calcu­ lation, 68% of the traits claimed were "masculine." 13. Nor were there significant differences when "other" traits were ex­ cluded from the analysis and when Independent candidates were omitted. 14. Nor were there Significant differences when "other" traits were ex­ cluded from the analysis and when Independent candidates were omitted. Bibliography Benze, James G. & Declercq, Eugene R. 1985a. "Content of television political spot ads for female candidates." Journalism Quarterly. 62:278-83, 28Ft Benze, James G. and Eugene R. Declercq 1985b. "The importance of gender in congressional and statewide elections." Social Science Quarterly. 66:954-63. 67 1999 Commonwealth.max Debelko, Kristen L. and Paul S. Hermson, 1997. "Women's and men's mmpnigns for the U.S. House of Representatives." Political Research Quarterly. 50:121-135. Dolan, Kathleen 1997. "Gender differences in support for women candidates: Is lhere a glass cdling in American politics?" Women & Politics. 17:27-41. Duerst-Lahti, Georgia 1998. "The Bottleneck: Women Becoming Candidates." in Sue Thomas and Clyde Wilcox, Eds., Women and Elective Office: Past, Present, and Future. NY:Oxford University Press. Ford, Lyrme and Kathleen Dolan. 1996. "Contemporary women state legislators: A diverse group with diverse agendas." In Lois Lovelace Duke, Ed. Women in Politics: Outsiders or Insiders? 2nd edition. NY: Prentice Hall. Grey; Lawrence 1994. How to win a local election: A complete step by step guide. NY: M. Evans and Co., Inc. , Huddy; Leonie and Nayda Terkildsen. 1993. "The consequences of gender ste­ reotypes for women candidates at different levels and types of office." Politi­ cal Research Quarterly. 46:503-525. Johnston, Ann and Anne Barton White. 1994. "Communication styles and fe­ male candidates: A study of the political advertising during the 1986 Senate elections." journalism \..2uarterly. 7}:321-329. Kahn, Kim Fridkin. 1994. "The distorted mirror: Press coverage of women candi­ dates for statewide office." Journal of Politics. 56:154-73. Kahn, Kim Fridkin. 1993. "Gender differences in campaign messages: The politi­ cal advertisements of men and women candidates for U.S. Senate." Political Research Quarterly. 46:4Rl-.'i02. Kahn, Kim Fridkin and Edie Goldenberg. 1991. "Women candidates in the news: An examination of gender differences in U.S. Senate campaign coverage." Public Opinion Quarterly. 55:180~199. Kahn, Kim Fridkin and Ann Gordon. 1997. "How women campaign for the U.S. Senate: Substance and strategy."In Pippa Norris, Ed., Women, media, and poli­ tics. Oxford: Oxford University Press. Proctor, David E., William J. Schench-Hanlin, and Karen A. Haase. 1994. "Ex_ ploring the role of gender in the development of negative political advertis­ ing." Women & Politics, 14:1-22. Raymond, Paul B. 1987. "Shaping the news: An analysis of House candidates' campaign communication." In Jan Pons Vermeer,Ed., Campaigns in the news: Mass media and Congressional elections. Westport, CT: Greenwood Press. Rule, Wilma. 1996. "Why are more women state legislators?" In Lois Lovelace Duke, Ed. Women in politics: Outsiders or insiders? 2nd edition. NY: Prentice Hall. Sheekels, Theodore F. 1994. "Mikulski vs. Chavez for the Senate from Maryland in 1986 and the 'rules' for attack politics." Comrnunicutiun Quurterly, 42:311-326. Williams, Lawrence 1998. "Gender, political advertising, and the air wars." In Sue Thomas & Clyde Wilcox, Eds. Women and elective office: Past, present, and ftl.ture. NY: Oxford University Press. Williams, Leonard. 1994. "Political advertising in the year of the woman: Did X mark the spot?" In ElizabethA. Cook, Sue Thomas, & Clyde Wilcox, Eds. The year of the woman: Myths and realities. Boulder, CO: Westview Press. Witt, Linda, Karen M.Paget, and Glerma Matthews. 1994. Running as a woman: Gender and power in American politics. NY: The Free Press. 68