Gender as a Factor in the Attribution of Leadership Traits


of 20
Gender as a Factor in the Attribution of Leadership Traits
Deborah Alexander and Kristi Andersen
Political Research Quarterly, Vol. 46, No. 3 (Sep., 1993), pp. 527-545
University of Utah Gender as a Factor in the Attribution of Leadership Traits Author(s): Deborah Alexander and Kristi Andersen Source: Political Research Quarterly, Vol. 46, No. 3 (Sep., 1993), pp. 527-545 Published by: Sage Publications, Inc. on behalf of the University of Utah Stable URL: . Accessed: 08/09/2011 03:11 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . 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 Sage Publications, Inc. and University of Utah are collaborating with JSTOR to digitize, preserve and extend access to Political Research Quarterly. Gender as a Factor in the Attribution of Leadership Traits DEBORAH ALEXANDER,SYRACUSE UNIVERSITY KRISTIANDERSEN, SYRACUSE UNIVERSITY The candidateevaluationliteraturehas emphasized the contributionof both candidate characteristicsand voter characteristics(e.g., party to. But on and identification) candidateappraisals. the literature attribution sex role stereotypessuggests that women candidatesmay be evaluated than theirmale counterparts. This paperpresentsthe resultsof differently a survey of 98 voters in which we explored the relationshipsamong of gender role attitudes,voters'attribution leadershiptraits,and support for male and femalecandidates.The surveyswere conductedin Syracuse, New York,duringthe 1990 campaigns, which includedthreemale-female races. Our results substantiate the hypothesis that when candidate information sparse,genderrole attitudesare consequential the initial is in evaluation of lesser known women candidates. Gender attitudes are important factors in candidate favorabilitywhen the candidates are women challengers.Secondly, we found that voters had a tendency to attributeparticular leadershipqualitiesand issue skills based on sex to if was In candidates, no otherinformation available. addition, hypothetical we found that the more egalitarian voters'gender role attitudes,the the more likely they were to evaluate favorablyactual women candidates. Finally,it was the case that all incumbents,male and female,were rated more positively on both "masculine" and "feminine" traits than were challengers. job. But there is Predictingelectionresults in the UnitedStates is a hazardous one predictionthat any election-eveanalyst could offer with supreme confidence . . . that the newly elected official would be male. (Hershey 1980: 179). Even during the much-touted "Year of the Woman" such a prediction would be a fairly safe one at higher levels of government: the 1992 elections, certainly a success for women, produced a House of Representatives which is only 11 percent female. This is despite the fact that many recent studies have found that voters are generally indifferent to a candidate's sex in making their vote decision (e.g., Carroll 1985; Darcy et al. 1987). The apparent lack of gender bias in the voting booth (at least as measured by aggregate voting statistics) should not lead us to assume that there 527 Research Political Quarterly are not differences in the way voters think about and evaluate male and female candidates and politicians. The fact that gender role stereotypes characterize and influence many decision-making domains suggests that voters might use stereotypes to attributedifferentskills and capabilitiesto men and women candidates. Women have not been a major focus of the extensive literatureon candidate evaluation,but some recent work has raised questions about voters'differentialperceptions of male and female candidates. Research designs examining voter sexual stereotyping have been both experimental (for example, Adams 1975; Gitelson and Gitelson 1981; Hedlund et at. 1979; Huddy and Terkildsen 1991; Mend et al. 1976; Sapiro 1981) and nonexperimental (e.g., Boles and Durio 1980, 1981; Hershey 1977); but almost all of them have used students as subjects and fictitious candidates as the objects of investigation. The research described here, in contrast,examines the attributionof traditional sex-typed leadershiptraitsto real candidates(three female-malepairs) by a small sample of voters exposed to their campaigns. We confirm past findings that hypotheticalmale and female candidates are attributeddifferent skills (based on sex roles and accompanyingskills and traits),and use survey data to ask several questions. First, do voters'perceptions of male and female candidates'skills and issue strengthsin actualcampaignsvary in the expected ways? Second, are voters'general evaluationsof male and female candidates, and the extent to which voters use gender stereotypes related to their gender role beliefs? And finally, does incumbency and/or voter familiaritywith the candidate seem to affect the extent to which voters stereotype candidates according to sex? OR EVALUATION? VOTERSEXISM DIFFERENTIAL Male dominance of political leadership has been challenged by the campign triumphs of women candidates in the last two decades. The 1990 campaign year saw record numbers of women candidates for political office, including 85 running for statewide executive seats, 8 for the U.S. Senate, 70 U.S. House of Representatives candidates,and 2064 women seeking state legislativeoffices For The American Woman And Politics 1991: 1-2). At the close of (Center the 1990 season, 31 women (including one non-voting delegate) had won election to Congress. Three states had women governors and the number of women in state legislaturesis more than four times largerthan it was twenty years ago. The numerical increase in women in public office has made possible research which has attempted to discredit one of the major theses about is that such underrepresentation due to voter women's underrepresentation: sexism. Darcy and his colleagues concluded that "in general elections the 528 Gender Leadership and Traits voter reluctanceto support female candidates, as observed in the 1950s and 1960s, had all but disappeared by the the mid-1970s" (Darcy et al. s1987: 55). However, researchalso exists which supports the argumentthat women and men are still perceived in stereotyped ways (Boles and Durio 1981, 1980; Brovermanet al. 1972; NationalWomen's PoliticalCaucus 1987) and that women in political roles still must deal with stereotypicalexpectations (Diamond 1977; Mandel 1981; Deber 1982; Sapiro 1983; and Sigelman, Sigelman,and Fowler 1987). Can it be the case that voters stereotype female candidates but simultaneously act to produce election outcomes which do not favor male candidates? Both claims could be true if within individuals, aspects of feminine stereotypes which are considered positive in terms of suitability for elective office and those which are considered negative "canceleach other out,"so to speak, so that vote decisions are made which look as if they are sex-blind. Or this "cancellingout"process may work across offices or structuralsituations, so that women are seen as more supportablefor particularoffices, or so that female incumbents are stereotyped positively as compared to female challengers, or so that the issues which characterizeparticularlocal races help produce negative stereotypes of women candidates which in turn limit their support. In the past, voter sexism has been conceptualized as hostility toward women as political candidates and consequent reluctanceto vote for women candidates. The existence of voter sexism has been measured by looking at electionoutcomes,e.g. by pooling many electionresults,controllingforvariables such as incumbency and party,and then testing the null hypothesis that male candidates have no advantageover female candidates. But if the possibilities described above are to be investigated, a closer look at voters' reactions to female and male candidates is critical. Experimentalresearch certainly supports the notion that voters may designate particularoffices as appropriatefor women or define certain political climates as more suitable for a woman's particular political skills. Gender role stereotypes may no longer relegate women to the domestic domain or block their entrance into elective office, but may constrain public expectations about women's areas of expertise and appropriatelevel of public office. In the present research,these questions are approached through the use of survey data collected in a particularpolitical context and measuringattitudes toward real candidates. EVIDENCE FORDIFFERENTIAL EVALUATION In recent years, various studies have assessed the effect of gender in evaluation and decision-making processes. These studies reveal that gender has been associated with differentialratings of elected officials'job performance 529 Research Political Quarterly of identical tasks at identical levels of achievement (Mend et al. 1976). Sex role stereotypes have been shown to affect perceptions of academic competence (Fidell 1970; Simpson 1970); perceptions of emotional maturity and social competence (Brovermanet al. 1972); attributionof success (Deaux and Emswiller 1974); ratings of artistic and authorship performance (Goldberg 1968; Phetersonet al. 1971); attributionof different issue expertise for male and female candidates (Sapiro 1981); and finally, differentialperceptions of candidate strength and power across gender (Gitelson and Gitelson 1981). Boles and Durio (1980, 1981), measuring perceptions about male and female politicians, found distinct differences in gender and political labels. Generally,the "electedwoman"label was evaluated more positively than the "elected man" on traditionallyfemale traits and women were perceived as equal or superior to male politicians in terms of the masculine characteristics of efficiency, stability, and vitality (Boles and Durio 1981: 4-12). This suggests that gender stereotypes may not necessarily slow the progressof women candidates in winning public office; nevertheless,it confirms suspicions that women candidates have to present themsleves as both "male" and "female" to satisfy voters'expectations. Evidence for differentialevaluationalso comes from public opinion polls. Polls confirm that women candidates tend to be seen as more compassionate and honest while men are seen to be better suited emotionally for politics. Women are also attributedan expertise in health care, education, and other "social" "domestic" or issues that male candidates don't have (Toner 1990). A Lou Harrispoll in 1972 revealeddistinctly differentappraisalsof the abilities of men and women in office. The public judged men better at directing the military, managing business and labor issues, strengthening the economy, and dealing with demonstrationsand internationaldiplomacy, while women were thought to be better on issues about children and family, education, the arts, health, poverty, and consumer issues (Sapiro 1983). Fifteen years later, voters in a national survey thought that women running for office were more compassionate, more caring, more honest and would do a betterjob handling social issues and holding down government spending, while male candidates were perceived to be more effective at dealing with militaryand trade issues (NationalWomen's PoliticalCaucus 1987). Women political leaders, candidates, and political consultants believe that their experiences confirm the endurance of voter stereotypes (Kirkpatrick 1974; Lake 1989; Mandel 1981). AND OF ATTRIBUTIONSKILLS TRAITS VOTER In a candidate-centeredage, a good deal of attention has been paid to the factors and processes that voters utilize in evaluating candidates. Research 530 Gender Leadership and Traits shows that perceived candidate qualities have become more salient to voters than political parties (Kagay and Caldeira 1975; Miller et al. 1986); consequently political scientists have begun to assess the impact of candidate images on electability and to pinpoint the characteristics that shape the candidate'simage in the voter'seye (Abelson et al. 1982). The research done by Lodge and his associates (1989) solidly demonstrates the importance of candidate-specificimpressions as part of the candidate evaluationprocesses. have frequently been assumed such as partisanship Perceiver characteristics to be the underpinnings of candidate evaluations.Just as partisanship is an important factorin the development of candidate images, so may one's gender ideology have an effect on the formationof candidate images. That is, a or sex voter's allegianceto "traditional" "egalitarian" role norms may have an on how candidates are perceived. Although gender beliefs important impact probably have no impact on political contests between males, it is likely that gender expectation and norms become salient in races between males and females. Hershey uncovered significant sex differences in college students' willingness to support female political candidates. Young men, particularly those with masculine (as opposed to flexible or feminine) sex role orientations held more negative views of female candidates. Pursuing the relationship between gender role attitudes and attitudes toward women in politics, Hershey (1977, 1980) confirmed that supporters of women candidates are most likely to have egalitariansex-role attitudes. NEW CAMPAIGN 1990: SYRACUSE, YORK In Onondaga County, in Central New York, three races in the fall of 1990 matched male and female candidates: New York State Comptroller (Ned Regan,Republicanincumbent vs. Carol Bellamy,Democraticchallenger);the 48th State Senate District (Nancy LarraineHoffmann, Democraticincumbent vs. Jack Luchsinger,Republicanchallenger);and the 27th CongressionalDistrict (James Walsh, Republican incumbent vs. Peggy Murray, Democratic challenger).These races provided us with an interestingvariety of incumbent men and women, well-known and unknown names and different levels and types of elective office. In the congressionalrace, RepublicanJames Walsh (son of a formercongressman and mayor of Syracuse, and himself a past city district councilor and president of the city council) was running for re-election to his second term as congressionalrepresentativefrom the 27th District, after an impressive victory in 1988. His Democratic competitor, Peggy Murray,was a firsttime candidate for public office and formerpresident of the CentralNew York for chapterof the NationalOrganization Women. LocalDemocratshad invested heavily in the 1986 and 1988 congressionalraces. By the 1990 campaign,the 531 Political Research Quarterly party and its workers were depleted of energy and resources. Murraywas unable to rally enthusiasm for her candidacy, and Walsh easily won a second term by a margin of 63 percent to 35 percent. DemocraticStateSenatorNancy Larraine Hoffmann,three-timewinner in a predominantlyRepublicanand ruralStateSenate district,had enjoyed larger margins of victory with each race. A minority member of the New YorkState Senate, Hoffmannwas targetedfor defeat by the StateSenate Republicanleaders, and her opponent, Jack Luchsinger, attorney and first-time candidate, was heavily financed by state Republicans.With combined expenditures of nearly $600,000 this race turned out to be the most expensive State Senate race ever in CentralNew York.In the end, Hoffmann survived Luchsinger's challenge,defeatinghim with 56 percent of the vote in the OnondagaCounty portion of the State Senate District. In the New YorkStateComptroller's race, DemocratCarolBellamy,former New YorkCity Council president and unsuccessful mayoral candidate, took on a two-term Republicanincumbent, Ned Regan,in a race for an office that few voters knew or cared about. Although overshadowed by more locally based elective offices and more familiarfaces, this race was a criticalstep for women interested in diminishing barriers to executive office. A record low turnoutin Bellamy'svoter base of New YorkCity resulted in the closest statewide race of 1990. But Bellamywas unsuccessful, receiving 47 percent of the vote statewide and only 40 percent in Onondaga County. DATA AND METHODS Sample The researchreportedhere is part of a largerproject which monitoredvoters' responses to campaign informationand examined the ways in which gender stereotypes were affected by such information.The researchreported here is based on survey data from 98 respondents.1 The sample was randomly selected from 1990 voter files purchased from the OnondagaCounty Boardof Elections. We sampled voters who lived in areas of the county where the 48th Senate District overlapped with the 27th Congressional District, and who had voted at least once in the past four years. The overlap area included half of the City of Syracuse,and the northern,eastern, and southern suburbs of Onondaga County. Three hundred names were used to reach 98 respondents. The sampling error is approximately +.08. The refusal rate was 23 1The larger project involved less structured, more intensive interviews with a subsample of respondents and candidates' staff members and content analysis of newspaper coverage of the campaigns. Analysis of these data is not reported here. 532 Gender Leadership and Traits percent; 12 percent of the phone numbers were disconnected; and 32 percent of the 300 were unreachable after five callbacks. Interviewers were trained undergraduates.Respondents were told that we were conducting a study about local political candidates and that the survey was not associated with any of the candidates or parties. The relativelysmall size of the sample makes it imperativeto assess its representativeness. Comparingsample respondents with the voting population, the respondent sample contains a somewhat lower proportionof Republicans(33 percent) than does the Onondaga County portion of the 48th Senate District (41 percent), and a slightly higher proportionof Nonenrolled(29 percent) and Democrats (37 percent) than the population (23 percent and 34 percent). The sample did not vary significantly from the population in terms of sex. We conductedtwo separatesurveysto avoid the possible problemsinvolved in including questions about specific male and female candidates and about gender issues on the same instrument.Conducted August 6-9, the initial survey probed media use, candiate familiarity,gender role beliefs, gender-typed capabilities and issues, and level of office associated with male and female candidates and officeholders, in addition to questions of political ideology, race, income, and education. Partyenrollment,age, and sex were provided by the voter rolls. The second survey, conducted September4-6, asked respondents to rate all candidates on favorabilityand on seven traits, pre-tested to demonstratesex-typed attributions. KEY VARIABLES We constructed a Gender Role Ideology scale from six Likert-typeattitude statements about gender role attitudes. These variables are proxies for the assumed underlying theoreticalconstruct of gender role traditionalism.Four of the items were adapted from the 1988 NORC General Social Survey and two items were adaptedfromBroganand Kutner's (1976) workon the construction of a normativegender role scale. Examples of items are: "Itis more important for a wife to help her husband's careerthan to have a careerherself"; men are bettersuited emotionallyfor politics than are most women." and "Most The reliability coefficient (Cronbach's alpha) for the scale was 0.79.2 For some of the analyses, respondents were classified into three "genderbelief' groups: Traditionals (n=20), Moderates (n=63), and Egalitarians(n=14). 2 A principal component factor analysis of the sex role variables extracted only one factor, supporting the unidimensionality of the scale. Our sample's responses to the items were fairly similar to those of the NORC sample: for example, when presented the statement "A preschool child is likely to suffer if his or her mother works," 54 percent of our sample and 51 percent of the NORC sample disagreed. 533 Research Political Quarterly Six candidates were rated on seven traits: honesty, ability to handle a crisis, emotional stability, compassion, decisiveness, ability to compromise, and competence.3 Three of the traits-honesty, compassion, and ability to "feminine" capabilities.Threeother traits compromise- measuredtraditionally emotional stability-are tradithe ability to handle a crisis, decisiveness, and tionallyassociated with men and leadership.The last trait-competence-was assumed for this research to be gender neutral. Feminine and masculine indices were constructed for each candidate by summing, respectively, their additive feminine masculine traitscores or feminine traitscores. Furthermore, and masculine indices were developed for each sex group of candidates (see Appendix for scale and alpha coefficients).4 ANALYSIS The analyseswe performedon the data were designed to examine respondents' perceptions of male and female candidates in the abstract;to see how well known and favorablyregardedthe actual male and female candidates were; to see whether respondents'gender ideologies affected candidate favorability; and to see whether gender ideology seemed to produce candidate stereotyping. Issues and Traits Gender-Associated We were interested in seeing whether the Syracuse sample, like the experimentalsubjects and survey respondents in the researchdiscussed above, associated particulartraits, capabilities and issue with female candidates or officeholders. Table 1 presents a listing of issues and the percentages of respondents who indicated that either a woman or man candidate would do a betterjob dealing with the issue when in office. Althougha majorityreplied that there were no differences between the sexes' ability to deal with several issues, the net difference between males and females on most of the issues 3 We are gratefulto MarieMorse,ResearchDirectorat the NationalWomen'sPolitical 4 PublicOpinionResearch Hickmanof the Hickman-Brown Caucus(NWPC)and Harrison some of these firm(both of Washington, DC) forallowingus to borrowand paraphrase questions from a model questionnaire preparedfor NWPC'snationwidesurvey, "The New PoliticalWoman," releasedin 1987. Most of the researchabout women candidates and public perceptionof leadership Women'sPolitical Caucus(1984a, 1984b, traitshas been commissioned the National by 1987, 1989; and Williams 1987); see also Bolesand Durio(1980, 1981). The research reportsthat women candidatesare more likely to have an advantageon the traitsof of on honestyand compassion,and males are likely to have an advantage the attributes out and handlinga crisis and emotionalstability."Working compromises" decisiveness or producecontradictory unclearfindings. 534 Gender and LeadershipTraits testify to the continued stereotypingof men and women, e.g., the female candidate would do a betterjob with day care, education, helping the poor and needy, AIDS, health care, environment and civil rights; the male candidate would do a better job with military spending, foreign trade, agriculture,and taxes. There were, however, some surprises and indications of change in expected patterns.Votersalso indicated that they thoughtwomen would do a better job with government spending and the federal deficit; and the malefemale advantageon arms control was only about 10 percent. At least in the abstract,women have a comparable"playingfield"for most domestic issues and, in fact, appear to have a broaderand more diverse issue repertoirethan their male counterparts. Table1 WITH ISSUES ASSOCIATED FEMALE MALECANDIDATES AND ". tellme whether thinkthe manor the woman[candidate] would,mostof the time,do a you betterjob dealing with the issue when in office."Figuresare percent of respondentssaying "man," or woman," "nodifference." Issue Man Woman No Difference Woman's Advantage Day Care Helpingthe poor Healthcare Education Environment AIDS Civil rights Government spending Federaldeficit Drug abuse 1.0% 2.0 3.1 5.1 4.1 3.1 6.1 15.3 16.3 15.3 82.7% 46.9 44.9 44.9 43.9 35.7 36.7 42.9 30.6 23.5 16.3% 50.0 50.0 48.0 51.0 55.1 54.1 39.8 51.0 54.1 Man's Advantage Military spending Foreigntrade Agriculture Arms control Taxes N=98 53.1% 39.8 38.8 39.8 24.5 16.3% 4.1 10.2 29.6 21.4 28.6% 55.1 49.0 24.5 52.0 Table 2 is a similar presentationof candidate traits or capabilities.Again, respondents were asked if they associated the word or phrase more with a hypotheticalmale or female candidate. Even at a cursory glance, it is apparent 535 PoliticalResearchQuarterly Table 2 CAPABILITIES ASSOCIATED WITH MALE AND FEMALECANDIDATES ... tell me whetheryou would, most of the time, associateit [words and phrases]more with the man candidateor woman candidate." or "woman," Figuresare percentof respondentssaying "man," "nodifference." Trait/Capability Woman'sAdvantage Compassionate Struggled to get ahead Handles family responsibilities while serving in office More liberal Speaks out honesty Work out compromises Moral Gets things done Stands up for what they believe Hardworking Man Woman No Difference 1.0% 12.2 7.1 9.2 8.2 8.2 6.1 11.2 7.1 14.3 67.3% 57.1 54.1 45.9 44.9 42.9 35.7 33.7 32.7 26.5 29.6% 30.6 36.7 40.8 45.9 44.9 56.1 50.0 58.2 56.1 Man's Advantage More conservative Tough Handles a crisis Emotionally stable Decisive Better qualified N=98 35.7% 28.6 27.6 25.5 23.5 10.2% 24.5% 15.3 22.4 17.3 19.4 10.2% 35.7% 54.1 45.9 55.1 53.1 67.3% that voters still believe that male and female candidates possess distinct skills and capabilities. By large margins, women are believed to be more compassionate, moral, hardworking,and liberal. Women, more so than their male counterparts,are also thought to have stuggled to get ahead, be able to handle family responsibilities while serving in office, speak out honestly. and stand up for what they believe. Men, on the other hand, are believed to be tougher, more able to handle a crisis, more emotionally stable, more decisive, and more conservative, although the percentage margins are narrower for the traits. "maleadvantaged" capabilitiesthan the marginsfor "femaleadvantaged" Candidate Favorability To begin our assessment of the extent to which stereotypes and gender role beliefs affected voters' perceptions of these six candidates, we asked how 536 Gender Leadership and Traits favorablyrespondents viewed each of the candidates; Table 3 shows these "favorability" responses. In September, CongressmanJames Walsh and State Senator Nancy LarraineHoffmann were well known to our sample. In fact, both incumbents were known by all voters in this sample; in contrast, their opponents, Peggy Murray and Jack Luchsinger, were unknown to roughly one-half to one-third of the voters in September.Although the New YorkState Comptroller,Ned Regan, was a longtime incumbent, his name was not as familiar to the voters as the two other incumbents. The comptroller'schallenger,CarolBellamy,was unknown to almost half (46 percent) of the sample voters. Luchsinger,who began an agressive television ad campaign in early summer, was given the highest favorableand unfavorableratings of the challengers, most likely reflectiveof the very negative tone of his advertisingcampaign. On the other hand, his opponent, StateSenatorHoffmann,enjoyed the highest favorableand the lowest neutral or unfavorablepercentagesof all six candidates. We expected that high neutral and/or unknown scores would indicate a greater use of stereotypes to make judgments about less wellknown candidates. Table3 FAVORABILITY PERCENTAGES CANDIDATE an ... could you tell me on a scale of one to five how favorable impressionyou have of this person?" September Walsh (R/M/I) (D/F/C) Murray Hoffmann(D/F/I) (R/M/C) Luchsinger Regan(R/M/I) Bellamy(D/F/C) Favorable* 29.5% 6.3 53.7 12.7 18.9 10.6 Neutral 44.2% 26.3 29.5 26.3 45.3 25.3 Unfavorable* Unknown 26.3% 17.9 16.9 25.3 21.1 17.9 0.0% 49.5 0.0 35.8 14.7 46.3 N=98 Key: R=Republican D=Democrat M=Male F=Female =lIncumbentC=Challenger *Combines favorable" "favorable" and "very responses **Combines unfavorable" "unfavorable" and "very responses and GenderIdeology Candidate Favorability Is favorabilityassociated with gender beliefs? We posed this question by lookingat the associationbetween genderrole beliefs and candidatefavorability. Table 4 shows the correlationcoefficients for gender role beliefs and each candidate'sfavorabilityscores. Only in the instances of the two women chal537 PoliticalResearchQuarterly lengers (Peggy Murray and Carol Bellamy) are gender beliefs significantly related to candidate favorability.In other words, the more egalitarian the voters' gender beliefs, the more likely they were to rate these two women positively as the campaign began. When we looked at the ratingsthemselves (rather than the correlations),for example, Peggy Murray'smean favorability ratingamong Traditionalsin Septemberwas .74; among Moderates1.26; and 2.57 (all between-groupdifferences significantat p< .05). among Egalitarians Gender role beliefs were not significantly associated with evaluation of the male challenger.This finding suggests that when other candidate information is sparse, gender role beliefs may be consequential in the initial evaluationof less-known women candidates. Table4 RATINGS OF WITH CANDIDATEFAVORABILITY CORRELATION GENDER ROLE BELIEFS Congress James Walsh (R/M/I) (D/F/C) PeggyMurray State Senate (R/M/C) Jack Luchsinger N. Larraine Hoffmann(D/F/I) Comptroller Ned Regan(R/M/I) CarolBellamy(D/F/I) N=94 **p