Spectrum of Gender Self-Presentations among Women Candidates

Spectrum


INTRODUCTION
The 2018 U.S. elections were characterized by many representational "firsts" in terms of the path-breaking diversity among the candidates elected to office. Women won more seats in Congress than ever before, with 25 women serving in the Senate, and 101 women (with a record 43 women of color) serving in the House of Representatives (Center for American Women and Politics, 2023). More openly lesbian, gay, bisexual or trans people were elected than previously, including Colorado's Jared Polis (D; the first openly gay governor to be elected in any state) and Kyrsten Sinema (D-Arizona; the first woman U.S. Senator from her state and first openly bisexual U.S. Senator in the country; Zraick, 2018). Likewise, Ilhan Omar (D-Minnesota) and Rashida Tlaib (D-Michigan) were the first Muslim women elected to Congress (Zraick, 2018).
Women also made inroads into male political space in executive positions, which are often more difficult to obtain because candidates have to appeal to a wider constituency. To be elected in U.S. state-wide elections, candidates need to appeal to rural, urban and suburban voters, who often represent more varied ideological perspectives. Sixty-one women filed as candidates for governor in 2018, which was double the 2014 rate (Center for American Women and Politics, 2018). In the statewide elections for executive seats, nine women were elected as Governors (including incumbents), seven as Attorney Generals, and 11 as state-level Secretaries of State (Michigan made history by electing women for all three; Center for American Women and Politics, 2023). There were eight open, non-incumbent seats for governor with women candidates (Georgia, Idaho, Kansas, Maine, Michigan, New Mexico, South Dakota, and Wyoming; See Table 1), five of whom were elected (Kansas, Maine, Michigan, New Mexico, and South Dakota; See Table 1), and three of whom became the first woman governors of their state (Maine, South Dakota, and Iowa; Center for American Women and Politics, 2018). A record number of women were also elected to state legislatures, increasing the nationwide total by 29 percent from just two years earlier (Center for American Women and Politics, 2018). By 2019, women comprised a historically unprecedented 28.9 percent of all state legislators, with 25.5 percent of these women of color (Center for American Women and Politics, 2018).
The 2018 election year was exceptional for several reasons. In addition to women candidates being elected to office in record numbers, women politicians and voters played a significant role in flipping the U.S. House to a Democratic majority. Women of color challenged electoral biases and were elected in districts with a majority of white voters. Additionally, mothers of young children and young women challenged norms by running for office. Finally, in the words of Dittmar (2019), women "embraced gender as an electoral asset instead of a hurdle to overcome". That is, in contrast to previous years, in which women candidates typically emphasized gender neutral self-presentations, 2018 was the first election cycle to witness broad-scale attention to women's gender as an advantage rather than a liability (Aronson and Fleming, 2023). In fact, this election paved the way for subsequent elections in terms of both the demographics of candidates and their gender self-presentations (Aronson and Fleming, 2023).
Previous research has typically examined gender self-presentations in narrow, gender dichotomous ways. This approach highlights whether women candidates embrace or reject gender stereotypes and often obscures the full complexity of self-presentations. Elsewhere, we developed an innovative theoretical framework for understanding the gender self-presentations of U.S. candidates (Aronson, Oldham and Lucas, 2020). In that work, we suggest that viewing gender self-presentations on a spectrum more accurately reflects nuances in how candidates present themselves to voters. In the present paper, we build on and extend our previous work on this spectrum of gender self-presentations among candidates by considering a broader number of races and their election outcomes. In doing so, we compare election results of those candidates who drew on a wider and more multifaceted range of gender self-presentations to those whose self-presentations were more narrow. In a political context in which gender continues to matter significantly, this application of our theoretical framework contributes to a more nuanced understanding of how politics is gendered.

LITERATURE REVIEW
Women candidates at all levels face a no-win situation: many voters view them negatively if they violate gender stereotypes or if they are portrayed as too nurturing and sensitive (Bauer, 2015;Dittmar, 2015). As a result of the masculinity assumed in politics, women candidates and their strategists carefully weigh whether to draw on, minimize, or challenge gender stereotypes in their campaigns (Dittmar, 2015). In fact, they often aim to strike a "balance" between "feminine and masculine traits and issues" when it comes to both self-presentation and policy stances (Dittmar, 2015: 81).
Previous studies find that women candidates running for executive offices typically adopt a "masculine" campaign style that emphasizes "toughness" and experience in male domains like defense, while minimizing attention to women's issues (Carroll, 2009;Dittmar, 2015). At the same time, in order to conform to norms surrounding femininity, women candidates pay close attention to their tone (i.e., "to be tough but not mean") and appearance (i.e., dress professionally to "neutralize gender"; Dittmar, 2015: 89, 105). They may even have an advantage when they conform to gender norms that emphasize "feminine" traits such as honesty and authenticity, or highlight their "outsider" status (Dittmar, 2015). Some studies suggest that women candidates appeal to women voters when they emphasize stereotypical women's issues, such as education and child care (Herrnson et al., 2003) and that their websites are typically more likely to be "congruent" than "incongruent" with gender stereotypes (Schneider, 2014). "Feminine stereotypes" emphasize women's traits as "warm, nurturing, and sensitive" or "communal" (Bauer, 2015, p.691). In contrast, "masculine" traits are "aggressive," "agentic" (Bauer, 2015) or focus on "leadership" (Schneider, 2014: 265). "Strategic stereotype theory" suggests that candidates draw on gender stereotypes that provide them benefits (Fridkin and Kenney, 2015). For women, this could involve strategically drawing on stereotypes that women are caring (Fridkin and Kenney, 2015).
Yet West and Zimmerman (1987: 125) theorized that "doing gender" can be complex and involves a "routine accomplishment embedded in everyday interaction." As in person, online gender self-presentations are also "performed" and represent an "accomplishment" (West and Zimmerman, 1987). Despite this complexity, prior studies of women in politics emphasize a gender dichotomy, where candidates either draw on or reject gender stereotypes. This dichotomous way of understanding gender in electoral politics fails to capture the full range of possible self-presentations. The 2018 midterm elections were distinct in the ways that women candidates were "doing gender" in their campaigns. Across the U.S. and for a wide range of races, advertisements featured candidates pregnant or breastfeeding, highlighted their experiences with sexual assault or harassment, emphasized their military credentials, or showed images of them protesting Trump's inauguration (Aronson and Fleming, 2023). Dittmar (2019) examines several notable examples of 2018 campaign messages that "draw upon distinctly gendered experiences." Although prior research suggests that women candidates can counteract gender stereotypes, it emphasizes gender dichotomies and does not examine a full range of gender self-presentations or how they are combined. One exception is Lawrence and her collaborators (2016: 203, 197), whose study of three women candidates found a "hybridized" and "diverse strategic deployment of femininity," some of which was traditional and some of which challenged traditional gender roles. Thus, while previous research focuses on stereotypical self-presentations (e.g., Dittmar, 2015;Herrnson et al., 2003;Lawrence et al., 2016;McGregor et al., 2017;Meeks, 2016), examinations of a full range of gender self-presentations in campaigns remains understudied. The present paper pushes this analysis forward by applying our spectrum of gender self-presentations to understand 2018 U.S. women candidates' selfpresentations on social media and websites.
In our previous work, we build on feminist theory to expand our understanding of self-presentations beyond a masculine-feminine dichotomy. In particular, we draw on classic work by Ferree and Hess (1994: 131), who define gender consciousness as "politicized consciousness or identification of women as a group with common interests and a collective identity as women." Scholars have extended Ferree and Hess' (1994) definition to posit that women can be classified on a continuum of consciousness and identification, ranging from traditional on one side to feminist on the other (Aronson, 2017). That is, conservative women's activism affirms gender-traditional roles, yet progressive activists may also draw on gender-traditional frames for strategic purposes or blend traditional and nontraditional views of gender (Aronson, 2017). For example, the Mexican American community activists in Pardo's (1995) study did not identify as feminists, and they both reinforced and altered traditional self-definitions of what it means to be a woman through their activism. In addition, despite empowering women and making visible women's previously invisible health concerns, some activists within the breast cancer movement distance themselves from a feminist label and instead reify gender differences (Blackstone, 2004). As Naples (1991: 490) argued in her study of women community workers in low-income neighborhoods, there are "contradictions between women's performance of traditional female roles and the revolutionary actions they take on behalf of their families and communities." Likewise, while some women embrace a feminist identity, defined as alternative visions for gender relations based on a collective identity, others reject gender inequality but do not identify themselves as feminists (Aronson, 2017). Historically, women of color, working class women, and lesbians were excluded from the second wave women's movement and have argued at times that the movement does not address their concerns (e.g., Collins, 1991;Ferree and Hess, 1994;Reger, 2012). Research also suggests that contemporary young women often exhibit an awareness of inequalities while rejecting a feminist identity (Aronson, 2017).
Seeing women's activism along a continuum provides a model for our work on women political candidates. This approach is more useful than previous dichotomous understandings because it recognizes greater nuance and the potential to blend different approaches. Places on the continuum represent an Ideal Type in that they are theoretically distinct despite their complexity in real life.
Our prior examination found that most candidates drew on multiple self-presentation approaches. Starting with two ends of a continuum, traditional and feminist, our prior work developed a typology that represents multiple points on a gender self-presentation spectrum and applied them to the 2020 U.S. Democratic presidential primary election (Aronson, Oldham and Lucas, 2020). This research found that for both men and women candidates, traditional self-presentations were the least common and gender neutral self-presentations were the most common. Women's nontraditional self-presentations were broader than men's, which tended to stay solely in the domain of fatherhood. While both male and female candidates emphasized feminist issues, the men often did so while simultaneously drawing on traditionally masculine language or highlighting gender stereotypes as male protectors. Overall, we found that women's self-presentations were wider and more multifaceted than men's (Aronson, Oldham and Lucas, 2020).
Although our prior research advances our understanding of the spectrum of gender self-presentations in elections, it only covers one primary race in a progressive party. It also does not consider how self-presentations might be linked to election outcomes. The present paper applies concepts from this spectrum to the 2018 U.S. midterm elections. In doing so, we expand the number, level, and political party of the races considered. We focus on women candidates and also consider the election outcomes in the races examined.
In this paper, we analyze gendered self-presentations in the pivotal 2018 elections. As we will see, all of the candidates examined here strategically blended self-presentations. Our approach aims to capture how "elites" and institutional actors in politics "facilitate a movement's ability to produce cultural change" (Van Dyke and Taylor, 2018). We view candidates' self-presentations as both reflective of, and creating, women's movement frames (Aronson and Fleming, 2023). As such, self-presentations have the potential to create broader change through cultural diffusion (Aronson and Fleming, 2023;Van Dyke and Taylor, 2018). In the analysis that follows, we thus ask the following research questions: how do women running for governor in 2018 in open seat elections, and those running for key executive and legislative offices in the swing state of Michigan, self-present gender in their campaign websites and Twitter bios? To what extent do these self-presentations fit into gender traditional, neutral, nontraditional and/or feminist perspectives? Finally, what were the election outcomes based on the selfpresentations adopted by each candidate?

METHODS
We examine gender self-presentations in the 2018 U.S. midterm elections in two main areas: 1) all 8 of the open U.S. gubernatorial seats that had women candidates in 2018 1 and 2) the state of Michigan as a case study 2 . Nineteen of the 20 races featured men as opponents; for the one race with two women candidates, we analyze both candidates' self-presentations. We examine gubernatorial races because there were a record number of women running for this office in 2018. We focus on Michigan because it was an epicenter of the political divisions facing the country. In the 2016 primary election, Democratic voters nominated Bernie Sanders over Hillary Clinton, yet Donald Trump won the general election in the state. Michigan is also an ideal case study as a result of its geographic, political, and racial and ethnic diversity. It has a major urban center, Detroit, with high income suburbs, as well as major university towns, and many rural communities. Thus, Congressional candidates within the state faced vastly different electoral constituencies. In 2018, voters in Michigan's "pink wave" elected a woman for Governor, Senate, Attorney General, Secretary of State, and five U.S. House of Representatives seats; the new Attorney General is the first openly gay person to hold a statewide elected office. While state-wide and Congressional elections differ in their constituencies, studying these races together allows for a broader range of candidates to be considered, including those of diverse religious, racial and ethnic backgrounds, as well as sexual orientation.
We conducted a qualitative content analysis of owned digital platforms, specifically campaign websites and Twitter bios, for the candidates studied. For the websites, we examined candidate accounts that were live in October 2018. Although other aspects of the websites reveal how the candidates represent particular issues in their campaigns, we focus our analysis on the website bios, which contain the most information about self-presentation. The Twitter bios, which were not part of the original data collection and were added to this study later, were collected in April 2019. As a result of their later inclusion in the study, we examined both personal and professional Twitter accounts.
For the Twitter bios, we used each word included on the candidates' biography as our unit of analysis. For the candidate websites, we used each sentence as our unit of analysis. Each data source was coded by two researchers to assess inter-rater reliability. For the Twitter bio classification, there was 93.75% agreement between the two raters. For the website sentence classification, there was 96.9% agreement between the two raters.
Due to the short nature of the Twitter bios, we created a simple yes-no classification to evaluate whether or not each candidate used labels that fit into each self-presentation on the spectrum (See Tables 1 and 2). For the websites, which contain more text, we were able to complete a more comprehensive analysis. We counted all of the self-presentations on each website, keeping in mind that each sentence could contain multiple selfpresentations on the spectrum (Only 12 out of 445 sentences were classified in multiple ways.). For the website analysis, we examined the percentage of different self-presentations present for each candidate (See Table 2). As we will explain, the categories on the spectrum are distinct and divergent from each other, however most candidates blended multiple self-presentation approaches. Finally, we examined election outcomes.
The analysis of data started inductively with open coding for major themes. That is, these modes of selfpresentation emerged from the candidate websites and Twitter bios. We used qualitative data analysis methods to enter codes, highlight passages of text, and examine co-occurrences of codes. The coding scheme and analytical framework was continually refined throughout the analytic process (Miles and Huberman 1994). In particular, we classified self-presentations according to a full spectrum of possible approaches: traditional, neutral, nontraditional, or feminist (See Table 1).

FINDINGS Traditional Self-Presentations
Women candidates are subjected to judgment based on their appearance much more than men candidates, and the candidates considered here were aware of and sometimes denounced this double standard (Aronson and Fleming, 2023). At the same time, these candidates sometimes emphasized stereotypically feminine gender roles, such as their caretaking and family obligations (See Table 1). They also presented themselves as "listeners," which fits into a gender traditional framework. However, women in politics face a "classic double bind," as stereotypical gender self-presentations can be risky and make a candidate appear "incompetent" (Meeks 2016;Lawrence et al., 2016;McGregor, 2017). The 2018 candidates avoided this problem by blending their traditional self-presentations with feminist and non-traditional ones. Democratic women candidates also put gender-traditional issues at the forefront of their campaigns, including health care, education, and early childhood education (Aronson and Fleming, 2023).

Twitter bios
As a distinctly American electoral winner-take-all process, gubernatorial races obviously take place within established political structures. As such, candidates must have widespread voter appeal. In fact, traditional appeals to voters in ideologically conservative states suggests that the deployment of traditional self-presentations may be strategic. Our analysis of Twitter Bios revealed that forty-five percent (n=9) of the 20 candidates used gender traditional labels to represent themselves in their Twitter bios (See Table 2). Seven of the 8 gubernatorial candidates, who were primarily from conservative or Republican states, used traditional gender descriptors in their Twitter bios. For example, Gretchen Whitmer, of Michigan, claimed the "mom" label in both her personal and official gubernatorial account. Her primary account simply stated "49th Governor of Michigan. Proud mom." This gender-traditional self-presentation, which also appeared on her campaign website, established her maternal qualities and reinforced her ability to act as a caretaker of the state. Similarly, in her personal account, Laura Kelly of Kansas, used "wife," "mother," "champion for kids," and "Governor of Kansas" to describe herself, in that order. In ideologically conservative states such as these, the use of traditional self-presentations may be strategic. The order in which the self-presentations are listed on the Twitter bios also matters. For example, Michelle Lujan Grisham of New Mexico listed her traditional labels before her work labels. Her official governor account calls herself "mother," "grandmother," and "dog owner," followed by her title as the Governor of New Mexico. In her personal account, she listed her status as a "grandmother" first, and then her New Mexican identity, the latter of which has local appeal to voters. In the fourth and final label, she stated that she is the "32nd Governor of New Mexico." Similarly, Mary Throne of Wyoming, also used traditional labels, but unlike Kelly and Whitmer, she listed her work experience first: "Wyoming Public Service Commission, Deputy Chair, lover of all things Wyoming. Mother of 3 wonderful sons, grateful wife of Kevin." These four gubernatorial candidates portrayed the most traditional gender self-presentations. This approach appears to be strategic, as three of the four states voted Republican in the 2016 presidential election. Conservative and Republican voters are more comfortable with traditional gender roles (Khazan, 2017) and such selfpresentations have the potential to win over such voters.

Candidate websites
Overall, 11.9 percent (n=53) of the total number of perspectives on the candidate websites are classified as traditional self-presentations (See Table 2). This number ranged from 0 on the low end (Abrams, Bizon and Epstein) to 30 percent on the high end (Noem; see Table 2). Many of the candidates highlighted their immediate family on their websites, including listing the names of their children and grandchildren. Senator Debbie Stabenow's website stated: "She has two grown children, Todd and Michelle; a daughter-in-law, Sara; a son-in-law, Scott; and five amazing grandchildren." Similarly, Maine's Governor Janet Mills' website stated: "Janet and Stan moved back to Farmington, and she became a full-time mom to five daughters who she helped raise while working full-time herself. She is now the proud grandmother to three grandsons and a granddaughter." Throne of Wyoming said: "With her husband, Kevin Boyce, she started her family in Cheyenne." Kelly of Kansas emphasized her relationship with her husband: "Laura and her husband Ted Daughety, a doctor of pulmonary and sleep disorders, live in Topeka and have been married for 34 years." Several candidates also emphasized caring for other family members. For example, Whitmer stated: "During her first term in the state legislature, Gretchen took care of her mother at the end of her life while giving birth to and caring for her first daughter." Some candidates drew on their family credentials to emphasize their approach to policy. Mills emphasized the connection between her husband's death and her stance on health care: "After a devastating stroke, Stan passed away in 2014. In the last year of his life, Janet learned firsthand what thousands of Maine families go through every year. … This is why Janet is so determined to ensure that every Maine person and every Maine family has access to affordable, high-quality health care." Similarly, Kelly linked her experiences "as a mom" to her commitment to affordable child care. These examples suggest that candidates highlight gender traditional roles as a form of experience that shaped their policy positions.
Candidates also emphasized their skills as "listeners," a traditionally feminine quality. For example, Congresswoman Elissa Slotkin stated: "I will listen to you, make a plan to improve your lives in concrete ways" (emphasis ours). Similarly, Governor Whitmer's website discusses her campaign trail as follows: "That's why she's visited all 83 counties in Michigan to listen to community members about the issues that they're facing" (emphasis ours). This role as "listeners" reinforces gendered expectations.
In sum, traditional self-presentations centered on candidates' roles as mothers, wives, caregivers, and "listeners." While almost half of the candidates use traditional self-presentations for their Twitter bios, only about 1 in 10 of the website sentences have gender traditional elements. This finding may reflect the limited number of words available on the Twitter bios, as candidates must carefully select the words that best represent the images that they want to convey. At the same time, successful candidates did not stay within a traditional framework; they drew on the ideological spectrum of self-presentations to strategically craft a complex and multifaceted image.

Neutral Self-Presentations
In gender neutral self-presentations, candidates downplayed their gender and instead emphasized their accomplishments or policy positions. These accomplishments typically sought to downplay the importance of gender in the campaign in order to have widespread voter appeal. It is especially notable that all of the Twitter bios and nearly three-quarters of the website perspectives include gender neutral language. This approach was typically blended with others in order to create a multi-faceted self-presentation. In fact, prior work reveals that genderneutral self-presentations provided a space for the support of voters, especially men, who do not embrace feminist and non-traditional self-presentations (Aronson and Fleming, 2023).

Twitter bios
In their Twitter bios, all of the candidates included gender neutral or gender blind labels as a part of their selfdescription. Thirty percent of the candidates we analyzed used exclusively gender neutral language, featuring instead their political identities and accomplishments. For example, Dana Nessel's Twitter bios, both personal and official, were completely gender neutral. "Michigan Attorney General" is the only label on her personal account, and her official account is equally straightforward. She used the bio to set a serious and accomplished tone, which aligned with her elected position. This approach stood in sharp contrast to her notable campaign advertisement, where she dramatically emphasized gender by encouraging people to vote for the candidate "without a penis" (Schwartz, 2017).
Similarly, incumbent Congresswoman Debbie Dingell also kept her profile gender neutral. As she put it: "Proudly serving the people of Michigan's 12th Congressional District." Likewise, incumbent Senator Debbie Stabenow's profile simply read, "Representing the people from the Great State of Michigan in the United States Senate." The rest of the candidates included other descriptors alongside gender neutral ones. For example, Gretchen Driskell of Michigan listed a series of work identities, including "Former Congressional Candidate," "Commercial Realtor," "Former State Rep 52nd District," and "Former Mayor of Saline, MI." Although Michigan's 2018 "Pink Wave" received a lot of attention and put gender at the forefront in the races, many candidates appeared to strategically create gender neutral self-presentations in order to have broad appeal in an ideologically diverse swing-state (See Table 2).

Candidate websites
Like the Twitter bios, every candidate used neutral self-presentations on their websites. In fact, 70.6 percent (n=314) of the self-presentations were classified as neutral. In particular, many candidates featured lengthy paragraphs about their policy positions and/or accomplishments during their political careers. Senator Stabenow said: "She authored the bipartisan 2014 Farm Bill, which streamlined over 100 programs, saving more than $80 billion and making historic investments in Michigan agriculture, small towns, and local food systems." Similarly, Kimberly Bizon of Michigan stated her policy positions: "She will push to revitalize and reinvent our community colleges to train a new generation of students for high-tech manufacturing jobs that can be created in the Thumb [of Michigan] and keep our talent right here at home." Another type of neutral self-presentation highlighted educational attainment accomplishments. For example, Slotkin emphasized her Ivy League education: "After college at Cornell University and a few years working for non-profit organizations, I went to graduate school at Columbia University in New York City." Similarly, Congresswoman Haley Stevens' website stated: "She earned a master's degree in social policy and philosophy and a bachelor's degree in political science from American University with honors and distinction." This approach draws voters' attention to the merit of the candidate and their qualifications for political leadership.

Nontraditional Self-Presentations
The 2018 candidates also displayed gender non-traditional self-presentations. These candidates drew on women's movement frames in their demonstration of anger toward inequalities (Aronson and Fleming, 2023). They also assumed stereotypically masculine gender roles, by using aggressive language (sometimes even swearing), and by presenting themselves as "tough" and "fighters." Swearing, and the anger behind it, deviates from gender stereotypes and is typically not accepted for women politicians. This approach also included nontraditional career experiences, such as work in traditionally male occupations. Although research before the 2018 election assumes that women candidates face a no-win situation (Bauer, 2015;Burns, Eberhardt and Merolla, 2013;Dittmar, 2015), 2018 saw new ways for candidates to present themselves (Aronson and Fleming, 2023).

Twitter bios
Half of the Twitter bios included nontraditional self-presentations, especially the word "fighter" (See Table 2).
In an assessment of political candidates, Klein (2019) described that while "men are assumed to be fighters. … Women have to overcome suspicions of weakness, which means they have to be much more explicit about their willingness to fight." Mills, Lawrence and Jordan all labeled themselves as "fighters." For example, Mills, running in Maine, wrote she would "Fight for affordable health care, a stronger economy, and a world-class education for all Maine kids" (emphasis ours). Similarly, Congresswoman Brenda Lawrence had a neutral Twitter bio except for the word "fight." Cathy Albro, also running for Congress, had a gender neutral bio, except for the word "fighter." Although one might claim that the term "fight" is ubiquitous in politics and therefore not gendered, we argue that it should be viewed as gender nontraditional. "Fighting" goes against gender stereotypes that view women politicians as cooperative and getting along well with others.
Many of the candidates blended traditional and nontraditional labels in their Twitter bios. Paulette Jordan, from Idaho, first identified as "Country strong," a non-traditional self-presentation, followed by the traditional "proud mom." Jordan also included the non-traditional career self-presentation of "Business Founder/Exec". She closed with "fighting for Idaho always" (emphasis ours).
Similarly, Republican Kristi Noem, who ran for Governor in South Dakota, used the traditional feminine descriptors "wife" and "mom" as well as her state identity in her gubernatorial account. In her personal account, she included "wife," and "mother," but also described herself as a "farmer," "rancher," and "small business owner," traditionally masculine occupations. While many of the other gubernatorial candidates in Republican states focused on gender traditional labels, Noem's nontraditional ones stand out. Likewise, Jocelyn Benson, Michigan's new Secretary of State, used a blend of labels. She listed her primary identity as "mom." Her second label was "Michigan Secretary of State," and her third was "marathoner," a nontraditional self-presentation. Benson was a dedicated runner, averaging two marathons per year. She made headlines after completing the Boston Marathon while eight months pregnant (The Detroit News, 2016).
Similarly, Stacey Abrams, who ran for Governor in Georgia, presented gender in a nontraditional way in her Twitter bio. In her post-campaign account, she named herself first as a founder of two voter equality organizations, then as an author, and then as the founder of two apps, and closed with her ghostwriting alias. Her bio is distinct in that it excludes any specific mention of gender, but still painted a strong contrast between her identity as a founder and entrepreneur and her identity as a writer of romantic suspense novels. Slotkin, who unseated a Republican in her run for Congress in Michigan, presented herself solely in gender neutral and gender nontraditional ways: "Proud Michigander. Representative #MI08. Former Acting Assistant Secretary of Defense, @CIA." Michigan Congressional District 11 was unique in that it featured two women on the ballot in the general election: Stevens (D) and Lena Epstein (R). In their Twitter bios, both drew on multiple self-presentations, including gender nontraditional. Stevens' bio began with her new title, "Congresswoman for MI-11. Metro-Detroit," and then stated "Millennial. Woman in manufacturing. Passionate advocate. Here to work for you!" She was the only candidate to use a generational label and used her work identity to portray a nontraditional gender self-presentation that is connected with a strong state identity. Manufacturing (particularly the auto industry) has a long history in Michigan, and is historically male dominated. Stevens' opponent, Epstein also blended traditional and nontraditional self-presentations: "Business Leader," "Conservative," and "Wife & Mother." The latter traditional self-presentations are often expected from a Republican candidate in a conservative district. In sum, half of the candidates utilized gender nontraditional self-presentations in their Twitter bios, blending these them with other approaches on the spectrum.

Candidate websites
The 2018 election in general, and gubernatorial races in particular, were distinct because women candidates did not shy away from traditionally masculine language, images, emotions, or assertiveness. This type of anger breaks out of gender stereotypes for women. In doing so, they often captured, and propelled forward, the anger of voters, especially women and people of color (Aronson and Fleming, 2023).
Almost 11 percent (n=48) of the website self-presentations could be classified as nontraditional, ranging from 0 to 37.5%, depending on the candidate. Yet eight candidates did not utilize gender nontraditional selfpresentations at all (See Table 2). For example, Whitmer stated on her website: "She'll fight for the things that matter to people and put them first" (emphasis ours). Similarly, Albro stated that she: "has made it her life's work to fight for the health and education of our nation's children" (emphasis ours). Likewise, Driskell said: "I'll fight to stop foreign trade deals like TPP that outsource our jobs, fix bad deals like NAFTA that have forced our kids to leave Michigan to find work, and fight to protect Medicare and Social Security" (emphasis ours). Jordan linked "fighting" with her Native American heritage: "proud member of the Coeur d'Alene tribe, I was raised to fight for the needs of my community and to protect our priceless natural resources" (emphasis ours). Finally, Tlaib, known nationally for saying "we're going to impeach the motherfucker" at her swearing in, used "fighter" language: "Rashida has an unparalleled record fighting for her constituents and her values, taking on billionaires and multinational corporations, and winning" (emphasis ours). Although not featured directly on her website, Whitmer made "fix the damn roads" a central campaign slogan (Aronson and Fleming, 2023).
The candidates used other nontraditional language on their websites, such as "roll up her sleeves." Albro's website says: "Cathy is ready to roll up her sleeves and get to work building the social movement towards civility, respect, and a balance of power in our beloved country" (emphasis ours). Tlaib put it similarly, "she will roll up her sleeves to make sure her residents are cared for, no matter how big the challenge" (emphasis ours). While this image is not inherently masculine, it does conjure up a picture of a man rolling up his button down dress shirt sleeves in order to get to work immediately on important issues.
Gender nontraditional self-presentations included the "tough" self-presentation of Slotkin, who campaigned as a former CIA analyst and defeated a Republican male incumbent in a close race. Slotkin had the highest use of nontraditional self-presentations on her website, at nearly 40 percent (see Table 2). Slotkin: "I would join the intelligence community and work to prevent future terrorist attacks against the United States. I was recruited to join the Central Intelligence Agency as a Middle East analyst, and within a year of joining the agency, I was deployed to Baghdad to serve alongside America's soldiers and diplomats, doing my part to help in a very complicated war. I served a total of three tours in Iraq over a span of five years." This deviation from stereotypical gender selfpresentations was blended with other approaches to highlight a multifaceted appeal to voters. In a context in which women candidates pushed back against the status quo, research suggests that many voters, especially women, supported these nontraditional roles and candidates' anger (Aronson and Fleming, 2023).

Feminist Self-Presentations
Embedded within feminist self-presentations was an emphasis on women's collective power, the recognition of the intersection between electoral politics and the #MeToo movement, and an awareness of intersectional feminism. Reflecting larger social and cultural changes that were underway, many candidates also used explicitly feminist self-presentations (Aronson and Fleming, 2023). Feminist self-presentations are classified as such because they emphasize women's social and political interests as women and/or include an awareness of gender inequalities and patriarchy (Aronson, 2017). This approach was undertaken least often in our study: only 1 candidate featured a feminist self-presentation in her Twitter bio and 6.7 percent (n=30) of the website self-presentations were classified as feminist. However, the presence of these self-presentations, while lower than the others, is notable.

Twitter bios
One type of feminist self-presentation is an emphasis on women's issues. Tlaib was the only candidate who explicitly used feminist language in her Twitter bio. In her personal account, Tlaib referenced the campaign slogan and autobiography title of Shirley Chisholm, the first black woman elected to Congress, and the first woman to run for the Democratic presidential nomination, when she called herself "Unbossed Congresswoman." Tlaib's second label is "Mama working for justice," suggesting an activist approach that is linked with motherhood. Suggesting a feminist intersectional perspective, she mentioned her ethnic identity ("Palestinian American") and religious identity ("Proud Muslima"). This explicitly feminist self-presentation fit within Tlaib's district, which historically favors Democrats by a very wide margin. In this context, she presented feminist ideology and activism and an intersectional identity in ways that many candidates likely were not able to do as a result of efforts to appeal to a more conservative voter base.

Candidate websites
Nearly half of the candidates (9 of 20) explicitly utilized feminist self-presentations on their websites, which covered 6.7 percent (n=30) of the website self-presentations. Many candidates emphasized women's issues, including the #MeToo movement. Whitmer made public, and drew on, her own experiences with sexual assault as part of the reason for her policy position: "On the floor of the state legislature, I shared my own story of surviving sexual assault to speak out for all the women they silenced by refusing to hold a single public hearing. As Ingham County Prosecutor, I implemented a new domestic violence unit, and am committed to bringing together university officials, law enforcement, students, and legislators to end sexual assaults on college campuses." Although we are not systematically analyzing images, Whitmer's notable website featured a photo of her at the Michigan capitol protesting Trump's inauguration in January 2017. Likewise, Dingell's website stated that "she is a recognized national advocate for women and children. … She successfully fought to have women included in federally-funded health research, and advocated for greater awareness of issues directly related to women's health, including breast cancer and women's heart health." Like Dingell, Abrams also combined the nontraditional "fighter" language with women's issues, particularly abortion and birth control: "Stacey Abrams is a true fighter for women and families. … As governor, Stacey will look out for women and defend our right to access basic healthcare, including contraception and abortion." These feminist self-presentations thus typically include policy statements and personal history.
Some self-presentations centered on an intersectional approach. For example, as the first openly gay Attorney General elected in the country, Dana Nessel illustrated both her personal interests and policy stance on LGBTQ rights. Nessel stated that she "lives in southeast Michigan with her wife, Alanna Maguire, their twin sons, Alex and Zach, along with various cats." Nessel's website also stated: "Dana is also recognized as one of the premier litigators of LGBTQ issues in Michigan." Her website mentioned her role in same-sex couples' custodial rights, second parent adoption, and employment discrimination.
Many of the candidates illustrated feminist self-presentations in their emphasis that they were the "first" woman or woman of color to hold their positions. For example, Lawrence stated that she "became the first female and first African-American Mayor of Southfield … and was the first African-American female major party nominee for that position in Michigan's history" (emphasis ours). Tlaib stated: "Rashida made history in 2008 by winning her race for State Representative and becoming the first Muslim woman to ever serve in the Michigan Legislature" (emphasis ours). Likewise, Abrams said: "In 2010, Stacey became the first woman to lead either party in the Georgia General Assembly and the first African American to lead in the House of Representatives" (emphasis ours). Similarly, Mills' website stated: "For the past five years, Janet has served as the Attorney General of Maine-the first and only woman to hold the job" (emphasis ours). Thus, emphasizing the collective rights of women in their feminist selfpresentations, candidates made links to the women's movement, feminist and intersectional issues, the #MeToo movement, and the historic nature of their candidacy.

Gender Self-Presentations and Election Outcomes
Considering self-presentations on websites in open and challenger races, we compared those seats that flipped from Republican to Democrat with those that stayed Republican in the 2018 election (See Table 3). Considering the gubernatorial races, we find that the four races that flipped the seats to Democrat (Grisham, Kelly, Mills and Whitmer) had a wider range of gender self-presentations than the four seats that remained Republican (Abrams, Jordan, Noem and Throne). Republican-staying seats were more likely to be gender neutral, at 77.4 percent (See Table 3), in contrast to the 59.8 percent of website self-presentations of the flipped seats. Conversely, flipped seats have more traditional (17.5 vs. 12.9 percent), nontraditional (15.5 vs. 8.1 percent), and feminist (7.2 vs. 1.6 percent) self-presentations. We see a similar pattern with the Michigan races. The five Democratic candidates (Whitmer, Bensen, Nessel, Slotkin, and Stevens) who flipped their seats utilized a wider range of self-presentations than those who did not flip seats (Albro, Bizon and Driskell; See Table 3). In a context of bitter ideological divisions, these results are suggestive. In more conservative, Republican areas, women candidates appear to embrace gender neutral self-presentations, while those in progressive or mixed areas exhibit a wider range.

CONCLUSIONS
In this paper, we move the scholarship on gender and politics forward by examining women's gender selfpresentations in Twitter bios and websites in a key election year in U.S. politics. We extend our previous research on the spectrum of self-presentations by focusing on a broader range of races in the 2018 midterm U.S. elections. This area of research will be increasingly important in future elections as women continue to expand their role in politics and push the boundaries of typical self-presentations.
Races obviously take place within established political structures. As such, candidates must have widespread voter appeal. Successful candidates during the 2018 contests drew on a spectrum of self-presentations to strategically craft a complex and multifaceted image. All of the candidates examined here strategically blended together one or more types self-presentations. Gender traditional self-presentations included their status as caretakers and their roles as "listeners", which may attract conservative voters. Gender neutral self-presentations, the most common approach, de-emphasized gender and focused instead on accomplishments or policy positions. As a strategy, this approach may have the widest voter appeal, especially in conservative or ideologically mixed areas. Nontraditional self-presentations drew on stereotypically masculine language, traits, or occupations in new ways in 2018. Finally, feminist self-presentations emphasized a connection to women's movement issues and concerns, the #MeToo movement, and feminist intersectional approaches. The range of candidates' gender selfpresentations were unprecedented and reflect a unique historical moment. In particular, feminist and nontraditional images and language contributed to a sense of women's power and agency. Women candidates ran for office in the midst of a "gender revolution" in which new self-presentations, most notably gender non-traditional and feminist, were put to the forefront of campaigns (Aronson and Fleming, 2023). Thus, the 2018 election illustrates the interplay between social movements, political power, and cultural change.
There are several limitations to this study that should be addressed by future research. First, we are limited in the number and type of races considered. Future research should expand the positions considered to develop a more systematic understanding of gender self-presentations in elections. In particular, it will be important to examine whether the broader use of these self-presentations influence election outcomes in a wider number of races. Although our small sample size allows us to provide an in-depth analysis of the self-presentations, it does not allow statistical comparisons of candidates who were successful with those who were not. A broader sample could help to disentangle the many factors at play (such as party affiliation, type of seat, and local dynamics) in determining the connection between gender self-presentation and election outcomes.
Although our examination has allowed us to consider some intersections between gender, ethnicity, race and sexual orientation, future research should deepen the comparison between different political parties, as well as statewide and local elections, as the constituencies, and therefore, self-presentations, may differ. It would also be interesting to compare the women candidates with their opponents, to discover how men candidates represent gender in their races. For example, the Governor of Georgia, Brian Kemp, may have pushed Abrams into more gender neutral approach in response to his hyper-masculine self-presentation. Finally, a historical and cross-cultural comparison with previous elections and elections in other contexts could help untangle the extent to which these self-presentations are unique to the midterm election in the U.S..
Despite these limitations, our study illustrates important dynamics that emerged in candidate presentation in the 2018 election. The gendered nature of politics today often becomes obscured in a context of extreme ideological and partisan polarization. However, gender self-presentations are an important component of campaigns, as new candidates and resistance after Republican successes are challenging norms for women candidates. These candidates are continuing to defy traditional expectations and starting to alter our expectations for how women in politics present themselves, in digital media and beyond.