Don't Read the Comments: Women's Rhetorical Strategies in the Comments Sections of News Articles

Scholars have taken a range of positions on comments sections, ranging from discussions of whether they have democratizing, deliberative potential (Chen & Lu, 2017; Diakopoulos & Naaman, 2011; McCluskey & Hmielowski, 2011; Rowe, 2015; Ruiz et al., 2011; Waddell, 2018; Weber, 2013) to sounding the alarm about how comments sections are hostile to marginalized users, including women (Clinnin & Manthey, 2019; Løvlie, Inlebæk, & Larsson, 2018; McKee, 2002; Vie, Balzhiser, & Ralston, 2014; Wu & Atkin, 2017). Research has also considered a much broader range of topics related to online discourse than can be fully considered here, including the motivations behind commenting (Lee & Kim, 2015; Wu & Atkin, 2017), how and when users are likely to engage with one another (Cheng, Danescu-Niculescu-Mizil, & Leskovec, 2015; Ksiazek, Peer, & Lessard, 2016), the effect of moderation on the comments sections (Løvlie et al., 2018; Sherrick & Hoewe, 2016; Suh, Lee, Suh, Lee, & Lee, 2018), and more. Work on trolling also explores how behaviors that are framed as "harmless" and "fun" can have unintended and damaging consequences (see Phillips, 2015, 2019). Although this project does not look at comments as examples of trolling, Whitney Phillips (2015, 2019) has provided a useful perspective on how trolls mimic and exaggerate broader cultural issues.

A substantial body of work also focuses on quantitative analysis of comments sections (Cheng et al., 2015; Ksiazek, et al., 2016; Løvlie et al., 2018), while qualitative approaches like the one taken by this webtext and by Clinnin and Manthey (2019) add specific detail and depth to the discussion of what is happening within the comments. For example, while Thomas B. Ksiazek, Limor Peer, and Kevin Lessard (2016) found that commenters are more likely to respond to one another on less popular news videos on YouTube, the content and tenor of those comments are not explicitly discussed. Like Clinnin and Manthey (2019), I advocate for reading and analyzing comments themselves as a way of understanding the discussions as part of a "technological experience within a larger cultural, social, and political system" (p. 37).

While scholars still occasionally herald the comments sections as elements of a digital public sphere that can "facilitate highly deliberative political discussions" (Rowe, 2015, p. 540), a more complex picture of how comments sections function has also emerged. Brett Sherrick and Jennifer Hoewe (2016) found that comment moderation can signal bias and lead to a backlash in opinion against remaining comments and the article itself, while T. Franklin Waddell (2018) illustrated how incivility in comments decreased readers' sense of the importance of the issue and of the article's credibility. Gina Massulo Chen and Shuning Lu (2017) also found that disagreement between users in the comments, whether civil or uncivil, led to increases in negative feelings and aggressive intent, although only uncivil replies led to further incivility (p. 121). While the comments sections do provide a space for deliberation, as many scholars argue, the conversations are often riddled with incivility that can have a negative effect on participation.

Clinnin and Manthey (2019) have offered a rhetorical technofeminist response to what they call "toxic commenting culture," a culture in which off-topic responses, trolling, and outright harassment, abuse, and hate speech flourish (pp. 32–33). Other scholars have observed that women experience a greater degree of sexist responses to their posts in the comments sections and become less likely to participate (Løvlie et al., 2018), or describe the comments sections as "hotbeds for prejudice" (Wu & Atkin, 2017, p. 62). Despite the acknowledgment of how the comments sections can lead to the targeting of women and other marginalized users, there is little qualitative analysis of how (or if) commenters respond to sexist comments or replies, a gap that the present study attempts to address, in line with other scholars in rhetoric and writing studies who have engaged in qualitative analysis of comments sections.

Leigh Gruwell (2017), Heidi McKee (2002), and Stephanie Vie, Deb Balzhiser, and Devon Fitzgerald Ralston (2014) have all considered the role of online spaces in public discourse and the impact of harassment on deliberation and discussion. Gruwell (2017), although she does not focus on comments sections, warned educators to "avoid relying on a Habermasian understanding of the public sphere," especially when teaching writing courses that involve public networks (para. 5). Both McKee (2002) and Vie et al. (2014) have engaged in qualitative analysis of comments in context, and both of their studies acknowledged the shortcomings of comments sections as spaces of public discourse, particularly for marginalized participants. McKee's work explored misconceptions about flaming through an analysis of student discussion boards, wherein inflammatory posts were written in neutral-seeming language and replies challenging those stereotypes would, according to the usual scholarly definition, have been considered flames. McKee complicated our understanding of what constitutes hostility in written online discourse, while Vie et al. considered how comments sections themselvesparticipate in the construction of conversations.

Vie et al. (2014) explored fat-shaming in the comments sections of Jezebel.com, YouTube, and Reddit. Their work provided a qualitative counterpart to the more quantitative analysis of comment moderation and guidelines from scholars such as Justin Cheng, Christian Danescu-Niculescu-Mizil, and Jure Leskovec (2015), Sherrick and Hoewe (2016), and Kil-Soo Suh, Seongwon Lee, Eung-Kyo Suh, Hoseong Lee, and Jaehoon Lee (2018). Oppression within comment threads occurs in complex ways determined, in part, by how comment sections are built and how closely social norms and terms of service are enforced. Clinnin and Manthey, Gruwell, McKee, and Vie et al. illustrated the need for qualitative analysis—and activism, in Clinnin and Manthey's case—around how oppression functions online; this webtext continues that work by directing our focus to gender in the comments and to users' rhetorical strategies in the face of gendered hostility.

Public perception of the comments also tends to be that they are hostile or toxic; many major news media publications have done away with their comments sections entirely in recent years (Finley, 2015). The internet is often conceptualized as an actively or potentially hostile space, as research into uninhibited online behavior flaming has shown (Herring et al., 2002; Jane, 2015; Kayany, 1998; Kiesler, Zubrow, Moses, & Geller, 1985; Lange, 2006; Lapidot-Lefler & Barak, 2012; Lea, O'Shea, Fung, & Spears, 1992; Lee, 2005; O'Sullivan & Flanagin, 2003). Research on the role of gender in shaping online interactions and hostility sheds additional light on why paying attention to sexism online—including in the comments sections—remains relevant.

Gendered online harassment has been an interdisciplinary area of focus within scholarship over the past decade, with numerous scholars from law, media and communications, social and political sciences, women's studies, and more considering what such harassment looks like and what effect it can have (Citron, 2009, 2014; Herring, 1999; Jane, 2014, 2015; Megarry, 2014). The combination of hostile sexism with its overt bigotry and violent threats with benevolent sexism and its covert bias and positive-sounding essentializing is a common focus of study, along with the effects of sexist abuse and its effects on women's presence and participation online. However, research on gendered online harassment has rarely looked at the role gender plays within the comments sections—and much research on gender online has had to be geared towards demonstrating there is a problem with harassment at all. Little scholarship has yet been attentive to how women respond to online sexism or the rhetorical strategies employed to create spaces for being heard.

Women's participation in discourse, online and off, is another thread of scholarship worth considering. Understanding the role of gender in shaping conversational norms has been a research problem for numerous scholars (Brescoll, 2011; James & Drakich, 1993; Sussman & Tyson, 2000). A major meta-analysis of research into who speaks more in mixed-gender discourse environments offline revealed that, despite the stereotype of women as chatterboxes, men tend to take control of conversations in nearly every scenario (James & Drakich, 1993). Further analysis by Victoria L. Brescoll (2011) provided insight into organizational gender imbalances. However, much of the existing research on such imbalances focuses only on face-to-face interactions.

Nan Sussman and Dianne H. Tyson (2000) provided an early challenge to the idea that the internet is a gender-neutral space, finding strong interactions between gender and power in shaping the amount of talk. Turning a similar lens to the comments sections of news articles can reveal whether or not such patterns persist in the discourse communities formed by user-generated interactions and contributions. As with much of the research on gendered online harassment, Sussman and Tyson's research was conducted in Web 1.0. Web 2.0 both enables and demands active engagement from users as readers, producers, and observers of content. Further studies on how communication and gender function in a Web 2.0 environment is therefore essential, given the radically different nature of the environment.

Research into the comments section, uninhibited behavior and flaming, gendered online harassment, and women's participation in discourse has been valuable, but remains incomplete. Studies of the comments sections paint a picture of the complex range of interactions that can happen in the spaces beneath news articles, but analysis of the rhetorical dimensions of responses to hostility remains necessary. Examinations of gendered online harassment have not looked specifically at how women engage with news and with other commenters. Women's participation, online and off, is undeniably shaped by sexism, gender roles and expectations, and the types of socializationexperienced by people regardless of gender. Few studies, therefore, have looked closely at how sexism can affect how women participate in discourse in the comments sections. This webtext is an attempt to draw attention to those gaps, and to begin shedding light on some of the rhetorical strategies women use when encountering sexismonline.

In conducting my analysis of the comments sections, I relied on sociological and feminist rhetorical criticism to identify women's subversive rhetorical strategies in the comments sections of various news articles, and identified codes and major themes through multiple rounds of coding. I approached the research for this project with a deep awareness of my own multiple positions in relation to it: a feminist, a scholar (more specifically, a scholar who studies online harassment), a white, cisgender woman, a woman who has experienced online and offline harassment and abuse, a rhetor, a relative outsider to the community spaces I studied, a frequent participant in similar communities, and so on.¹ In acknowledging my positionality and subjectivity, I move away from any claim to objectivity: another scholar, given the same data, could come to different conclusions. However, I also claim my positionality and subjectivity as strengths in doing this research: my own multiple and layered experiences, combined with the methods and methodology I chose, have shed light on women's rhetorical strategies in spaces that, through truisms like "Don't read the comments," are often opaque.

I chose sociological and feminist rhetorical criticism as the best methods for analysis of the comments. Sociological and feminist rhetorical criticism, as explained by Bernard L. Brock, Robert L. Scott, and J. W. Chesebro (1990), attempt to "describe, interpret, and evaluate the power differentials" associated with things like gendered language, such as sexist language, or women's responses to gender-based hostility (p. 301). Feminist rhetorical criticism rests on "the belief that human issues are best understood as the creation of societal systems" rather than behaviors divorced from external social influence—as a result, women's rhetorical strategies are understood and studied as components of larger discursive structures influenced by gender, power, and so on (Brock et al., 1990, p. 281). The framework of feminist rhetorical criticism therefore allows for an examination of women's rhetorical strategies within the comments sections while also contextualizing those behaviors as part of gendered social patterns of discourse. Case studies allow for deeper analysis of particular interaction, which quantitative analysis can obscure or miss, and stand in as representative examples of trends within the full data set.

The use of case studies also draws on the methods used by Susan C. Herring (1999) in "Rhetorical Dynamics of Gender Harassment On-Line." Herring developed her analysis using two examples of extended sexist rhetoric and gender-based harassment, along with sociolinguistic rhetorical criticism. In describing her own methodology, Herring wrote that she chose two separate sources for data collection because "if significantly similar gender dynamics are found between two otherwise dissimilar data samples, the chances that this is due to coincidence are greatly reduced" (p. 153). The three websites selected for the current study were chosen for a similar reason: if shared patterns were found across BuzzFeed, MSNBC, and Fox News, the chances of the similarities in women's rhetorical strategies occurring due to chance or the idiosyncrasies of a particular audience population are likewise reduced.

Further, the theoretical framework and justification for the focus of the research draws on Feminist Rhetorical Practices, in which Jacqueline Jones Royster and Gesa E. Kirsch (2012) pointed out that rhetorical research has centered on "power elites by race, class, and gender, that is, rich, famous or infamous, white males; and attention directed toward public domains… that is, arenas in which white, elite males have dominated historically" (p. 30). The study moves the focus away from men and onto how women rhetorically navigate spaces that are made hostile to women's participation by the presence of sexist language. Although Rosyter and Kirsch might have noted that the comments sections of news websites have still "been shaped or controlled by power elites," scholars like Michael McCluskey and Jay Hmielowski (2011) might have responded that they are spaces for discourse that occur outside the historical power structure of communication with news organizations (p. 30).

I also draw on Royster and Kirsch's (2012) description of a feminist rhetorical practice that is interdisciplinary in nature, and involves critical work "that is helping us to know more broadly and deeply the nature, scope, impacts, and consequences of rhetoric as a multidimensional human enterprise" (p. 41). In drawing on research from multiple disciplines to create bridges across the gaps between them, and analyzing women's rhetorical strategies to more clearly understand the "nature, scope, impacts, and consequences" of those strategies, I align myself with the feminist rhetorical practices Royster and Kirsch explored.

Turning to the concrete application of methods, I selected three news websites to analyze in my attempt to develop a better understanding of women's rhetorical strategies in the comments sections. The study focuses specifically on BuzzFeed, Fox News, and MSNBC: news websites that have different political leanings, different reader/commenter demographics, and comments sections with different functionalities.² I used a Google search for each website to allow for searching within specific date ranges (a function not available on the sites themselves) and key words related to women and feminism to identify articles where sexism, and thus women's responses to sexism, were more likely to be present. One article per month per website from the year 2016 were randomly selected from the Google results, leaving me with a pool of 36 articles. Six articles were randomly set aside for double coding.

I was left with a pool of 5,532 comments, all of which were read to identify comments employing sexist language, and comments responding to sexism.³ Of those, 676 (or 12.2%) contained some type of sexism, and 136 (or 2.4%) were women's responses to sexism. The remaining comments included non-sexist discussions of the topic, numerous examples of spam, off-topic conversations, and comments that contained other forms of bias and bigotry (racism, homophobia, xenophobia, ableism, and others) but did not contain sexism (the focus of the present project). I identified comments containing sexism as those including sexist stereotypes, objectification, sexist or objectifying language or images in a username or profile picture, generalizations or essentializing remarks about women, hostility towards individual women or groups of women that included gender as a component of the hostility, and threats of violence. I identified comments responding to sexism as those left as a direct reply to comments containing sexist language, comments mentioning a pattern of sexism within the space, and comments that included the username of someone who had left a sexist comment but may not have been a direct reply.

Determining the gender of commenters online is a fraught business. The quasi-anonymous nature of commenting on Fox News and MSNBC in particular made determinations challenging: while site users must register to leave comments, user names and profile pictures are chosen at the time of registration and may be as plainly gendered as John Smith or as unidentifiable as FlyBird. Even on BuzzFeed, where comments are left via Facebook profiles, determining gender is an uncertain task. Names and profile pictures are not reliable indicators of what someone's offline gender might be. Katherine DeLuca (2015) described additional difficulties associated with identifying gender in online spaces in her analysis of Pinterest users, including the ways that spaces perceived as dominated by women are often devalued. She also noted that "gender identity and gendered rhetorical spaces are composed and created in online spaces" ("Gendered Rhetorical Spaces on Social Media"). Similarly, my interest for the purposes of this study is not whether each commenter can definitively be said to be any particular gender; instead, my interests are the rhetorical strategies of commenters who seemed to present themselves as women, and who were responded to by other commenters as if they were women. For the purposes of this project, therefore, the true offline gender of commenters is not relevant. As a result, I used names, profile pictures, self-identification, context, and patterns of response to determine the presented/perceived and constructed online gender of commenters, with the acknowledgment that such a method is likely to include at least some false positives.

Once identified as women's responses to sexist language, the comments were read again, taking note of tone, common language patterns, and methods of response to begin identifying themes and rhetorical strategies. Initial categories included broad frameworks like explanation of the problem, sarcasm, identifying sexism, reaffirming feminism, and silence. During further rounds of coding, categories were refined and finalized in terms of feminist theory and feminist rhetorical terminology, leading to final codes related to kairos, polyvocality, validation, sarcasm, standpoint claims, and silence, all of which help to frame the responses to sexism as rhetorical actions within the often-hostile space of the comments sections. What I found is that women's rhetorical strategies create a constellation of distinct but related methods of responding to sexism—networks and nodes of rhetorical movies that can show us patterns of discourse, if we only take the time to look.

A final note on methods and methodology: The narrow focus of this project limits the value of the results, given the many intersectional factors at play in online discussions, including comments about race, ethnicity, sexuality, disability, and more. Additional work taking an even more intersectional approach to how marginalized groups use various rhetorical strategies in the comments sections will move the conversation forward; my hope in beginning the conversation is that such work will be undertaken. It can also be argued that there is no neutral or at least less obviously biased publication represented in the current study, which is undeniable. However, application of the patterns observed in this study to the comments sections of websites that are regarded as more neutral will enable comparison of how different discourse communities might respond to sexism and other -isms, and how participants in other spaces engage in subversive rhetorical practices.

See the Appendix for a breakdown of all articles used and comments on each article.

¹ Several things should be noted about this project. The first is a limitation: while I address some intersectional concerns in the rhetorical strategies I describe and the specific examples that highlight them, gender was the primary lens through which I analyzed the comments. Gender was an easier (although not easy) category to identify in the absence of real names or profile pictures. While a more intersectional analysis of behavior and rhetorical savvy in the comments sections is necessary, this project starts with gender as its overarching focus. Further, what I found when reading women's comments challenges the notion that comments sections can be considered sites of democratic discourse, as they are often described in scholarship: women's comments are far less common than men's, subjected to far more hostility when they do appear, and are rarely raised as a challenge to sexism. What I also found, however, was extensive rhetorical richness and versatility in the strategies women did use to respond to sexism: seizing kairotic moments, participating in polyvocal discussions, validating one another, employing sarcasm, using standpoint claims, and, often, choosing silence. The wide variety of women's strategies for addressing sexist comments suggests that the comments sections are valuable sites for additional, focused analysis—and ripe for potential interventions that may encourage more equitablecontributions and reduce instances of abuse and hostility. (Back to Text)

² Since comments are published in a nominally public forum and no intervention was applied to the commenters themselves, IRB approval was not required. However, consideration was given to commenter privacy. Commenters know their contributions become part of public discourse, but may not be comfortable when their contributions become the focus of academic attention. The names of commenters have been changed except where a user name is relevant to the analysis. These instances are noted when they occur. (Back to Text)

³ The Internet never stops changing: many of the comments sections once visible on the selected websites have since disappeared; some conversations have likely continued since the dates of my data collection. I copied the 813 comments selected for coding into a spreadsheet, but for many of the articles, the spreadsheet is the only remaining record. Changing website policies and attitudes towards comments sections are the likely reason for some sections having been deleted. (Back to Text)