Online aggression represents a serious, and regularly occurring, social problem. In this piece the authors consider derogatory, harmful messages on the social media platform, Twitter, that target one of three groups of women, Asians, Blacks, and Latinx. The research focuses on messages that include one of the most common female slurs, “b!tch.” The findings of this chapter reveal that aggressive messages oriented toward women of color can be vicious and easily accessible (located in fewer than 30 seconds). Using an intersectional approach, the authors note the distinctive experiences of online harassment for women of color. The findings highlight the manner in which detrimental stereotypes are reinforced, including that of the “eroticized and obedient Asian woman,” the “angry Black woman,” and the “poor Latinx woman.” In some exceptions, women use the term “b!tch” in a positive and empowering manner, likely in an attempt to “reclaim” one of the common words used to attack females. Applying a social network perspective, we illustrate the tendency of typically hostile tweets to develop into interactive network conversations, where the original message spreads beyond the victim, and in the case of public individuals, quite widely. This research contributes to a deeper understanding of the processes that lead to online harassment, including the fortification of typical norms and social dominance. Finally, the authors find that messages that use the word “b!tch” to insult Asian, Black, and Latinx women are particularly damaging in that they reinforce traditional stereotypes of women and ethno-racial minorities, and these messages possess the ability to extend to wider audiences.
Felmlee, D., Rodis, P. and Francisco, S. (2018), "What a B!tch!: Cyber Aggression Toward Women of Color", Segal, M. and Demos, V. (Ed.) Gender and the Media: Women’s Places (Advances in Gender Research, Vol. 26), Emerald Publishing Limited, pp. 105-123. https://doi.org/10.1108/S1529-212620180000026008Download as .RIS
Emerald Publishing Limited
Copyright © 2019 Emerald Publishing Limited
Online interactions allow individuals to communicate with one another, read the news, share information, and take part in a global network. However, not all cyber interactions are positive, and one look at social media reveals that a vast number of divisive and nasty interchanges are publicly available. Due to its prevalence, the potential for anonymity, and the ease with which harassing messages can spread, cyber aggression represents a growing social problem. Furthermore, a recent study by the Pew Research Center (2017) finds that online harassment is directed regularly toward those with readily accessible characteristics, such as race/ethnicity and gender. Therefore, this study aims to examine online harassment that is oriented toward targets that are both easily accessible, and historically subjected to discrimination, that is, women of color. Our goals are to examine the thematic content and the network spread of these types of aggressive messages via the social media platform, Twitter.
Aggression in Social Media
Aggression and violence toward women in the media remains extensive and problematic, as found in numerous studies. Reoccurring themes include that females are more likely than males to be sexualized (e.g., Dill & Thill, 2007; Miller & Summers, 2007), treated as objects (e.g., Burgess, Stermer, & Burgess, 2007; Stankiewicz & Rosselli, 2008), and receive less attention in many media forums (Downs & Smith, 2010; McCabe, Fairchild, Grauerholz, Pescosolido, & Tope, 2011; Women’s Media Center, 2017). Goffman (1979) noted the subtle nonverbal cues in magazine advertisements that highlight women’s role as submissive, and recent studies of contemporary advertisements continue to document significant differences between the portrayal of women and men. For example, compared to men, women are more likely to be depicted with flawless skin, as passive, taking up less space, and with dismembered body parts (Conley & Ramsey, 2011). Racism also persists within popular media, with Black characters frequently linked to violence, a pattern that can reinforce stereotypes that Blacks are violent among video game players (e.g., Yang, Gibson, Lueke, Huesmann, & Bushman, 2014).
Furthermore, the depiction of minorities within advertising and other forms of media reflects dominant, cultural stereotypes about these groups (e.g., Ki-Young & Joo, 2005; Taylor, Lee, & Stern, 1995). For example, Asian women tend to be represented as submissive and docile, but at the same time, also exotic and oversexualized (e.g., Ki-Young & Joo, 2005; Mok, 1998). Black women are likely to have the most diversity in their media representation, while still generally portrayed within advertising as less educated and in more token roles (e.g., Taylor et al., 1995). On the other hand, Latinx individuals are highly underrepresented in mainstream advertisements in regard to their relative proportion in the United States population (e.g., Ki-Young & Joo, 2005; Taylor et al., 1995). Moreover, when Latinx women are visible, the media often depicts them as romantic, overtly sexual, faithful, self-sacrificing, and family oriented (e.g., Ki-Young & Joo, 2005; Merskin, 2007; Taylor et al., 1995).
Social media represents one of the newest forums for media aggression, and one where hostility is experienced up close and personal. Aggressive attacks within such media domains are frequent. Based on a nationally representative survey (Pew Research Center, 2017), approximately 41% of Americans report having personally experienced some form of online harassment, with 18% describing particularly severe behaviors such as physical threats and sexual harassment. Blacks (25%) and Hispanics (10%) were more likely than Whites (3%) to report being targeted because of their race or ethnicity, and about twice as many women (11%) as men (5%) recount being harassed online due to their gender (Pew Research Center, 2017). Facebook and more recently, Instagram, offer some of the most popular venues for online aggression (Gibbs, 2017). Furthermore, aggressive content remains more common within socio-political forums that take place online as compared to those that transpire offline (Rost, Stahel, & Frey, 2016).
The current project examines social interchanges that occur on the digital social media venue of Twitter, with an emphasis on negative content. Twitter represents a popular micro-blogging service in which people exchange short messages, or “tweets,” with over 330 million active users per month in 2017 (Bray, 2017). In addition, Twitter is another frequent social media location for harassment, with at least 15,000 bullying “tweets” occurring daily, according to one report (Xu, Jun, Zhu, & Bellmore, 2012). Furthermore, the timing of the bullying messages is not uniform and occurs most often during the evenings when individuals may not be at work or school (Bellmore, Calvin, Xu, & Zhu, 2015). Aggressive incidents based on race, gender, or sexual orientation also are readily and quickly accessed in this form of social media (Sterner & Felmlee, 2017). Our focus is on instances of cyber aggression on Twitter.
Cyber aggression, which refers to intentional, online messaging with the aim of insulting or harming someone, poses a serious social problem, and one that extends worldwide. Both victims and perpetrators of digital aggression are at risk of a host of negative behavioral and psychological outcomes (e.g., Nansel et al., 2001). Being the victim of peer aggression has particularly deleterious consequences and is associated with anxiety, depression, and poor academic performance (e.g., Faris & Felmlee, 2014; Nansel et al., 2001; Willard, 2007). Youth who were the targets of cyberbullying also had more suicidal thoughts and were more likely to attempt suicide than those who had not experienced this type of aggression (Hinduja & Patchin, 2010). Furthermore, the harm unleashed by derogatory messages on Twitter extends beyond adolescents in school and college. A number of cases in the news report instances in which public figures deactivate or delete their Twitter accounts after experiencing incidents of repeated bullying. Yet, as we will address subsequently, some well-known individuals choose to become further involved in these large, public conversations instead of retreating from social media.
Social Networks of Cyber Aggression
One key, troublesome characteristic of cyber aggression within social media is its ability to extend beyond the original target to reach a wider audience. In cases involving public figures, this pattern can rapidly expand to result in the dissemination of hundreds or more derogatory messages and responses to those messages. One of the main goals of this research, therefore, is to examine the social networks of message interchanges that emanate from cases of cyber aggression on Twitter. We investigate the web of tweets, “retweets” (in which one person forwards the same message), “replies” (in which one person directly responds to a previous message), and “likes” (when an individual indicates, electronically, that they like a tweet) that emanate from incidences of aggression on Twitter. A social network framework facilitates an understanding of the interactions that develop in response to cyber aggression, and allows us to depict visually these online interactions. More generally, a network perspective underlines the fact that hostile digital interchanges do not represent solitary, individual incidents, but that they remain rooted in an extensive, social relational, context.
Social Processes in Cyber Aggression
Research identifies two of the social processes that contribute to the development of cyber aggression, and these include the enforcement of social norms and the establishment of social hierarchies (Felmlee & Faris, 2016). For example, the high levels of online harassment aimed at individuals who do not conform to the traditional, societal expectations of heterosexuality (e.g., Felmlee & Faris, 2016; Hinduja & Patchin, 2010) illustrate the endorsement of the social norms associated with “compulsory heterosexuality” (Rich, 1980). Furthermore, the motivation to increase one’s status among peers represents another fundamental mechanism involved in the development of online and offline aggression (e.g., Sijtsema, Veenstra, Lindenberg, & Salmivalli, 2009). Much of school peer victimization occurs among relatively popular students, for instance, many of whom are jockeying for similar positions, grades, and respect from peers (e.g., Faris & Felmlee, 2014).
According to Ridgeway (2011), status arrangements remain fundamental to the production of inequality within the societal gender system, and we extend that argument more specifically to women from underrepresented ethno-racial groups in America. We argue that cyber aggression aimed at women of color stems from hierarchical social systems, and in particular, the systems that place these women toward the very bottom rungs. Demeaning messages that target minority females represent attempts to reinforce the subordinate position of these women in our society. Perpetrators likely view such attacks as avenues for enhancing their own status and respect in the eyes of their peers and online followers, who they believe will admire their ability to harass and embarrass women judged as threatening the dominant hierarchy. Furthermore, we expect that women of color, who are doubly marginalized based on both race/ethnicity and gender, are apt to be particularly susceptible to this genre of aggression.
We apply an intersectional lens (e.g., Collins, 2000; Crenshaw, 1991; MacKinnon, 2013) to examine experiences of online aggression toward women of color. An intersectional perspective is fundamental to the study of gender and race, because it emphasizes that an improved understanding of these socially constructed distinctions arises from consideration of the ways in which multiple social categories, such as gender and race, interact with each other (Shields, 2008). Based on an intersectional standpoint, we not only consider the effects of the singular processes of racism and sexism, but a more specific blend of the two that goes beyond simply adding racism and sexism together.
We begin by noting that women, and women of color more specifically, are not a homogeneous group. There is apt to be a qualitative departure between the digital-based experiences of Asian women, Black women, and Latinx women. The concept of “misogynoir” (Bailey, 2010), which refers to the distinctive condition of violence and stereotyping aimed at Black women, highlights the unique, intersectional blend of hostility experienced by this particular subset of minority women. Furthermore, all three of the groups we examine represent broad and diverse ethno-racial categories. Not all Latinxs are the same, for instance, and differ widely in ethnic heritage and along other social identities that likely influence their involvement in online interactions. Here we narrow our focus to shed light on the unique, aggressive electronic encounters that occur at the intersection of two highly salient social dimensions, those of gender and ethno-racial identity.
We examine the occurrence of aggressive, harmful Twitter messages that are directed toward several groups of women of color, including Asian, Black, and Latinx women, with a focus on one of the most common female slurs, b!tch (Felmlee, Inara Rodis, & Zhang, 2018). Next, we discuss regular themes that emerge within these types of communications and note the distinctive experiences of women in each of our three groups. Finally, given that negative messages on Twitter have the potential to spread beyond the original target, we describe illustrative cases in which these hateful messages form interactive networks that extend past the original victim.
In this project, we utilize the term “b!tch” because it directly counters the common stereotype of submissiveness for females. “B!tchy” women are considered aggressive or forward. That is, such women do not portray the typical lady-like and feminine ideals of passivity, selflessness, or subservience, and the slur “b!tch” brands them for acting in ways that defy these cultural traditions. On the other hand, recent popular cultural trends (e.g., Celious, 2002; Fairclough, 2008; Westmoreland, 2001) reveal a second interpretation of the term “b!tch,” in which case the word is meant to be inspiring, rather than offensive.
Positive uses of the word “b!tch” often represent what is termed “reappropriation,” or “reclaiming” (Cedar, 2008; Croom, 2011). Reappropriation of insulting terms occurs when individuals who have been targeted by a particular slur, typically members of an oppressed minority, attempt to re-define the word in a positive and empowering manner. Reclamation involves vulnerable individuals who are emphasizing their own ability for independence and self-definition (Cedar, 2008). For example, Celious (2002) cites the case of famous and successful female rappers, primarily women of color, who self-identify as “b!tches” as a display of power or self-determination. According to Westmoreland (2001), feminist groups and organizations such as Riot Grrrl and Lilith Fair also use the term “b!tch” in their communications to celebrate a modern femininity that subverts patriarchy and redefines negative stereotypes. Therefore, while always connoting strong and often hostile traits, the word “b!tch” can be used either in a negative, insulting manner or in a positive act of reappropriation. In our data set, there were examples of both uses of “b!tch.” Here we focus primarily on the dominant, negative use of the term, but we also note and provide an illustration of a network of interchanges in which the word, “b!tch,” appears to be reclaimed.
We collected data from the Twitter API using the Twitter keyword search function of the social network program, NodeXL (Smith et al., 2010), over a period of one week at the end of February 2017. We downloaded small samples of tweets at a time, all of which were published publicly on Twitter. At the time of data collection, these tweets were limited to a maximum of 140 characters. We selected our sample by combining two types of keywords in the same search. The first set of keywords identified messages that concerned ethno-racial minority groups, and the second selected messages that contained words frequently used to insult females. In order to choose tweets related to our three groups of minority women, we searched for keywords derived from a collection of words to describe Asians, Blacks, or Latinxs, as shown in the first column of Table 1. The second set of keywords referred to typical derogatory female slurs. Previous research finds that the word “b!tch” represents one of the most frequently used terms to insult females in online aggression (Felmlee et al., 2018), and therefore our current study focuses on messages that include that particular feminine slur.
|Target Group||Terms Combined to Download Data||Total Tweets Downloaded (Percent Total)||Time to Find Aggressive Tweet from Downloaded Data (s)|
|Minority Term||+||Female Term|
|Asian Women||asian, ch!nk, ch!!ky, ch!!g ch!!g, Chinese, Japanese, Filipina, Indian, Korean, Oriental, Philippine, Vietnamese, Yellow||+||B!tch||6,111 (25.90%)||23|
|Black Women||Black||15,479 (65.59%)||18|
|Latinx Women||Beaner, Hispanic, Latina, Latinx, Mexican, Wetback||2,008 (8.51%)||16|
Note: Tweets collected through NodeXL.
In total, we downloaded and analyzed 23,598 tweets. Approximately two-thirds of these tweets were directed toward Black women (15,479 tweets), just over one-quarter targeted Asian women (6,111 tweets) and the remaining (2,008 tweets or roughly 9% of all tweets) were oriented toward Latinx women. There are several possible ways to explain why Black women compose the large majority of data. First, Black Americans constitute a racial minority group –with a long and complicated history within the United States, and one that is large in size (13.3% of the total population), second only to Hispanics/Latinx (17.8%), and with Asians composing a relatively small percentage (5.7%) (U.S. Census Bureau, 2016). Given the prominence of this group of Americans, it is not surprising that a large portion of the data references Black women. Second, in two notable content analyses of magazine advertisements (e.g., Ki-Young & Joo, 2005; Taylor et al., 1995), researchers find that Blacks tend to be the most represented minority group when compared to Asians and Hispanics, whereas Hispanic or Latinx individuals remain vastly underrepresented in comparison to their constituency in the overall population. Ki-Young and Joo (2005) find that Latinx individuals are 12.5% of the whole population but only 2.6% of the individuals in the advertisements (p. 665).
In addition, within the US, individuals of Black ethno-racial origins are more likely to identify as Black than Asian or Latinx individuals are to identify as Asian or Latinx. Recent immigrants from the latter groups are more likely to self-identify with a specific ethnic group (e.g., Vietnamese as opposed to Asian, or Peruvian as opposed to Latinx), due to social norms and expectations (e.g., Kiang, Perreira, & Fuligni, 2011; Lee, Wong, & Alvarez, 2009; Mok, 1998). Therefore, our data searches for “b!tch” tweets about Asians and Latinx likely underrepresent those groups of women. Finally, one of the stereotypes specifically ascribed to Black women is that they are strong (Beauboeuf-Lafontant, 2009), which may make them particularly susceptible to being attacked with the label, “b!tch,” suggesting that they are much too strong.
In order to provide anonymity to participants, we replaced Twitter handles or user account references with “@USER#” and website links with “URL.” However, portions of our sample contained messages in which the meaning was unclear, or provided mixed messages. Here we focused primarily on messages that were clearly negative in content and that fit our definition of cyber aggression – messages with the intent to insult or harm. Nevertheless, some tweets represented instances in which the key derogatory term seemed to be used in a positive manner, as mentioned earlier, and therefore we also depict an example of such a case.
Accessibility of Negative Messages
Our data do not enable us to provide accurate estimates of the frequency with which cyber aggression focuses on women of color on Twitter. Such messages likely comprise a very small portion of all the content that transpires on this forum, given that numerous messages can contain news updates, factual information, and a wide variety of impersonal content. Therefore, we instead investigate the ease of access to negative, insulting tweets.
In order to examine the accessibility of online aggression toward women of color, we recorded the time it took to locate such texts within Twitter messages. Following Sterner and Felmlee (2017), two of the authors read through the sample of downloaded tweets and recorded the time (in seconds) that it took to find the first instance of a clearly negative tweet, as opposed to one that was positive or neutral in content. They performed this task on three different occasions in order to calculate the average time to locate an aggressive tweet, based on our key search word (b!tch). Shorter times suggest that the negative content is more readily available to the public than longer periods.
Social Network Analysis
Next, we examine the ways in which online, intersectional hostility can spread beyond the two most immediate individuals (i.e., the bully or perpetrator and the victim) to involve others. Drawing from our sample, we identify three examples that are illustrative of online interactions that can arise in aggressive cases. We visualize these interactions as social networks in which the actors represent individuals involved in a particular “conversation” on Twitter that includes negative content. The ties between actors represent tweets, retweets, replies, and likes of both the aggressive tweet and any other responses to that tweet.
In addition, we investigate the social roles of individuals in each network (following the work of Salmivalli, 1999; Sterner & Felmlee, 2017; Xu et al., 2012). We identify the aggressor, perpetrator, or “Bully,” which refers to the individual tweet account who sends the offending tweet, and any individuals who support the original aggressive messages, whom we label as “Reinforcers.” We also identify the target or “Victim” of the tweet and the users who support the victim, that is, “Defenders.” We do not label either the gender or the race of the participants in the interaction, due to conflicting information, or a lack of data on those characteristics. In some of our particularly negative cases, for example, users employed cartoon characters or non-identifiable images in their descriptions, perhaps to hide their hostile messages behind fictional facades.
Exploring Intersectional Themes
We also discuss the overall themes highlighted by the incidents of cyber aggression in our data. We note that aggressive incidents often suggest that there are particular social expectations to which each of the three groups of minority women are held, and then admonished when viewed as not fulfilling those expectations. In addition, we note the specific stereotypes emphasized and consider the ways in which women of color specifically, and not just women or ethno-racial groups, become targets.
Accessibility of Negative Tweets
To consider the pervasiveness of hostility toward women of color, we examine the data systematically to note how quickly we could locate a tweet that was derogatory in content toward this subsample of women. It took an average of 19 seconds of searching through the data to find a tweet that used the term “b!tch” in an aggressive manner toward one of these three groups of women. Tweets directed toward Asian women were located after about 23 seconds, while it took 18 seconds for messages aimed at Black women and 16 seconds for those targeting Latinx women (see Table 1). In other words, harmful messages toward women of color were located extremely quickly, within less than half a minute, on this social media platform.
Conversation Network for Black B!tch
In one of our three types of data acquisitions, we searched for messages that contained the combination of the two keywords “black” and “b!tch.” As can be seen in Fig. 1, this combination of search terms yielded a particularly large social network of tweets, replies, retweets, and likes. As one of the most visible and entrenched minority groups in American history, Black women appear to be common victims of online aggression in our study. As found in previous research, this group of women are subject to prejudice and criticism for their strength of character and can be labeled as angry or sexually lascivious (e.g., Beauboeuf-Lafontant, 2009; Harris-Perry, 2011). On the other hand, not all of the messages in this extensive interchange contained intentional aggression. Some of them used the term b!tch in a positive, or humorous manner, and a number of others were opaque or contradictory, illustrating the complex conversations that emerge online. Nevertheless, the considerable mass of the network illustrates the common practice of using a primarily negative, feminine slur (Felmlee et al., 2018) to discuss this group of minority women within Twitter.
Networks of Everyday Instances of Cyber Aggression
Next, we illustrate examples of cyber aggression toward minority women, and we analyze the social networks of conversations that develop around these cases, noting the social roles of the actors. We choose three examples that highlight thematic trends, with one illustration from each ethno-racial group. While these examples do not encompass all the patterns we observe in our data, they do demonstrate several common ways in which women of color are insulted and the types of interactions that emanate from those insults. Moreover, these specific examples are not unique in their level of negative content, and they exemplify what we found to be routine, everyday aggression in our searches.
Cyber Aggression Toward an Asian Woman
In the first example, we examine the following hostile tweet toward an Asian woman: “Like ch!!ky b!tch sit down the only reason your hear is because your mum poked herself wi a spring roll [Winking Face Emoji].” In this text, we note the submissive and sexual overtones. The tweet’s injunction to “sit down,” and with the implication to “be quiet,” highlights submissiveness and passivity, while the comment of her “mum pok[ing] herself wi a spring roll” hints at Asian women’s stereotypical sexual exoticism.
Using the negative message as a starting point, we portray the aggressive tweet and all tweets associated with the original hostile message (i.e., retweets, replies, and likes) in the form of a social network, as shown in Fig. 2. In this network, we see that the Bully (indicated by a solid square) sent out an aggressive tweet. The network also consists of seven Reinforcers, that is, seven other individuals who expressed support for the hostile message, but who did not directly victimize the target or another person (Reinforcers indicated by solid diamond). Note that there were no Defenders in this “conversation.” That is, no one criticized the content of the nasty message, nor did anyone express support for the Victim. Furthermore, two of the Reinforcers both liked and retweeted the original message, therefore disseminating the derogatory message to a wider network of Twitter users.
Cyber Aggression Toward a Black Woman
Our second example explores an aggressive tweet that targets a Black woman with the use of our key feminine slur (as well as other slurs) in the following message: “@USER1 ugly black a!! b!tch YOU BROKE AND UGLY AND YOU JUST BROKE AF. With yo floating a!! ponytail [Skull Emoji Skull Emoji Skull Emoji] GET THE F!!K OUTTA HERE”. The content of this tweet demonstrates common perceptions of Black women as poor (repeating “BROKE” twice) and not fitting the ideal beauty standards (also using “ugly” twice).
The conversational network, displayed in Fig. 3, contains not one Victim, but two (indicated by solid triangles). In addition, three Defenders (i.e., circles) take the side of the Victims in reaction to the Bully’s (i.e., a square) direct, hostile tweet. Within the network, the Bully attacks both Victims. One of the Victims replies to the other Victim, moreover, and then responds to the Bully. Thus, the principal actors interact directly with one another. Furthermore, the content of the message appears to offend three Defenders, who offer their support for one of the Victims. One of the Defenders, for example, attempts to end the hostile interchange by telling the Bully to “stop talking about it.” In this illustration, we see that several individuals refused to reinforce the negative message presented in the Bully’s tweet, and instead, expressed their backing of the Victim.
Cyber Aggression Toward a Latinx Woman
In our third example, we analyze an aggressive tweet that was directed at a Latinx woman. The content of the aggressive tweet is as follows: “@USER1 @USER2 Shutup dumb b!tch go back to school and sell some crack dumb Texas B!!ner.” In this tweet, the content features stereotypes about Latinx women, namely being unintelligent. Here the sender of the original tweet, in addition to using the word “dumb” twice, tells the targeted Latinx woman to “go back to school,” suggesting that she is uneducated.
As shown in Fig. 4, the network includes one Bully who again targets two Victims. In addition, three Reinforcers buttress the Bully’s message by indicating that they liked the tweet. In this instance, no Defenders appear to align themselves with either of the Victims in an effort to oppose the Bully’s original tweet. The hostile message receives validation through the Reinforcers, however, whose actions suggest they like it. Moreover, neither Victim responds directly to the Bully’s message. This pattern of action, and inaction, emerges to leave the Bully’s original derogatory sentiment unchallenged. It suggests, too, that the followers of the Bully and the Reinforcers, none of whom chose to defend either Victim, support the racist and sexist attitudes in the Bully’s tweet.
Network Example of a Black Woman Reclaiming B!tch
Not all cases in our data contain negative content, as noted earlier. In some instances, tweets containing the term “b!tch” represent examples of reappropriation or reclaiming, in which a message is intended to be uplifting rather than insulting. One such example occurs in the following tweet: “Black Girls are Ugly ‘B!tch Where’ [Nonplussed, Disgruntled Face Emoji] Black Girls Are Goddesses, Queens & Powerful In All Shades #IAmMySistasKeeper [Raised Fist Emoji].” In this tweet, the author counters the derogatory, misogynoir, statement at the beginning of the message, “Black Girls are Ugly,” by describing Black women enthusiastically as “Goddesses, Queens & Powerful In All Shades.” The use of an emoji in the shape of a fist at the end of the tweet, following the powerful hashtag (#IAmMySistasKeeper), also emphasizes that this message is meant to be empowering.
The “interactional” network that emanates from this message, displayed in Fig. 5 , does not have a specific Victim or a Bully. Instead, the central node is the Defender, that is, the individual who sent the message that defends Black women who are insulted with the terms “ugly” and “b!tches.” The other actors in the network consist of people who also defend Black women, by approving of this message. In total, there are 37 people who support the message, with 10 retweeting the message and 27 liking it. In this example, the positive and empowering tweet is endorsed by others, suggesting that the message is well received and successful at reaching additional users.
Network of an Online Fight Against Aggression Toward Women of Color
In addition to the everyday instances of cyber aggression toward women of color illustrated above, we include one example of a Twitter social network that spread particularly widely. Several of the Twitter messages in our data set that contained our key slur term concerned a case of aggression aimed at a public figure, US Representative, Congresswoman Maxine Waters. The aggressive incident began not online, but on television, when political commentator and then TV host Bill O’Reilly responded to an address by Waters. O’Reilly was on Fox News watching a clip of Rep. Waters, a Black woman from California, speak to Congress when he declared that he did not listen to a word Waters said, because he was too focused on her “James Brown wig.” In this instance, O’Reilly directly attacked Waters’ appearance and negated her speech, with the use of an insulting statement regarding her hair. His comment sparked outrage from many audiences, especially on the Internet. O’Reilly later apologized for his statement, but by then, social media had dispersed the content extensively. In fact, both Waters and O’Reilly were targets of online aggression in reference to this incident. Rep. Waters was called a b!tch, and others noted how O’Reilly’s comments represented a prime example of misogynoir in America.
On the other hand, the online responses to the Waters incident also demonstrate the role social media can play in disseminating positive messages. Over time, it was apparent that the sentiment expressed on the Internet toward Rep. Waters became increasingly supportive. For example, affirmative and sympathetic messages toward Waters comprise the bulk of the extensive graph that we display in Fig. 6. This figure represents the Twitter network that developed from the key word search for “Rep Waters” the day following the initial aggressive incident. In response to her experience with O’Reilly, Waters also subsequently became part of a popular, positive social trend by tweeting #BlackWomenAtWork, which provided a space for Black women to share their difficult experiences in the workplace. The hashtag became a leading rally for a broad Twitter following that discusses Black women’s differential treatment in the workplace.
According to our study, derogatory messages attack women of color on a regular, daily basis on Twitter, and such messages are readily available to the public. In fewer than 30 seconds, we were able to locate an aggressive tweet oriented toward either an Asian, Black, or a Latinx woman in our sample of Twitter data. Furthermore, our search for a negative term relied only on the inclusion of a single, common, insulting feminine term (b!tch), whereas there are hundreds of other typical feminine slurs that could have been used to locate an even larger sample (e.g., c!nt, h!, sl!t). The ease with which we located aggressive messages based on one key insult suggests that women of color likely face much online hostility.
We find that aggressive messages between two people often develop into an online “conversation network,” in which others become involved in the interchange, frequently reinforcing the negative content and retweeting it to a larger audience. In some of these cases, users defended a Victim and rebuked the Bully, or attempted to end the interaction. In other situations, especially those involving public figures, an original insult was communicated widely to others, forming an extensive social network of online communications, with varying ratios of negative and positive commentary. These networks demonstrate one of the unique, and deleterious, characteristics of cyber aggression, which is its ability to spread quickly beyond the two initial, main actors. Even when the reaction to an attack by others generates a good deal of support, the dissemination of a hostile message likely further harms minority women victims who were insulted or humiliated by the message.
In the cyber aggressive cases we examined, we also see evidence of “intersectional aggression,” that is, unique shades of hostility that target women from differing ethno-racial backgrounds. Messages oriented toward Asian woman, for example, often contain sexual overtones, or expectations for silent and obedient women. On the other hand, Latinx women are denigrated for being poor and/or uneducated, and for their supposed involvement in the illegal transfer of people or drugs across the Southern US border. The overrepresentation of messages aimed at Black females that contained the word b!tch, furthermore, depicts an example of misogynoir (Bailey, 2010) that highlights the modern stereotype conveyed in these attacks of the “angry black woman” (Harris-Perry, 2011). Thus, the inclusion of other derogatory slurs (e.g., c!nt, sl!t), such as those based on sexuality or a lack of education, may uncover more examples that involved Asian or Latinx women, respectively. More generally, through exploring these aggressive messages toward women of color, we see the utility of applying an intersectional framework (e.g., Collins, 2000; Crenshaw, 1991; MacKinnon, 2013) to the study of electronic forms of bullying.
Two of the social processes involved in cyber aggression include the enforcement of traditional, social norms and the creation and maintenance of status hierarchies, according to prior research (Felmlee & Faris, 2016). Both of these processes contribute to the pattern of aggression documented within this study. Social norms that enforce traditional, negative race, and gender stereotypes, are evident in our data, as noted above. In other words, electronic bullying does not simply reflect individuals lashing out in anger or in revenge to harm another in an idiosyncratic manner. These incidents systematically reinforce deleterious, societal race and gender stereotypes. In addition, cyber bullying may increase the status and esteem of the perpetrator in the eyes of his or her followers. We find that people can and do retweet and express liking for mean attacks on a woman of color (e.g., those Reinforcers we identified within aggressive networks). Such actions further encourage the Bully and may convey respect on the part of followers for the Bully as a “leader” in online interactions. As Ridgeway (2011) notes, fundamental, status processes contribute to the construction and maintenance of societal gender inequality. Here we see ways in which bullying in social media serves to reinforce the low position of women of color in the dominant, societal hierarchy, and serves to maintain inequality located at the intersection of both race/ethnicity and gender.
Not all of our findings uncover negative uses of Twitter regarding women of color, however. Our results also demonstrate the ways in which social media communications can be employed constructively. In certain situations, for example, women can use tweets to reappropriate the term “b!tch,” and apply it in a way that underlines the importance of strength and toughness among underrepresented groups of women. Furthermore, in the case of a political figure, Rep. Waters, Twitter was used to generate a large following of support for Waters, following her harassment, and to create a new venue in which Black women can discuss and address instances of harassment and discrimination within the workplace.
As one of the first studies on this noteworthy topic, there are strengths to our research. Yet there remain a number of limitations, as well. For example, we highlight cyber aggression that attacks women of color, but we do not mean to imply that other groups of individuals are free from online aggression; they are not. On a related matter, we cannot conclude that minority women are significantly more likely to be targets than other groups, such as white women, men of color, or sexual minorities. Our sample is nonrandom, and it derives from the very limited portion of tweets that Twitter releases for public downloading. As a result, the networks, and the illustrations we analyze, may be missing actors and interchanges of tweets. Moreover, the interpretation of message content and demographic characteristics on Twitter can be quite subjective and prone to error (due to researcher coding or intentional misinformation supplied by the user). Finally, additional research is needed to examine the intersectional themes and network patterns we identify to examine the extent to which they, and other trends, occur in additional samples of cyber harassment.
In sum, aggression in online media remains a serious problem and one that can embroil women of color in its nastiness. At the same time, not all communication of this kind is negative in nature. As depicted in one of our illustrations, people can rally to support victims of harassment and this type of support also can “go viral.” Moreover, Twitter and other social media giants are attempting to institute measures to make it more difficult for people to use their services abusively (e.g., facilitate the reporting of harmful messages). One of the goals in our work is to raise the awareness of this social problem, and encourage both individual Internet users and social media platforms to explore additional ways of reducing harmful, online aggression.
Bailey, 2010Bailey, M. (2010). They aren’t talking about me…. The Crunk Feminist Collective, March 14. Retrieved from http://www.crunkfeministcollective.com/2010/03/14/they-arent-talking-about-me/. Accessed on February 22, 2018.
Beauboeuf-Lafontant, 2009Beauboeuf-Lafontant, T. (2009). Behind the mask of the strong Black woman: Voice and the embodiment of a costly performance. Philadelphia, PA: Temple University Press.
Bellmore, Calvin, Xu, & Zhu, 2015Bellmore, A., Calvin, A. J., Xu, J. M. & Zhu, X. (2015). The five w’s of “bullying” on twitter: Who, what, why, where, and when. Computers in Human Behavior, 44, 305–314.
Bray, 2017Bray, C. (2017, October 26). Twitter says it overstated monthly-user figures for 3 years. The New York Times. Retrieved from https://www.nytimes.com/2017/10/26/technology/twitter-q3-earnings-trolls.html.
Burgess, Stermer, & Burgess, 2007Burgess, M. C. R., Stermer, S. P. & Burgess, S. R. (2007). Sex, lies, and video games: The portrayal of male and female characters on video game covers. Sex Roles, 57(5–6), 419–433. https://doi.org/10.1007/s11199-007-9250-0
Cedar, 2008Cedar. (2008). ‘Tranny’ and Subversivism: Re-Reclaiming ‘Tranny’ (or Not) Part 1. Taking Up Too Much Space. Retrieved from https://takesupspace.wordpress.com/2008/11/10/tranny-and-subversivism-re-reclaiming-tranny-or-not-part-1/. Accessed on February 22, 2018.
Celious, 2002Celious, A. K. (2002). How ‘bitch’ became a good thing – or, at least not that bad. Perspectives, 8(2), 90.
Collins, 2000Collins, P. H. (2000). Black feminist thought: Knowledge, consciousness, and the politics of empowerment. New York, NY: Routledge.
Conley, & Ramsey, 2011Conley, T. D. & Ramsey, L. R. (2011). Killing us softly? Investigating portrayals of women and men in contemporary magazine advertisements. Psychology of Women Quarterly, 35(3), 469–478.
Crenshaw, 1991Crenshaw, K. (1991). Mapping the margins: Intersectionality, identity politics, and violence against women of color. Stanford Law Review, 1241–1299.
Croom, 2011Croom, A. M. (2011). Slurs. Language Sciences, 33(3), 343–358.
Dill, & Thill, 2007Dill, K. E. & Thill, K. P. (2007). Video game characters and the socialization of gender roles: Young people’s perceptions mirror sexist media depictions. Sex Roles, 57(11–12), 851–864.
Downs, & Smith, 2010Downs, E. & Smith, S. L. (2010). Keeping abreast of hypersexuality: A video game character content analysis. Sex Roles, 62(11–12), 721–733. https://doi.org/10.1007/s11199-009-9637-1
Fairclough, 2008Fairclough, K. (2008). Fame is a losing game: Celebrity gossip blogging, bitch culture and postfeminism. Genders, 48, 19.
Faris, & Felmlee, 2014Faris, R. & Felmlee, D. (2014). Casualties of social combat: School networks of peer victimization and their consequences. American Sociological Review, 79(2), 228–257.
Felmlee, & Faris, 2016Felmlee, D. & Faris, R. (2016). Toxic ties: Networks of friendship, dating, and cyber victimization. Social Psychology Quarterly, 79(3), 243–262.
Felmlee, Inara Rodis, & Zhang, 2018Felmlee, D., Inara Rodis, P. & Zhang, A. (2018). “Sl*ts, wh*res, and bi*ches: Targeting femininity in online aggression. Paper presented at the Annual Meetings of the American Sociological Association, Philadelphia, PA.
Gibbs, 2017Gibbs, C. (2017). Instagram is the worst social network for cyber bullying: Study. Retrieved from http://www.nydailynews.com/life-style/instagram-worst-social-network-cyber-bullying-study-article-1.3339477. Accessed on February 01, 2018.
Goffman, 1979Goffman, E. (1979). Gender advertisements. New York, NY: Harper & Row.
Harris-Perry, 2011Harris-Perry, M. (2011). Sister citizen: Shame, stereotypes, and Black women in America. New Haven, CT: Yale University Press. ISBN 9780300165548.
Hinduja, & Patchin, 2010Hinduja, S. & Patchin, J. W. (2010). Bullying, cyberbullying, and suicide. Archives of suicide research, 14(3), 206–221.
Kiang, Perreira, & Fuligni, 2011Kiang, L., Perreira, K. M. & Fuligni, A. J. (2011). Ethnic label use in adolescents from traditional and non-traditional immigrant communities. Journal of Youth and Adolescence, 40(6), 719–729.
Ki-Young, & Joo, 2005Ki-Young, L. & Joo, S.-H. (2005). The portrayal of Asian Americans in mainstream magazine ads: An update. Journalism and Mass Communication Quarterly; Thousand Oaks, 82(3), 654–671.
Lee, Wong, & Alvarez, 2009Lee, S. J., Wong, N.-W. A. & Alvarez, A. N. (2009). The model minority and the perpetual foreigner: Stereotypes of Asian Americans. In N. Tewari & A. N. Alvarez (Eds.), Asian American psychology: Current perspectives (pp. 69–84). New York, NY: Routledge/Taylor & Francis Group.
MacKinnon, 2013MacKinnon, C. A. (2013). Intersectionality as method: A note. Signs: Journal of Women in Culture and Society, 38(4), 1019–1030.
McCabe, Fairchild, Grauerholz, Pescosolido, & Tope, 2011McCabe, J., Fairchild, E., Grauerholz, L., Pescosolido, B. A. & Tope, D. (2011). Gender in twentieth-century children’s books: Patterns of disparity in titles and central characters. Gender & Society, 25(2), 197–226.
Merskin, 2007Merskin, D. (2007). Three faces of Eva: Perpetuation of the hot-Latina stereotype in desperate housewives. Howard Journal of Communications, 18(2), 133–151. https://doi.org/10.1080/10646170701309890
Miller, & Summers, 2007Miller, M. K. & Summers, A. (2007). Gender differences in video game characters’ roles, appearances, and attire as portrayed in video game magazines. Sex Roles, 57(9–10), 733–742. https://doi.org/10.1007/s11199-007-9307-0
Mok, 1998Mok, T. A. (1998). Getting the message: Media images and stereotypes and their effect on Asian Americans. Cultural Diversity and Mental Health, 4(3), 185–202. https://doi.org/http://dx.doi.org.ezaccess.libraries.psu.edu/10.1037/1099-9809.4.3.185
Nansel, Overpeck, Pilla, Ruan, Simons-Morton, & Scheidt, 2001Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., Simons-Morton, B. & Scheidt, P. (2001). Bullying behaviors among US youth: Prevalence and association with psychosocial adjustment. JAMA, 285(16), 2094–2100.
Pew Research Center, 2017Pew Research Center. (2017). Online harassment 2017. Retrieved from http://www.pewinternet.org/.
Rich, 1980Rich, A. (1980). Compulsory heterosexuality and lesbian existence. Signs: Journal of Women in Culture and Society, 5(4), 631–660.
Ridgeway, 2011Ridgeway, C. L. (2011). Framed by gender: How gender inequality persists in the modern world. New York, NY: Oxford University Press.
Rost, Stahel, & Frey, 2016Rost, K., Stahel, L. & Frey, B. S. (2016). Digital social norm enforcement: Online firestorms in social media. PLoS One, 11(6), e0155923.
Salmivalli, 1999Salmivalli, C. (1999). Participant role approach to school bullying: Implications for interventions. Journal of Adolescence, 22(4), 453–459.
Shields, 2008Shields, S. A. (2008). Gender: An intersectionality perspective. Sex Roles, 59(5–6), 301–311.
Sijtsema, Veenstra, Lindenberg, & Salmivalli, 2009Sijtsema, J. J., Veenstra, R., Lindenberg, S. & Salmivalli, C. (2009). Empirical test of bullies’ status goals: Assessing direct goals, aggression, and prestige. Aggressive Behavior, 35(1), 57–67.
Smith, Milic-Frayling, Shneiderman, Mendes Rodrigues, Leskovec, & Dunne, 2010Smith, M., Milic-Frayling, N., Shneiderman, B., Mendes Rodrigues, E., Leskovec, J. & Dunne, C. (2010). NodeXL: A free and open network overview, discovery and exploration add-in for Excel 2007/2010. Retrieved from http://nodexl.codeplex.com/ from the Social Media Research Foundation, http://www.smrfoundation.org.
Stankiewicz, & Rosselli, 2008Stankiewicz, J. M. & Rosselli, F. (2008). Women as sex objects and victims in print advertisements. Sex Roles, 58(7–8), 579–589.
Sterner, & Felmlee, 2017Sterner, G. & Felmlee, D. (2017). The social networks of cyberbullying on Twitter. International Journal of Technoethics (IJT), 8(2), 1–15.
Taylor, Lee, & Stern, 1995Taylor, C. R., Lee, J. Y. & Stern, B. B. (1995). Portrayals of African, Hispanic, and Asian Americans in magazine advertising. American Behavioral Scientist, 38(4), 608–621. https://doi.org/10.1177/0002764295038004010
U.S. Census Bureau, 2016U.S. Census Bureau. (2016). U.S. Census Bureau QuickFacts: United States. Retrieved from https://www.census.gov/quickfacts/fact/table/US/PST045217
Westmoreland, 2001Westmoreland, K. (2001). “Bitch” and Lilith fair: Resisting anger, celebrating contradictions. Popular Music and Society, 25(1–2), 205–220. https://doi.org/10.1080/03007760108591793
Willard, 2007Willard, N. E. (2007). The authority and responsibility of school officials in responding to cyberbullying. Journal of Adolescent Health, 41(6), S64–S65.
Women’s Media Center, 2017Women’s Media Center. (2017). The status of women in the U.S. media 2017. Annual Report. Women’s Media Center. Retrieved from http://www.womensmediacenter.com/assets/site/reports/the-status-of-women-in-u.s.-media-2017/10c550d19ef9f3688f_mlbres2jd.pdf.
Xu, Jun, Zhu, & Bellmore, 2012Xu, J. M., Jun, K. S., Zhu, X. & Bellmore, A. (2012). Learning from bullying traces in social media. In Proceedings of the 2012 conference of the North American chapter of the association for computational linguistics: Human language technologies (pp. 656–666). Association for Computational Linguistics.
Yang, Gibson, Lueke, Huesmann, & Bushman, 2014Yang, G. S., Gibson, B., Lueke, A. K., Huesmann, L. R. & Bushman, B. J. (2014). Effects of avatar race in violent video games on racial attitudes and aggression. Social Psychological and Personality Science, 5(6), 698–704.
The authors would like to thank Amy Zhang, Jordan Lawson, Robert Zuchowski, and members of the Twitter Research Group for their helpful feedback. Partial support for this research was provided by the National Science Foundation under Grant No. 1818497 and by the Social Science Research Institute.
- Women’s Places: An Introduction to Gender and the Media
- Part I Agency Affirming Places
- Chapter 1 War, Culture, and Agency Among Sahrawi Women Refugees: A Photo-Essay
- Chapter 2 From “Old Boy” To “Gender Progressive”: The Shifting Gender Story of Funeral Work in Trade Journal Publications
- Chapter 3 “Punk Fairytale”: Popular Music, Media, and the (Re)Production of Gender
- Chapter 4 “Trappin’ Ain’t Shit to Me”: How Undergraduate Students Construct Meaning Around Race, Gender, and Sexuality Within Hip-Hop
- Part II Overtly Hostile or Agency-Denying Places
- Chapter 5 Truth, Justice, Boobs: Gender in Comic Book Culture
- Chapter 6 What a B!tch!: Cyber Aggression Toward Women of Color
- Chapter 7 Mainstreaming Gender, Endangered, Ungendered? Analysis of Media Reports of the 2012 Case of Rape in India
- Chapter 8 Images of Trafficked Women: A Case Study ofMedia and Social Science Discourse in Moldova, 2003–2008
- Part III Covertly Negating Places
- Chapter 9 Mortality Salience, Terror Management, and Hollywood Film: Theorizing on the Absence of Anorexia as a Subject in US Mainstream Movies
- Chapter 10 Who is the American Girl? Analyzing Difference in American Girl Advice Books
- Chapter 11 Gender and Critical Evaluation in Popular Music