Abstract
Video game livestreams have become a significant component of gaming culture, offering an online platform where thousands gather to connect and interact with one another. However, like video game culture in general, previous research has highlighted a large presence of misogyny on livestreaming sites, particularly on Twitch.tv. This research explored instances of gendered harassment on Twitch.tv by examining differences in the chat environment curated by male and female video game livestreamers. To achieve this, the study used a qualitative approach that included a sample of seven male and seven female video game livestreams, selected across seven video game titles. Thematic analysis suggested that there was an observable difference between the male and female video game livestream experience, with female livestreamers experiencing greater criticism regarding their physical appearance and gameplay strategy. Interestingly, the most glaring examples of prejudice were not directed towards female livestreamers but were discovered in the male livestreamers’ chat, suggesting that this space provides an environment whereby men can establish their masculine dominance during the livestream and within the associated chat channel. These findings underscore the significance of examining how misogyny and prejudice manifest in emerging digital landscapes, such as Twitch.tv, which can subsequently aid in platform policy design to create a welcoming digital space for all.
Introduction
Since the 1980s, video games and their associated culture have rapidly gained popularity (Marston and Duro, 2020; Shaw, 2010). However, the harassment, sexualisation, and exclusion of women within this space has remained consistent (Dill and Thill, 2007; Rogstad, 2021; Ruotsalainen, 2022). This is often argued to be because of the historically male-dominated curation and consumption of video game culture, which has led to hostile reactions towards women attempting to enter this space (Cardillo, 2022; Schott and Horrell, 2000). A recent industry development is the rise of video game livestreaming, where individuals broadcast live video of themselves playing video games to viewers over the internet (Gelūnas, 2022; Leonhardt and Overå, 2021). Central to this new development is Twitch.tv (Twitch), a livestreaming platform that regularly attracts over one-hundred million viewers each month (Woodcock and Johnson, 2019). Although this development has been celebrated within gaming culture, previous research has consistently shown evidence of blatant racism, sexism, and homophobia occurring on the platform (Mihailova, 2020; Powell and Williams-Johnson, 2023; Ruberg, 2021). Therefore, it is reasonable to assume that the harassment female game players previously experienced has also become a prevalent aspect of the female gamer experience on Twitch.
This research investigates the ways in which gendered harassment occurs during online video game livestreams on Twitch. Prior research focused on gendered harassment on the platform has utilised a mix of qualitative and quantitative methods, but few focus on the live interactive dynamics of harassment during extended livestream observations. This study adopts a qualitative research approach, including the use of participant observation, to explore the experiences of female streamers and examine how streamers, moderators, and chat participants respond to and negotiate harassment in real time. Additionally, these findings illustrate how female streamers often experience more subtle forms of harassment, including frequent sexual comments and increased scrutiny regarding their appearance. It was also found that female streamers often receive overly critical messages regarding their gameplay and strategy, a trend that appears less prevalent in male livestream chat environments. However, while women were more likely to experience subtle forms of harassment and more criticism regarding their gameplay choices and strategy, this study also found that the most overt displays of harassment occurred during the male livestreams. This unexpected finding demonstrates that video game livestreaming is an environment in which heteronormative masculinity is reinforced through overt displays of heterosexuality and dominance to gain approval from other men. Furthermore, in relation to the techno-cultural affordances of Twitch, this research contributes significant findings towards how moderation practices, including the use of both automated and human moderation features, substantially affect the environment of the associated livestream chat channel. This, consequently, appears to influence how viewers and streamers respond to harassment during their livestream.
This article starts with a review of existing literature, focussing on masculinity and video games, femininity and video games, and harassment on Twitch. It will then detail the methodological approach, which is qualitative and employs non-participant observation to gather research material from 14 livestreams broadcast during June and July 2023. This study also draws on the features of computer-mediated discourse analysis, which supported the analysis of online communication, including the technical features that impact how communication channels operate on Twitch. The analysis adopted a thematic approach and identified five key themes: ‘Gender bias visualised through critical feedback’, ‘Gender variation in streamer appraisal’, ‘Hostility within male IRC channels’, ‘Differences across game genre’, and ‘Forms of community protection’, which included the subthemes ‘automatic moderation’ and ‘human moderation’. The article concludes by highlighting the significance of these findings and how they could be used to inform platform moderation policies and contribute to the general understanding of gendered interaction between streamers and viewers, and how technological infrastructure shapes online interaction.
Background
Masculinity and video games
Video games have historically existed within the male gender domain, a connection reinforced by their link to technology and computer science, fields traditionally dominated by men (Kuss and Griffiths, 2012; Lie, 1995; Rogstad, 2021; Schott and Horrell, 2000). However, before the 1970s, computer programming was considered a female profession. The term ‘computer’ was not associated with a machine, instead, it referred to women who wrote code and produced punch cards to operate the machines (Klinger and Svensson, 2021: 2077). The work itself, then referred to as coding rather than programming, a distinction that separated female coders from their male superiors who held the programmer title, was considered tedious, repetitive, and intellectually undemanding (Ensmenger, 2010: 14). Consequently, coding was viewed as a feminine activity that was tied to software and contrasted by the male computer engineer who worked with hardware (Ensmenger, 2010: 14). However, the demand for software began to grow, which led to the promise of higher salaries, and thus men began entering the field in greater numbers (Klinger and Svensson, 2021: 2078). Eventually, coding was reframed as highly complex, masculine work, emphasising mathematical skill and recasting coding as creative, active, and unpredictable (Ensmenger, 2010: 79). Thus, the image of coding became associated with the male artistic, anti-social, anti-authoritarian genius, and the hiring practices of the industry further reinforced and institutionalised the male subculture of computing (Ensmenger, 2010: 77).
As a result, women found themselves increasingly pushed to the margins of this space. Meanwhile, during this same period, the video game industry was beginning to take shape. However, early prototypes of video games were created and consumed mainly by men, as access to the computing power required to run and produce these games was restricted to advanced universities and government facilities, which were primarily accessible to men (Cote, 2015; Gelūnas, 2022). While progress has been made, this demographic imbalance has contributed to the exclusion of women from this space, both physically and virtually (Kuss and Griffiths, 2012). As studies show, male video game characters are more likely to occupy central plotlines and outnumber female characters at ratios ranging from 3:1 to over 5:1 (Blackburn and Scharrer, 2019). Additionally, while female characters are absent from 41% of games, when they are present, 28% are portrayed as sexual objects, and 21% experience extreme violence (Dietz, 1998). Such examples highlight the deep-rooted gender bias in video game culture.
Previous research has also demonstrated how masculine identities are promoted within video games and the surrounding culture, highlighting two distinct identities: ‘geek masculinity’ and ‘hegemonic masculinity’ (Adams et al., 2025; Crawford, 2005; Rogstad, 2021; Trinh, 2013). ‘Geek masculinity’ is commonly associated with video game culture and is an identity celebrating intellect, niche interests, and video games, fostering a fraternity of like-minded men (Salter and Blodgett, 2012). Likewise, while ‘hegemonic masculinity’ may appear distinct from ‘geek masculinity’, it is argued to be promoted through video games’ techno-cultural elements, where male characters value violence, strength, and the subordination of others (Rogstad, 2021). Furthermore, Kimmel (2008) argues that video games are central to ‘Guyland’, a contemporary life stage where men escape the expectations and pressures of family, girlfriends, jobs, and children (Kimmel, 2008: 155). For these men, video games present an ideal escape, as they are immersed within an alternative reality that can provide them with a sense of power that is potentially lacking in their daily lives and allows them to act without the oversight of authority figures and free from societal pressures for equality (Kimmel, 2008: 150). Consequently, video games and their surrounding culture create a space where masculine dominance is celebrated and unchallenged, enabling the expression of aggression and the virtual domination of others. This research aims to explore how masculinity is promoted through video game livestreaming on Twitch. Prior research has focused on how masculine identities operate within video game culture, however, fewer studies have examined this within the context of livestreaming platforms, which have become a crucial element of the online video game community (Taylor, 2018). Additionally, this research hopes to contribute to academic understanding of online sexism and potentially inform strategies which could help to foster a more inclusive digital environment.
Femininity and video games
As discussed, video games and the surrounding culture have been strongly tied to masculine identities, creating challenges for women who, while not explicitly excluded, face multi-layered gendered boundaries that reinforce their marginalisation within this space (Drenten et al., 2022). One mechanism sustaining male dominance is the underrepresentation and sexualisation of female game characters (Dill and Thill, 2007; Downs and Smith, 2010). Female characters often embody characteristics associated with emphasised femininity, positioning women as serving to support and sexually gratify men, while lacking intellectual thought, personal interests, or ambition (Connell, 1987; Dill and Thill, 2007; Rogstad, 2021). For instance, in Overwatch, female characters often occupy support roles, such as healers or medics, who aid male protagonists and whose narratives are defined by the men around them (Apperly, 2022; Valisalo and Ruotsalainen, 2022). Moreover, female characters are often hypersexualised, designed with unrealistic body proportions and revealing clothing. A case in point is the character ‘Ivy’ from Soul Calibur, who is visualised as a sexual dominatrix, characterised by proportionally inaccurate breasts, revealing clothing, and armed with a whip-like weapon (Biscop et al., 2019; Dill and Thill, 2007: 723). Although Ivy’s character alone may be unproblematic, when situated within a broader digital culture that hypersexualises women, the character becomes a stereotype (Biscop et al., 2019). This presentation of women is also argued to reflect idealised feminine beauty standards aimed at a predominantly heterosexual male audience, which consequently can confine women in gaming to the perception of being sexual objects (Charles, 2016; Drenten et al., 2022).
In real life, female gamers also face backlash, sexualisation, and objectification from their male counterparts (Blackburn and Scharrer, 2019; Drenten et al., 2022). A significant barrier impacting female gamers is their struggle to identify as ‘real’ gamers. The video game community often categorises games as either ‘casual’ or ‘hardcore’. Casual games, which are viewed as less legitimate because they are more forgiving, easier to learn, and lack violent or sexual content, are a genre dominated by female players (Blackburn and Scharrer, 2019: 313; Paaßen et al., 2016: 425; Phan et al., 2012: 1500). Therefore, while women frequently play video games, their preferred titles often hinder their ability to be considered ‘real’ gamers. Furthermore, when women do engage with ‘hardcore’ games, they often experience backlash from male gamers, receive less skill recognition, and greater criticism, reinforcing the perceived incompatibility of feminine identities and gamer identities (Kivijärvi and Katila, 2021; Olsson, 2018; Paaßen et al., 2016). Additionally, female gamers are often routinely sexualised, with their interest in gaming framed as a strategy to gain male sexual attention (Drenten et al., 2022: 10). This leads to increased antagonism against female players, especially in livestreaming or live-chat environments, where women face significantly higher levels of sexualisation and hostility in comparison to their male counterparts (Anciones-Anguita and Checa-Romero, 2024). Such research highlights how women cannot simply exist within video game culture; instead, they are regularly excluded, harassed, and objectified, all of which occurs in an environment where such behaviour is normalised (Blackburn and Scharrer, 2019; Charles, 2016). Few research projects have examined how this hostility manifests during extended livestreams and how streamers and viewers navigate such harassment. This research addresses that gap by examining the treatment of women on Twitch to better understand the gender dynamics within modern video game culture.
Harassment on Twitch
Livestreaming has become a celebrated aspect of gaming culture, with Twitch emerging as the central platform, and this has aided in transforming gaming from a personal experience into a source of public entertainment that encourages community interaction (Gelūnas, 2022; Leonhardt and Overå, 2021; Taylor, 2018; Woodcock and Johnson, 2019). A livestream typically combines gameplay footage, live video of the streamer, and real-time chat functions known as the Internet Relay Chat (IRC) channel onto one screen. Viewers can communicate with livestreamers and other viewers using the IRC channel, often posting text or ‘emotes’, which are small images that have a shared meaning for those within the community (Siitonen and Ruotsalainen, 2022: 93). Emotes are particularly useful in fast-paced chat environments such as Twitch, as they help users concisely express their opinions (Kim et al., 2022). While emotes can infer tone or sarcasm, they have also been used to convey homophobic, racist, or sexist sentiments (Kim et al., 2022). Although Twitch has platform policies intended to prohibit harassment and hateful content, streamers are also encouraged to moderate their IRC channels through automated systems and by appointing human moderators (Twitch, 2023a). Automated moderation tools can be configured to a streamer’s preference, enabling them to filter banned words and spam (Twitch, 2023b). Human moderators can work alongside these tools and manually oversee the IRC channel, removing flagged comments, hateful content, or spam, as well as banning problematic viewers (Mihailova, 2020; Twitch, 2023b). Additionally, human moderators can also enforce community boundaries, helping to cultivate a positive chat environment by addressing inappropriate behaviour (Mihailova, 2020). However, this system is not entirely effective, for instance, automatic moderation often only detects text comments, not emotes, which allows users to bypass filters (Drenten et al., 2022; Kim et al., 2022). Consequently, users are often still able to engage in homophobic, racist, or misogynistic behaviour within IRC channels.
Notably, as livestreaming has gained popularity, the harassment historically faced by female gamers has also surfaced on these platforms (Wolff and Shen, 2022). Female streamers face up to 11 times more sexual harassment compared to male streamers during livestreams (Ruvalcaba et al., 2018: 307). Likewise, this harassment also extends to the derogatory terms often used for female streamers, such as ‘cam-girl’ or ‘boobie streamer’, common terminology that acts to belittle their content and success (Wolff and Shen, 2022: 865). Therefore, although livestreaming typically involves showcasing one’s appearance and personality to attract an audience and gain success, the connection between this practice and sex work is solely reserved for female streamers (Wolff and Shen, 2022: 865). Additionally, women on Twitch have also experienced stricter content policing, with their attire and self-presentation being more regulated when compared to male streamers (Zolides, 2021). Furthermore, such policing reveals a clear double standard, as popular games such as Grand Theft Auto V feature sexualised content that would lead to bans for female streamers if replicated, indicating that women’s self-presentation is more heavily regulated than sexual representations in games (Ruberg, 2021). Moreover, hate raids have become a prevalent issue on Twitch, increasing in frequency in spring 2021 (Adams et al., 2025). Hate raids involve a sudden influx of users entering a streamer’s chat and posting hostile and hateful language, symbols, and threats, often of a homophobic, racist, misogynistic, and transphobic nature. Consequently, such raids aim to disrupt a streamer’s rapport with the live audience and work to signal to the streamer that they do not belong within the Twitch community (Adams et al., 2025).
This study extends existing research on Twitch by examining the conditions that impact how gendered harassment occurs within live streaming environments. By comparatively analysing chat environments across male and female livestreams and across a range of game genres, this research provides a context-sensitive and in-depth account of gendered harassment. While highlighting the presence of gendered harassment, the study also aims to capture the temporal dynamics of harassment, identifying peak moments linked to gameplay events, streamer visibility, and audience engagement. In doing so, this study advances understandings of gendered harassment on Twitch as it is produced and negotiated through live, platform-specific interactions.
Method
Research design and framework
This study uses a qualitative approach to explore how gendered harassment manifests within Twitch livestreams and how platform features impact this. Qualitative methods were employed as they provide an in-depth approach for examining how gender identities shape community interaction dynamics on Twitch (Smilde and Hanson, 2018). Specifically, this study was informed by computer-mediated discourse analysis (Herring, 2004), which allowed for an examination of both the content of chat messages and the distinctive technical features of Twitch that structure interaction. Herring (2004) identifies two variables to be considered within computer-mediated discourse analysis. Firstly, ‘technical factors’, which include the technical features of communication channels, such as messaging format, whether messages are anonymous, the number of characters allowed in a single message, or whether they are publicly available. Secondly, ‘situational factors’, which help generate a greater textual understanding of how messages are received, such as the type of conversation, demographic characteristics of participants, the goal of interaction, and the tone of conversation (Herring, 2004: 21). This framework helped structure the current study by improving understanding of the social aspects of Twitch messages, while also enabling an investigation into whether specific technological features influence the prevalence of gendered harassment.
Sample and data collection
Table showing the stream length and view count of selected streamers.
Data analysis
Thematic analysis was adopted to explore the experiences of female livestreamers and the approach followed Braun and Clarke’s (2006) six-step framework alongside guidance from Nowell et al. (2017). Manual data analysis was selected over computer software as it allowed for an unstructured approach, enabling themes to emerge organically (Braun and Clarke, 2006; Lochmiller, 2021; Sullivan, 2003). Following a process of identifying and revising a series of themes and subthemes to effectively illustrate the lived experience of female livestreamers within a male-dominated space, this study identified five themes and two subthemes: ‘Gender bias visualised through critical feedback’, ‘Gender variation in streamer appraisal’, ‘Hostility within male IRC channels’, ‘Differences across game genre’, and ‘Forms of community protection’ which included the subthemes ‘Automatic moderation’ and ‘Human moderation’.
Ethical considerations
There is a growing debate around how traditional ethical principles apply to online research, as existing ethical codes often do not reflect the realities of online interactions (Burles and Bally, 2018; Kozinets, 2010). Obtaining informed consent from anonymous users is particularly challenging; previous Twitch studies have found it impossible to locate and contact thousands of partly anonymous users (Joden and Strandell, 2021). Likewise, in this study, the constant flow of anonymous accounts entering and leaving the chat made it impossible to secure individual consent. Additionally, seeking the contact details of previously anonymous users without their permission could be seen as more intrusive than including their anonymous online content (Burles and Bally, 2018). As informed consent was not obtained, ensuring confidentiality was a priority. Streamers’ names were replaced with pseudonyms, and all chat usernames were anonymised. The issue of confidentiality was complicated as direct quotes could potentially be located through search engines, log files, or user profiles (Roberts, 2015). Therefore, this research employed several strategies when including direct quotations, such as ensuring that no usernames were attached to quotes and using aggregated quotations, whereby minor changes are made to the original quotations. This way, the intended meaning of the quote is preserved but is untraceable through search engines (Roberts, 2015).
Limitations
A key methodological challenge in this study was the anonymous nature of Twitch. While prior research suggests that most Twitch users are male, white, and under 34 (Pollack et al., 2021; Twitch, 2023c), this research was unable to obtain first-hand data to confirm this. Therefore, future research could include a survey to gather demographic information about Twitch users. Additionally, due to time and language constraints, this research focused on English-speaking streamers, limiting the analysis of gendered harassment to English-speaking Western countries. Furthermore, the chosen sample size aimed to balance depth and manageability, as the volume and speed of chat messages within IRC channels, especially those lasting over 4 hours, were considerably high. Consequently, while the small sample size may limit the representativeness of the study, increasing this would have exceeded the practical limits of the current study.
Findings and discussion
Gender bias visualised through critical feedback
This theme highlights negative comments regarding streamers’ gameplay, including direct insults, spamming, trolling, and critiques of their gaming decisions. Analysis of chat logs revealed a consistent pattern that, although male and female streamers received negative comments relating to gameplay, female streamers experienced more. The Overwatch livestreams were the only ones in which the male streamer received more negative comments than the female streamer. However, this anomaly could be attributed to the stricter moderation during the female Overwatch stream. Analysis revealed two key periods in which negative comments were most observed across the livestreams. First, negative comments often emerged following a streamer’s in-game mistake, death, or defeat, with this being more noticeable during the Fortnite, Apex Legends, and VALORANT livestreams. Second, negative comments also occurred during strategic portions of gameplay. This was most evident during the Diablo IV and League of Legends livestreams. Focussing on the former, negative comments of this kind mainly appeared as spam, often consisting of one-word phrases such as ‘F’, ‘noob’, and ‘skill issues’, all of which are negative terms implying that the stream is ‘fucked’ (f), that the streamer is new to the game (noob) and lacks gaming skill (skill issues). Moreover, female livestreamers appeared more likely to receive these comments. For example, the chat reaction to the female Fortnite streamers’ loss was considerably harsher than the male equivalent, who frequently received positive terms such as ‘NT’ (nice try) and ‘GG’ (good game). While this does not suggest all chat responses to female streamers are negative, it does suggest that viewers are more outwardly empathetic towards the male streamer.
Additionally, viewers appeared quicker to insult the female streamers’ gaming ability, often insinuating a lower skill level: FemaleDiablo_1: ‘I’m going to go out on a limb and say this is your first time playing diablo’ FemaleLeague_1: ‘Left click baby’ FemaleLeague_2: ‘Can’t even defend herself’
These comments reveal a noticeable pattern in which female livestreamers are perceived as less competent, which supports prior research that suggests that a feminine gender identity and a ‘gamer’ identity are perceived as incompatible (Kivijärvi and Katila, 2021; Olsson, 2018; Paaßen et al., 2016). As discussed, the ‘gamer’ identity is frequently framed within a gendered hierarchy of play (Kivijärvi and Katila, 2021). A ‘real’ gamer is depicted as someone who embodies a masculine identity, dedicates hours playing ‘hardcore games’, and asserts their superiority over ‘casual gamers’ (Kivijärvi and Katila, 2021; Ruotsalainen, 2022). Conversely, femininity is often devalued because of its association with ‘casual games’ (Blackburn and Scharrer, 2019; Paaßen et al., 2016). This distinction between ‘casual’ and ‘hardcore’ games often prevents women from being able to inhabit a ‘gamer’ identity, and when women do attempt to engage in ‘hardcore’ games, they are frequently dismissed by their male counterparts (Paaßen et al., 2016). Consequently, whenever female gamers attempt to partake in ‘hardcore’ games, such as Diablo IV and League of Legends, they are viewed as having lower levels of competence, and this belief can prevent female gamers’ skills from being rewarded with the same recognition as male gamers. The current findings also add an important dimension to previous literature, highlighting how bias against female livestreamers is not uniform across all video game genres. As the research material suggests, the more strategically complex a game is, the more pronounced the dismissal of female competence appears to be, implying that the gap between femininity and a ‘real’ gamer identity is more pronounced whenever strategic mastery is the most visible and valued skill. This was further supported by comments such as ‘You’re the best female apex player’ (FemaleApex_1), which illustrates the conflict between the feminine identity and gamer identity. Rather than being a straightforward compliment, the comment appears to suggest conditional recognition, acknowledging the skill within a gendered bracket while also implying that such skill would not compete with the male players of the same game. This suggests that within the gaming community, there appears to be an unwillingness to discuss female players in the same terms as their male counterparts, and thus female streamers cannot be recognised as equal to males.
In summary, this theme illustrates the limited space afforded to female gamers within gaming culture. Importantly, the findings of the current study confirm the existence of gender bias previously identified by researchers, but they also demonstrate how and where this bias operates. Gender bias appears to escalate at moments of visible failure, or strategic gameplay, and such bias against female video game livestreamers manifests not only through hostile comments, but through the lack of empathy that is seemingly awarded to female players in comparison to their male counterparts. This could be attributed to the male-dominated environment of video game culture, which can lead to antagonistic reactions from male players whenever women attempt to participate. This, combined with the perception that female gamers are inherently less competent and skilled, could explain the higher proportion of negative comments surrounding the female streamer’s gameplay.
Gender variation in streamer appraisal
This theme explored gender differences in positive comments received by male and female streamers and found that both genders received a large amount of praise, for instance: FemaleOverwatch_1: ‘I just want to say, thanks for inspiring me to play this’ FemaleFortnite_1: ‘UR MY IDOL!!!’ MaleLeague_1: ‘I respect him so much’ MaleVALORANT_1: ‘[Streamer] is being a literal God going through this stuff’
These comments reflect how viewers perceive viewer/streamer relationships, with viewers regularly expressing their admiration, respect, and adoration for their preferred streamer. This aligns with prior research suggesting that Twitch fosters ‘one-and-a-half’ way parasocial relationships through the technical features that enable communication between streamers and the associated chat community (Kowert and Daniel, 2021). Furthermore, Twitch users appear to prefer livestream content over traditional media, spending on average 106 minutes daily on Twitch (Leith, 2021). Streamers also broadcast as ‘themselves’, which can help generate real-feeling relationships between the viewers and the streamer (Tran, 2022). Consequently, this extended interaction, coupled with streamers sharing personal information about themselves to build a community, can help foster a perceived familiarity between the viewer and the streamer (Zhou and Farzan, 2021). Thus, this can help explain the comments expressing admiration, appreciation, and respect for the streamer. These sorts of comments were common across both male and female streams, indicating this type of interaction holds no gender distinctions.
However, a particularly valuable finding of this study was that female streamers consistently received more comments regarding their appearance. While such comments were often positive remarks regarding the female streamer’s hair or makeup, others focused on the streamer’s facial features and weight. Crucially, this pattern was observed across all female livestreams regardless of genre, suggesting that a woman’s most complimented and valued aspect within this space is their appearance. Previous studies have consistently linked the presence of female Twitch streamers to a higher frequency of appearance-based comments, with female streamers receiving more body-focused messages compared to male streamers, whose comments are primarily related to gameplay (Nakandala et al., 2017). Similarly, Ruvalcaba et al. (2018) found that when examining positive comments directed at both male and female Twitch streamers, 37% of positive comments directed towards women were based on appearance compared to only 7.5% for men. The current findings also support this, suggesting that while both genders receive positive comments and feedback, for female streamers, this praise is often also linked to their appearance and body, regardless of gaming skill demonstrated or video game genre.
Unsurprisingly, alongside comments regarding the female streamer’s appearance, there were also numerous examples of sexual comments directed towards female streamers: FemaleLeague_4: ‘What a nice ass’ FemaleDiablo_5: ‘You got to sell some thongs’ FemaleApex_3: ‘Frick I wanna play apex so bad when you’re being filthy like that’ FemaleApex_4: ‘I’d gagg her’
These comments highlight a significant trend that suggests statements directed towards female streamers tend to be more sexual in nature. Importantly, these comments were posted on a very public platform, visible to thousands of viewers as well as the streamer. As the commenters are openly engaging in this behaviour without fear of repercussions, it suggests that such messages have become a normalised aspect of interacting with female streamers. This also reflects the broader pattern of the sexualisation of female bodies within video game culture. For example, video game culture has historically sexualised and objectified female bodies within this sphere, with female video game characters being designed to appeal to heterosexual male desires (Blackburn and Scharrer, 2019; Charles, 2016; Dill and Thill, 2007). Consequently, female characters have been historically portrayed as sexual objects, a sentiment that has been normalised within the community (Blackburn and Scharrer, 2019; Drenten et al., 2022). However, this study’s findings extend this, suggesting that this normalisation has migrated seamlessly from virtual characters to real women on Twitch.
Taken together, the findings within this theme reveal two systems of appraisal which are seemingly divided across gendered lines. Male streamers are primarily praised for their actions, skill, strategy and gameplay, whereas female streamers are frequently praised for their appearance, with this often also including sexual comments. This distinction is important as it again reflects how female streamers are often reduced to their physical attributes and sexual allure rather than the quality of their gaming performance.
Hostility within male IRC channels
This theme explores one of the most unexpected findings of this study: that most hostile messages, particularly those containing slurs or rooted in prejudice, were found within the male IRC channels. The male Minecraft stream is notably exempt from this trend, as this research did not uncover such comments during analysis. Alternatively, games within the first-person shooter and online battle arena genre, including Overwatch, Apex Legends, Fortnite and VALORANT, often featured hostile comments. It is important to emphasise that most of these comments were not explicitly directed towards the male streamer but rather towards the streamer’s teammates, enemies, or the characters in the game. As will be discussed in greater depth, this distinction is significant as it implies that such comments could function as a cultural performance of masculine heteronormativity to other viewers: MaleValorant_2: ‘faggggg’ MaleFortnite_1: ‘IF HE GAY SHOOT HIM’ MaleValorant_3: ‘Hold your breath until you ace or you’re gay’
A key contribution of this study was the finding that homophobic language exclusively occurred within the male livestreams, something which has not been widely documented in prior livestreaming research. This discovery could be explained by how a stronger adherence to masculine norms predicts greater support for homophobic language or content. This finding has also been found true when examining homophobic content during massively multiplayer online games (Ballard and Welch, 2017; Salter and Blodgett, 2012). Additionally, within the framework of ‘Guyland’, homophobia or anti-gay rhetoric is understood as an extension of gender policing, where homosexuality is essentially viewed as anti-masculine (Kimmel, 2008). Likewise, the use of the word ‘gay’ is also associated with negative connotations, as it can serve as a synonym for ‘wrong’, ‘dumb’, or ‘stupid’ (Kimmel, 2008: 76). This is reflected within this study’s findings, as the context in which the word ‘gay’ is used locates the term as inherently negative. As such, whenever ‘gay’ is employed as an insult the comment becomes less of an attack on the individual’s sexual identity and more about challenging their masculine status, effectively suggesting that the individual is not man enough (Kimmel, 2008: 76). Notably, none of the male streamers within the current study openly identified as part of the LGBTQ+ community, suggesting that such comments were not directly intended as an attack on the streamer’s sexual identity. Furthermore, it is encouraging that these findings align with prior work suggesting that homophobic beliefs or statements within gaming environments are not exclusive to those with LGBTQ+ identities, with such comments almost equally experienced by heterosexual and cisgender players (Gillin and Signorella, 2023).
This reinforces this study’s view that homophobic language within these spaces is less about the interpersonal dynamic between the streamer and his viewer, nor is it an attack on the streamer’s sexuality; instead, such comments function as a tool of masculine norm enforcement. This interpretation is further supported by the absence of homophobic rhetoric during female Twitch livestreams. This is an important observation and contributes to academic knowledge regarding masculinity within livestreaming environments. Further research would be valuable in exploring how masculine norms are policed and enforced on Twitch, a space in which masculine dominance is a major feature, as this suggests that this form of hostility is specifically tied to the performance of masculinity in male-dominated spaces, rather than being a general feature of livestreaming chat culture.
Sexual and misogynistic remarks emerged as the second most prevalent type of comment within this theme, for instance: MaleFortnite_2: ‘Unreleated but I would destroy [female character’s] pussy’ MaleApex_2: ‘I blew my load to [female character]’ MaleValorant_4: ‘Once we have ai sex bots its over for these bitches’
As illustrated, male IRC channels often included sexual content when mentioning women. This could be explained by research suggesting that heterosexuality and expressing it are key to successful masculine performances (Duckworth and Trautner, 2019; Kimmel, 2008). Male heterosexual desire in this context seems uninterested in connection, love, or pleasure but rather in demonstrating masculinity, asserting dominance over women, and gaining positive social feedback from other men (Kimmel, 2008: 169). This was especially evident during the male Apex Legends and VALORANT livestreams when full-body images of female game characters appeared. Following this, messages reading ‘Smash’, a colloquial term that implies sexual attraction to the woman or female character on screen, flooded the IRC channel. Within this setting, it was assumed that these messages were acting as a declaration of heterosexuality, with users expressing their sexual desires towards these female characters and thus proving their masculine status to other male viewers. However, one consequence of the overt expression of heterosexual desire being central to the successful performance of masculinity is that women can become viewed as sexual objects that are used to enhance a man’s social image and attain a positive masculine status (Duckworth and Trautner, 2019). Thus, within this sphere, women are not perceived as fully rounded human beings but rather as sexual objects whose objectification is necessary for the successful performance of masculinity (Roberts et al., 2020).
Additionally, this is further reinforced by the derogatory terms Twitch viewers use to describe women, with chat users frequently referring to women as ‘bitches’, ‘cunts’, ‘hoes’, and ‘whores’. These terms are argued to perpetuate traditional misogynistic assumptions, whether this is intentional or not (Yildirim, 2022). For example, the use of the word ‘bitch’ has been argued to reinforce the idea that women are inherently different from men, while ‘hoe’ is a term that further sexualises and objectifies women, and ‘cunt’ is argued to imply that a woman has no redeemable features (Felmlee et al., 2019; Yildirim, 2022). The casual and normalised use of such language suggests that the devaluation and objectification of women within this space is routine rather than an anomaly. This finding aligns with previous research on video game culture, which shows that women’s presence is rarely perceived as neutral and is often contested and sexualised within male-dominated gaming spaces (Kuss and Griffiths, 2012). The present study extends this work by showing how this trend persists within video game livestreaming environments, suggesting that Twitch often functions as an extension of broader gaming culture, whereby exclusionary gender norms are reproduced and normalised.
Following computer-mediated discourse analysis, it is essential to attend not only to the content of chat messages but also to their tone and framing. While some comments were overtly misogynistic, homophobic, or hostile, others were presented as jokes. This distinction is important as prior research has highlighted the central role humour plays within video game and livestreaming culture (Johnson, 2024). For instance, Twitch is argued to provide a structural environment in which chaotic forms of humour, that are also frequently hostile and exclusionary, can thrive. Consequently, humour within this space often carries a darker undertone, often crossing boundaries into practices associated with trolling or bullying (Johnson, 2024). These dynamics were evident within the current study, as while homophobic and misogynistic comments were presented as playful discussion, these jokes were often framed at the expense of LGBTQ+ individuals, women, or the male streamer. Additionally, Twitch has been recognised as an environment in which boundary pushing and moral ambivalence is normalised, a pattern that is troubling given that the platform already operates as a contested arena shaped by unequal power relations between dominant and marginalised groups (Johnson, 2024). Consequently, those already marginalised within the space are more likely to experience such humour as alienating, delegitimising, or threatening. Crucially, this study demonstrates that framing these comments as humour does not neutralise their harm and instead provides a mechanism through which prejudice can be expressed while enabling the user to evade accountability. In this way, the findings reinforce and extend existing research by demonstrating how humour within livestreaming environments is mobilised to reproduce the marginalisation of other groups, including women and LGBTQ+ individuals, who are repeatedly positioned as targets of ridicule within spaces that structurally enable and protect such behaviour.
Forms of community protection
As demonstrated, streamers regularly face unpleasant, hostile, or prejudiced messages. However, streamers regularly devise strategies for such experiences, which for the most part involve the use of different forms of moderation. The two subthemes ‘automatic moderation’ and ‘human moderation’ explore how effective moderation strategies are in protecting streamers from harmful content.
Automatic moderation
This study identified automatic moderation across all 14 livestreams, with most streamers using third-party chat tools, such as ‘Nightbot’ and ‘StreamElements’. These were identified by green sword icons, which marked messages from these automated moderators, signalling their presence to other users within the chat. While the main purpose of these extensions is to effectively manage IRC channels by filtering chat messages and enforcing community rules and norms, they also serve secondary promotional purposes, for example, providing links for their Discord communities, YouTube channels, merchandise stores, or sponsored content advertisements. However, crucially, this research found that the sole reliance on automatic moderation does not provide meaningful protection against harassment. This is because while automatic moderation was employed by every streamer, sexualised comments, prejudice, harassment, spam and sexism were still present across most livestreams. This observation, however, aligns with previous findings that highlight the complexity of employing automatic moderation features on Twitch. This is because much of the responsibility of moderation enforcement still falls on the streamer (London et al., 2019). For example, tools like ‘StreamElements’ offer customisable chat moderation functions, allowing the streamer to define which words should be banned, specify the maximum character length in comments, regulate the amount of capital letters allowed, and set their preferred spam protection level, ranging from no protection to maximum security. Consequently, the level of effectiveness can vary depending on the streamer’s settings, and as this study found, if a streamer is not taking the necessary steps to prevent certain content from being posted, IRC channels can often foster a negative chat environment. Additionally, users were often able to bypass automated moderation by manipulating the formatting of certain words, including incorporating or alternating letters (i.e. ‘faggggg’ and ‘Cunk’). Consequently, this study directly challenges any assumption that automatic moderation constitutes sufficient safeguarding against harassment. Therefore, it becomes evident that additional layers of policing in the form of human moderators may be warranted to discern the subtleties of such messages and detect offensive comments that automated systems may fail to realise.
Human moderation
There was a notable difference in the range of human moderation across the livestreams. On average, streamers maintained an active team of three human moderators on top of automatic moderation tools. The female Overwatch streamer had the highest number of human moderators with seven, followed closely by the male and female Minecraft streamers, who each featured six. In contrast, the male VALORANT streamer refrained from including any human moderators and, throughout the livestream, repeatedly voiced his aversion to doing so. Interestingly, discernible patterns emerged when comparing the chat environments between those with fewer human moderators and those with more, confirming that an increased presence of human moderators helps cultivate a positive chat environment. Firstly, focussing on the female Overwatch IRC channel, almost no comments were observed criticising the streamer’s gameplay, there were very few instances of prejudice, and a scarcity of sexual comments. It was clear that a high number of human moderators played a pivotal role in shaping the community, as moderators were often seen engaging in active conversation with viewers, greeting viewers specifically by their username, recognising when a viewer had been absent for a while, and thanking viewers for watching. This consequently appeared to create a personalised and quite intimate viewing experience for chat users, enabling the creation of a positive online community.
This was further highlighted during one notable exchange within the female Diablo IV livestream. At one point in the stream, a user comments, ‘Were you sexually abused as a child?’ (FemaleDiablo_3) in reference to the female streamer. Immediately, the chat responds negatively towards this user, criticising them for this comment, before one of the chat moderators informs the viewers that this user has been removed from the chat and is thus unable to make future comments. This collective response to negative content, from not only the moderators but also other chat members, resonates with prior research. Studies have shown that this strategy is frequently observed within Twitch subcultures and has proven efficient in shaping and reinforcing community norms, while deterring others from engaging in inappropriate behaviour (London et al., 2019). Likewise, the ongoing provision of feedback through synchronous interactions in the Twitch IRC channels can play a crucial role in curtailing toxic behaviours, mitigating cyberbullying, and providing essential emotional support for victims of harassment (Zhou and Farzan, 2021). Additionally, the reinforcement of friendly interactions can help develop a sense of loyalty among viewers, meaning that regular viewers are more inclined to continue watching and uphold the communities’ associated values, while at the same time deterring perpetrators of harassment and bullying (Zhou and Farzan, 2021). Within this study, it thus became apparent that the presence of active human moderators helped to cultivate a positive chat environment. This finding represents a promising contribution, as it recognises inconsistencies in the effectiveness and deployment of automated Twitch moderation features, highlighting the importance of human moderation. As consistent, friendly interaction between viewers and moderators cultivated viewer loyalty, with regular viewers more likely to uphold community values and deter potential perpetrators. A finding that has practical implications for how streamers may wish to design their community environments.
Conversely, in the absence of moderators, particularly when combined with a streamer who endorses harassment and prejudice, this was shown to actively produce harmful community cultures. For instance, focussing on the male VALORANT livestream, there was an instance following homophobic and sexist comments whereby one chat user spoke out against this, criticising the users within the chat and the streamer for enabling them by not employing human moderators. The streamer responded, ‘I can’t control what people say’ and was seen engaging with chat members by laughing at and encouraging comments that were directed towards the user who condemned this behaviour. The endorsement of this behaviour from the streamer appeared to empower the chat and encourage them to persist in their actions, with users commenting ‘Chat would be aids with her as a moderator’ (MaleValorant_6) and ‘Fuck off from the stream if you’re so offended then’ (MaleValorant_7). This observation again aligns with previous studies, which have documented how Twitch livestreams can function as online ‘third places’, whereby communities engage and integrate around shared experiences and streamers play a pivotal role in the emergent culture of the community (Brewer et al., 2020). For example, as demonstrated, when a streamer is willing to implement the necessary tools to regulate IRC channels and encourage the creation of a positive community environment, this can significantly decrease levels of harassment, prejudice and prevent cyberbullying (London et al., 2019). Therefore, when a streamer, such as the male VALORANT streamer from this study, is reluctant to moderate their IRC channels, this can encourage the creation of toxic chat environments. Likewise, failing to establish clear rules and norms within a Twitch community can encourage outsiders to amplify negative behaviour. Previous research has demonstrated that newcomers who are unfamiliar with established community norms are more likely to engage in inappropriate behaviour and, therefore, to assist newcomers and prevent such behaviour, streamers often employ rule-setting and moderation strategies to deter such conduct (Zhou and Farzan, 2021). Thus, applying these findings to the current research, the lack of rule-setting and moderation strategies within the male VALORANT livestream, coupled with the streamer’s endorsement of negative behaviour, allowed newcomers entering this chat to observe inappropriate behaviour and believe that such conduct was not only acceptable but replicable.
Consequently, this finding contributes to academic understanding of how streamer behaviour and moderation practices actively shape the culture of Twitch communities. Firstly, this study identified how human moderators contribute beyond the simple enforcement of community guidelines, demonstrating how they engage in active conversations with viewers, greeting them by username, acknowledging their absence, and expressing gratitude on behalf of the streamer. These behaviours produced a distinctly personalised and intimate viewing experience, which helped to foster genuine community cohesion. This study also extends this finding by demonstrating how the opposite, that is, the absence of rules, human moderators, and the presence of a streamer who encourages hostility, can normalise and encourage harassment. Consequently, this has implications for platform moderation policies, suggesting that accountability for toxic IRC environments should not rest on automatic moderation, nor should it rest on streamers alone, instead suggesting that a more robust system of interventions may be warranted in livestreaming environments.
Differences across game genre
The decision to focus on a range of video game genres, spanning seven video game titles, enabled an exploration into whether gendered harassment varies by game genre. Interestingly, analysis demonstrated that there were indeed trends in differing levels of gendered harassment, profanity, negativity, and bigotry across the game genres that stayed consistent when comparing male and female streamers. For example, games within the first-person shooter and online battle arena categories repeatedly showed higher levels of gendered harassment, bigotry, and negative responses to gameplay. In comparison, games existing outside of these categories, such as Minecraft, Diablo IV and League of Legends, exhibited IRC channels that were more positive. The male Minecraft stream was notably peaceful, with the streamer and audience engaging in a 5-h livestream of intricate storytelling. Within the chat, users contributed to a collaborative narrative, interacting with the characters the streamer was playing as, rather than with the streamer himself.
As a result, the associated IRC channel consisted mainly of positive interactions free from sexual content or bigotry. It is encouraging that these findings corroborate those from prior research, which have found that Minecraft is a game that facilitates and promotes positive social interaction (Bukvic et al., 2014; Riordan and Scarf, 2017). This is because, unlike first-person shooter or battle arena games, Minecraft has no specific aims or goals, and so players are given the creative ability to build their own world and immerse themselves in a narrative of their own creation (Riordan and Scarf, 2017). This also lends itself to positive socialisation, as the player can create a community that supports creative invention within this virtual world, and in the case of Twitch, this involves viewers being given the ability to engage with the streamer and help build the in-game narrative (Riordan and Scarf, 2017). Therefore, the comments and positive IRC channel environment observed during this study could be seen as representative of the positive socialisation that Minecraft has been shown to promote.
Alternatively, within online battle royal game livestreams like Fortnite and Apex Legends, and the first-person shooter game livestream, VALORANT, IRC channels in both male and female streams were often filled with prejudice, gendered harassment, sexual comments, and profanity. This trend aligns with previous studies that found online multiplayer games encourage hostile behaviour, with competitive gameplay believed to increase the likelihood of aggressive or harmful behaviour (Ballard and Welch, 2017; Shores et al., 2014). Therefore, this could explain why online multiplayer games in this study consistently had more negative comments within the associated IRC channel. However, this study could not confirm that League of Legends promoted higher levels of trolling and harassment (Aguerri et al., 2023; Neto et al., 2017). Within this study, League of Legends streams had a clear lack of trolling and harassment that was consistent across both male and female IRC channels.
This finding contradicts previous studies, but this could be due to the small sample size, as only one male and one female League of Legends livestream was observed, which most likely does not represent the entire League of Legends community. While this study suggests that game content influences whether the IRC channel is positive or negative, it’s important to note that this effect is mainly evident among male streamers. Female streamers likely face hostility and gendered harassment regardless of game genre. Therefore, this study argues that game genre alone can’t explain why streamers’ IRC channels become rife with harassment, bigotry, and gendered harassment. Instead, the gender of the streamer as well as the employment of moderation functions may be more influential in mitigating or encouraging such content.
Conclusion
This study has highlighted the gendered dynamics present within Twitch livestreams, demonstrating that female streamers appear to face greater scrutiny of their gameplay, a greater focus on their appearance, and are more frequently sexualised compared to their male counterparts. This finding adds to current literature that suggests that female players, especially in livestreaming or live-chat environments, face higher levels of sexualisation and criticism in comparison to their male counterparts (Anciones-Anguita and Checa-Romero, 2024). Interestingly, this study also suggested male IRC channels exhibit higher levels of sexist language and homophobia, suggesting that harassment on Twitch is not solely a reaction to female participation but also reflects broader cultural norms regarding masculine policing within male-dominated gaming communities. These findings, specifically the presence of homophobic content, build on previous research that suggests male-dominated spaces involve the policing and reinforcement of masculine norms (Kimmel, 2008). This is an important observation and contributes to academic knowledge regarding the enforcement of masculinity, specifically within video game livestreaming environments. Furthermore, this study also demonstrates that the nature of chat environments is not only shaped by the gender of the streamer but also by moderation practices. This study notes that human moderation practices, rather than automatic moderation tools, have a greater impact on mitigating instances of gendered harassment and influence the cultivation of a positive or negative chat environment. However, this effectiveness also relies on a streamer’s engagement, because, as seen, when a streamer is unwilling to actively manage their chat, disruptive, bigoted, or prejudiced behaviour is allowed to proliferate, curating and reinforcing hostile community norms. This finding has significant implications for Twitch’s harassment policy. It is clear that automatic moderation is not enough to protect streamers and chat users from harassment, and thus, changes to community guidelines and policy could help create safer and more inclusive spaces for all users. Finally, while prior research has indicated significant differences in community norms related to game genre, this research cannot be certain of this (Aguerri et al., 2023; Ballard and Welch, 2017; Neto et al., 2017; Shores et al., 2014). This is because while there were some differences in game genre across the seven different titles, the IRC channel appeared to also be affected by moderation practices, the type of engagement streamers have with their viewers, as well as the gender of the streamer.
This research provides crucial insights, but several limitations must be acknowledged, including the relatively small sample size, the exclusion of non-English speaking streamers, and the presence of other forms of harassment, such as racialised abuse. Future research could build on this work by investigating whether patterns of female discrimination and exclusion persist in non-English speaking livestreams. Additionally, future studies could focus on other livestreaming sites like Kik or YouTube to determine the impact of moderation practices on other livestreaming platforms. Furthermore, given that homophobia and misogyny were clearly present throughout this study, it is reasonable to assume that other forms of harassment, including transphobia and racism, are also present, warranting further exploration. Finally, a larger, potentially quantitative study could examine differences within game genres to confirm whether this is a major factor affecting IRC channel environments or if other contributing factors have a greater impact. Despite these limitations, this research aims to deepen our understanding of online sexism and how gendered harassment manifests within livestreaming environments, illustrating both the technological and cultural factors that perpetrate these inequalities. Beyond contributing to current academic knowledge, this study provides a foundation for fostering a more equitable and inclusive digital space, highlighting the role of moderation, community norms, and streamer behaviour in shaping safe online environments. By illustrating these dynamics, the findings offer potential avenues for both researchers and online platform designers to develop interventions, policies, and educational initiatives aimed at mitigating harassment and promoting respectful engagement, ultimately supporting a more welcoming and sustainable culture for all participants within video game livestreaming environments.
Footnotes
Ethical considerations
This study was approved by the School Research Ethics Committee, Queen’s University Belfast (Approval Number: REF 102_2223). All research materials were handled in accordance with the ethical guidelines for research using online communities.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The research material used in this study was drawn from publicly accessible online forums. Due to ethical considerations, exact URLs and usernames are not disclosed to protect user anonymity. Further methodological details are available upon request.