Live-streaming Phenomena on Twitch: A Systematilicencec Literature Review of International Scholarly Research

Exclusion criteria:

1. Studies that are not available in open access or are not accessible via the different institutions with which the authors are affiliated.

2. Research for which Communication Studies is not the main discipline.

3. Any records other than published articles, book chapters, technical reports or conference proceedings.

4. Publications that do not deal with the streaming phenomenon, or with streamers and/or influencers as the main object of study.

5. Publications lacking a well-defined research design, with replicable procedures and a solid theoretical framework that supports their objective and their findings, as these elements are considered essential to ensure the quality of the contributions to the field of study.

2.2 Performance of searches

After obtaining the results, a two-step manual selection process was undertaken, with the application of the inclusion and exclusion criteria (complete diagram in Figure 1). The total number of records extracted from the two databases was n=247, with n=123 from Scopus and n=124 from WoS. The first phase involved reading the abstract of each article to determine whether it met the inclusion criteria. This process reduced the sample to n=85. The records excluded were articles in other languages (n=2), publications that were not one of the selected types (n=14) or did not deal with the phenomenon as the main object of study (n=28), and articles outside the field of communication studies (n=46) or not openly accessible (n=1). Duplicates (n=55) were also eliminated. The second phase involved reading the full text of each publication, applying the fifth exclusion criterion (n=6), i.e., studies that lacked a rigorous methodology or theoretical basis. This resulted in a final sample of n=73: n=38 from Scopus and n=35 from WoS.

Figure 1. Flow diagram for the searches and the selection of publications

Source: created by the authors.

2.3 Analysis

A data sheet was developed to present the information related to the research methods, author affiliations and other aspects of the publications, complemented by a content analysis using analytical inference, for the specific purpose of developing the review outlined in section 3.5.

The research design adopted to analyse the object of study employed the categories proposed by Cabeza-Ramírez et al. (2021) in their systematic review, which adapts the main elements of Shannon’s mathematical theory of communication (1948). Combining the methodological proposals of these authors with those of Harpstead et al. (2019) and Li et al. (2020), it was determined whether each record focused on the transmitter, receiver, channel or transmission process. Equivalences were established between these categories and the elements of live-streaming, the transmitter being the streamer or social media influencer, the receivers being the audience or viewers, the channel being the Twitch platform, and the transmission being the codes and message. This framework made it possible to categorise the results obtained (Table 1).

3. Results

3.1 Evolution of scholarly output

The analysis of the 73 records that make up the sample of this study reveals a growth in scholarly output over the period that can be divided into two stages. Although the searches covered the period from 2008 to 2023, the first publication found was in 2012: a study of video game live-streaming (Kaytoue et al., 2012). Between 2012 and 2017, there were no publications that met the search and selection criteria; however, since 2017 there has been a trend of constant growth, apart from a drop observed in 2021.

As Figure 2 shows, two years stand out over the rest. The number of studies doubled in 2019, the year when the live-streaming format became popular in the gaming community (TwitchTracker, 2024). Scholarly output peaked in 2022, when a total of 18 papers were published (24.66%), more than double the number in the previous year. As shown in Figure 3, the country that is home to the largest number of affiliated authors is the United States (28 records), the same country where the Twitch platform was created. In second place is Spain, followed by Germany and the United Kingdom.

Figure 2. Scholarly output data per year (2012-2023)

Source: created by the authors.

Figure 3. Affiliation of authors by country

Source: created by the authors.

3.2 Methodologies used

Qualitative research methods are used in more than half of the studies analysed (53.42%). The rest use quantitative (38.36%) or mixed methods (8.22%).

As shown in Figure 4, the most commonly used qualitative data collection methods are interviews (15 records), digital ethnography techniques (13) and content analysis (12). In the case of quantitative studies, only two methods are used: surveys, in 16 studies representing 57.14% of the total; and data clustering, in 12 records (16.44%). Mixed methods are used in six papers: [ID03] use a methodological triangulation approach that combines the monitoring of data on a streaming channel on Twitch with interviews and participant observation; [ID05] combine the development of multi-tier architecture to relate data quantitatively with a qualitative content analysis; [ID14] combine an online survey with direct non-participant observation; [ID36] complement a case study in their research with a survey; [ID61] use both in-person surveys and semi-structured interviews; and [ID72] combine data mining in a data cluster with video content analysis.

Figure 4. Comparison of qualitative and quantitative methods

Source: created by the authors.

3.3 Objects of study and associated disciplines

The records selected take a variety of approaches to their objects of study (Figure 5). Applying the categories proposed Cabeza-Ramírez et al. (2021), more than half of the studies (54%) focus on Twitch (19%), streamers (21%) or viewers (13%). There is only one study focusing on the message [ID125] while the rest (46%) explore a combination of objects of study and approaches.

Figure 5. Objects of study

Source: created by the authors.

Given that communication studies is the academic discipline that concerns this review, it is not surprising that the majority of the publications are in this area (40 records, 54.8%). A notable disciplinary sub-category is human-computer interaction (HCI), with 21 records (28.77%), while computer science and psychology each represent much smaller proportions of the sample (5.4% and 4.1%, respectively).

3.4 Theoretical frameworks of articles in the sample

A distinction was identified between articles adopting a specific theoretical framework and those that included literature reviews, the latter being the predominant approach (64.38%). A total of 15% of the publications combine a review of the literature with the analysis and application of social theories, while only 2.7% (2 articles) are based exclusively on social theories. Table 2 presents a summary of the theories used in the sample.

3.5 Relevant results of the studies analysed

This systematic review has made it possible to identify key contributions, potential contradictions and opportunities for further research, while also determining the consensus on the object of study and explaining the scholarly interest it has aroused. Table 3 presents a summary of the main points addressed in the articles that make up the sample, which are discussed below. The data package available includes the complete list of the articles, ordered according to the numbering system used in the table.

3.5.1 Importance and consequences of social interaction

Social interaction is key to understanding the live-streaming phenomenon, as it effectively underpins its activity [ID52] and is recognised as a source of its success [ID69]. Both socialisation and the emotional connections established in the community are crucial for attracting and retaining audiences [ID49], [ID15], [ID27]. Live-streaming elicits a high level of audience engagement through its combination of audiovisual content with paratextual dialogue [ID69], creating a sense of belonging, affective networks, socialisation and entertainment, all of which has a positive psychological impact on viewers. There are various factors that influence socialisation on Twitch, including gender [ID40], age [ID14], the influencer’s personality [ID12]; [ID27]: [ID53], the viewer’s role [ID61] and the platform’s community guidelines [ID08]. The combination of these factors may have either negative or positive social and economic effects.

Positive effects

There is evidence that socialisation in this context is not inherently negative. Consuming limited amounts of different content results in a better entertainment experience and the communities created can offer emotional support and positively influence their members (Table 3). Attending online events, participating regularly in chats and sharing personal information in an environment regulated by moderators can help build a healthy, engaged community [ID30]; [ID49].

From the Foucauldian perspective adopted by [ID70], receiving gratifications, establishing relationships and taking part in shared experiences in the same space are ways that users exercise their freedom. In this context, subscriptions and donations to the channels are founded on users’ decision-making capacity. Consequently, the revenues that streamers generate from live-streaming are higher and more consistent than revenues from other sources.

Negative effects

According to [ID05] positive social effects play a crucial role in the clinical evaluation of pathological video game use, which can be extrapolated to the platform. This condition is associated with three factors: playing video games or watching live-streams as a way of escaping reality, experiencing intense negative feelings when unable to do so; and spending a gradually increasing amount of time gaming or viewing content.

Social interaction supported by gamification techniques [ID22] serves as a strategy for live-streamers to increase engagement and thus to monetise their activity. The more time users spend viewing live-streams, the more likely they are to spend money on donations or subscriptions, which are the channels’ main sources of revenue [ID15]. Subscriptions, donations and cheers on Twitch are driven by a social motivation associated with gaining prestige within the community, although their economic basis can distort the perception of their emotional worth, resulting in conflicts when economic value and social value are confused.

The motivations for belonging to competitive gaming communities on live-streaming platforms are associated with an increased likelihood of offensive comments. Protection against these behaviours is limited due to the vagueness of the platform’s community guidelines, which can result in negative social interactions depending on the user’s personality [ID05]; [ID08].

3.5.2 Inequality, the gender factor and harassment

Women continue to have their legitimacy as participants undermined on Twitch’s gaming spaces [ID08] and a clear gender bias is evident on the platform, with women tending to focus more on social interaction and to address humorous topics [ID49]. Only 16.9% of Twitch users are women [ID15]; [ID53].

The findings of this review reveal a sexist, heteronormative dynamic, as evidenced in research by [ID29]. On the one hand, it has been found that male viewers are more likely to watch female streamers who wear provocative clothing, and to make donations to streamers of the opposite sex [ID07]. On the other hand, users are more likely to engage in parasocial relationships and make donations to streamers of the same gender, and content creators who actively promote gender homophily generally attract a bigger audience [ID54].

[ID39] identified a greater tendency toward objectification of streamers in chats on female channels than on male channels. [ID40] describes this systemic sexual discrimination in live chats as subtle, he also acknowledges that there are also occasional blatant cases of aggression, such as devaluing the work of female influencers on Twitch [ID55]; [ID57], employing humour to express harmful attitudes [ID19] and using explicit terms such as “titty streamer” [ID47].

Another significant problem on the platform is harassment, which can be broken down into five categories: sexual harassment; expressions of hate (racism, sexism, homophobia and transphobia); offline harassment; physical attacks; and “backseating” [ID58].

3.5.3 The streamer

Streamers themselves are often an object of study. Audience management, leadership and the challenges associated with public exposure are the main topics explored. However, although leadership and personification are characteristic features of this form of communication [ID02]; [ID17]; [ID53], content creators rely on a whole social structure that supports and sustains them [ID44]. The community created around streamers feeds off the activity and participation of their audience [ID14].

The main factor determining the decision to subscribe to a channel is the visual design of both the channel and the website, as it reflects the personality and style of the streamer [ID53]. Other qualities such as the streamer’s gender, skills and communication style also have a significant influence on this decision [ID61].

3.5.4 Influence and evolution of Twitch

Twitch has emerged as an essential platform in the gaming and marketing industries [ID22]; [ID23]; [ID27]; [ID72]; [ID73]. Perhaps surprisingly, the platform has begun making use of traditional media formats such as radio and television ([ID36]; [ID45]; [ID51]. This move towards other media formats, which is especially notable in the Spanish-speaking community, reached a watershed moment in 2020 in the context of the confinement and isolation resulting from the COVID-19 pandemic lockdowns [ID51]. Despite the content and format changes, key aspects of live-streaming remain the same, such as the chat function, which facilitates collective, participatory media consumption [ID16].

3.5.5 The moderator

As characteristic features of online communities [ID61], moderators have attracted considerable scholarly attention. The moderator’s role is of vital importance to social interactions in live-streaming communities as they not only monitor compliance with the community guidelines but also enhance the dynamic quality of the streaming experience. As mentioned above, their importance to the community makes moderators role models for other users [ID30] and they promote positivity through active conversations. Other important roles played by moderators include assisting the streamer with technical issues and combating harassment in order to maintain a safe space for all users.

3.5.6 Potential of live-streaming for education

Live-streaming offers possibilities for collaborative work. The use of the platform for educational purposes in the field of computer science constitutes an example of the potential for its development as an alternative form of education [ID29]. Gamification has become the key feature of the educational experiences offered on the platform, with the design of work rooms that facilitate the simultaneous participation of all users in real time. Different interactive modes such as the chat function and the integration of images are combined to enrich the user experience through a higher level of interaction than what is offered on other virtual platforms [ID03]. [ID69] highlight the multi-screen experience that allows users to view different kinds of live content and interact at the same time.

4. Discussion

This discussion focuses on the findings of the systematic review of the literature on live-streaming on Twitch, stressing its significance as a constantly evolving sociocultural phenomenon. As Twitch becomes consolidated as a space for social interaction, complex dynamics are emerging that affect both content creators and their audiences. The possibility of interacting in real time with streamers and other users creates affective bonds and reinforces the feeling of community among viewers. This potential for facilitating engagement explains much of the platform’s appeal and its continued popularity. However, some of the studies reviewed warn of the risks associated with prolonged use, which may lead to pathological behaviours such as social isolation or emotional dependence, exacerbated by the use of gamification techniques aimed at monetising online activity.

The research also highlights persistent gender inequality and harassment on the platform, which call for critical attention. In particular, women face higher levels of harassment, as confirmed by numerous studies documenting cases of objectification and verbal aggression towards them in chats. This harassment not only affects the perception of safety on the platform but also limits the development of an equitable participatory experience. Although measures have been implemented to moderate user behaviour on the platform, current policies appear to be insufficient to effectively address the sexist dynamic present in the community.

Moreover, the transformation of the streamer into a character with a hybrid role as both an entertainer and a marketer raises questions about the sustainability of this model in the future. This development reflects Twitch’s transformation into a space where streamers not only create content but also manage communities and establish intense parasocial relationships with their audiences. This model allows them to professionalise their activity and earn revenue through donations, subscriptions and sponsorships. The studies reviewed confirm that streamers benefit from their capacity to engage their audience emotionally, thereby consolidating their position in the digital economy.

Twitch has ceased to be a platform focused exclusively on gaming to become a diversified media outlet offering services ranging from education to cultural entertainment. Its evolution towards formats more characteristic of traditional media, such as television and radio, reflects its ability to adapt and grow beyond the gaming world. This transformation reinforces Twitch’s role as a key player in the reconfiguration of the digital media ecosystem, offering new opportunities for content creators and advertisers alike.

5. Conclusions

The live-streaming phenomenon is an object of study that is gaining increasing scholarly interest. The articles selected for this review date back only as far as 2017, indicating that this is a particularly recent body of academic literature. Despite the nascent quality of this field of study, a few main lines of research can be identified: (1) social interaction, which is the biggest focus of the scholarly output, with special attention to the factors shaping the relationship between users and streamers; (2) the notion of the streamer as a social media influencer; and (3) the live-streaming platform as a new media format with a huge influence on the gaming industry.

The Twitch platform reflects the shift towards thematic diversification being taken by these technology companies, which are embracing traditional media formats such as radio and television. Live-streaming is rapidly expanding, as is scholarly output on this phenomenon. This highlights the need to establish consistent, systematic terminology related to conventional media, social media, online platforms and content creators.

The new media environment and the emergence of new media formats are changing our understanding of what defines and delimits a media outlet and its main agents. This systematic review confirms that Twitch is one of the most innovative platforms in the world of digital audiovisual communication. Further research into the live-streaming phenomenon is needed to understand the future of audiovisual communication and to clarify where this media trend, fashion or revolution is headed.

A number of opportunities for future research on live-streaming and Twitch have been identified in this review. In particular, two lines of research could greatly enrich the work done so far in this area. The first involves theoretical approaches exploring asymmetrical relationships and user behaviour on Twitch. These could include studies based on the uses and gratifications theory, the self-determination theory or the self-expansion model, given that these do not take into account the unique media features of online live-streaming. The second would be studies of the profile of the compulsive consumer and the pathological use of the platform associated with gaming. For example, an analysis of users’ gaming skills and the progress they make while playing has rarely been included in studies of disruptive behaviours and how they can influence communication and social interaction processes.

This review has identified a number of gaps in the research that offer future opportunities for study in this field. These include:

1. Identification of the key elements that distinguish live-streaming platforms from other social media platforms. Twitch is a hybrid of a media outlet and a virtual community (based on the idea of a video game “clan” or “guild”) with live social interaction and different levels of exclusivity for users (Yoganathan et al., 2021).

2. Analysis of the potential of streamers beyond their popularity or the size of their audience, from the perspective of micro-communities and digital ethnography. Little is known about the social effects of platform mechanisms such as the “host” and “raid” functions, which content creators use to interact with other communities and to increase the visibility of other streamers. Researchers could examine what motivates streamers to use these mechanisms, how streamers on other channels are selected or how cooperation is perceived on the platform itself.

3. Studies of the relationship between player skills and time invested on the live-streaming channel, and how this affects social interactions and the pursuit of the objectives of the communication processes on the platform, such as monetising streamer activity or to improving the interaction.

4. In-depth research into gender differences in Twitch’s most obvious elements (streamers, viewers, and the social interactions between them), in view of the rapid evolution of these platforms and the entry of users of different backgrounds into the live-streaming community.

5. Studies of digital marketing used in live-streaming, from the perspective of advertisers. The reasons for its use and the anticipated benefits and profits that prompt advertisers to choose one content creator over another may be of vital importance for understanding the success or failure of activities on the platform [ID22].

6. Analysis of new trends and user profiles, taking into account the current thematic diversity of channels on Twitch, which include art, talk shows, education, travel, nature and sports.