The theory of reasoned action (TRA, Fishbein and Ajzen, 1975) has been widely used as a model for the prediction of behavioral intentions and behavior in technology adoption. Considering TRA model’s long development history is often related to several other popular theoretical models, we will first briefly review them and discuss our choice of TRA model.

As e-sports online engagement requires esports consumers to continuously adopt and use new technology, including both hardware (i.e., smartphones and computers) and software (i.e., live-streaming software), a plethora of e-sports studies have applied the technology adoption theories (i.e., Technology Acceptance Model and Unified Theory of Acceptance and Use of Technology) as the theoretical framework to examine the antecedents and consequence associated with esports engagement (Jang and Byon, 2019; Jang et al., 2021; Meng-Lewis et al., 2022). For example, Meng-Lewis et al. (2022) integrated the Expectation Confirmation Model (ECM) with the Unified Theory of Acceptance and Use of Technology (UTAUT) to identify the determining factors predicting users’ intention to watch e-sports games on the live-streaming platform. To emphasize the hedonic value (intrinsic motivation) of technology users, Venkatesh et al. (2012) proposed UTAUT2, which incorporates three additional constructs into UTAUT: hedonic motivation, price value, and habit. UTAUT2 application studies involved users’ engagement in plethora of technologies such as gaming on mobile devices (Ramírez-Correa et al., 2019), social network sites (Herrero et al., 2017) and mobile tv (Wong et al., 2014). Majority of the UTAUT2 application studies examined user adoption decisions since they examined new to the market technologies in nascent stages of their product life cycle (Tamilmani et al., 2021). Similarly, Xu et al. (2023) used the technology acceptance model (TAM) to investigate how sports customers adopted over-the-top (OTT) services to consume sport content online. The TAM (Davis, 1989) is actually an adaptation of the theory of reasoned action (TRA) proposed by Fishbein and Ajzen (1975) to explain and predict the behaviors of people in a specific situation (Legris et al., 2003).

While TAM is widely recognized for its application in studying the adoption of online and mobile technologies, we have chosen to utilize the TRA as the theoretical framework for our study, motivated by two distinct reasons. On one hand, watching e-sports games through the live-streaming service may not be accepted as a process of embracing new technology in China. The first esports live streaming platform in China, Douyu TV, began operating in 2014 (Meng-Lewis et al., 2022), which means esports online spectatorship may not be perceived as novel by Chinese e-sports fans, but rather an integral component of their daily lives. Furthermore, Buabeng-Andoh (2018) indicated that TAM is less general than TRA to determine technology usage behavior. Therefore, our study would take the TRA as the theoretical framework to explore the antecedents associated with esports engagement.

The TRA theory posits that behavioral intentions are the immediate antecedents to the behavior (Madden et al., 1992), while one’s attitude toward certain behaviors and one’s perceived social pressure (i.e., subjective norms) are two of the main factors that directly correlate with one’s intention to perform an action (Ajzen and Fishbein, 1977, 1980; Xiao, 2020). The attitude toward the behavior refers to the degree to which a person has a favorable or unfavorable evaluation or appraisal of the behavior in question, and the subjective norm refers to the perceived social pressure to perform or not to perform the behavior (Ajzen, 1991). In the context of this study, we seek to define e-sports audience’s attitude towards e-sports game online spectatorship through the degree to which they have a favorable or unfavorable evaluation on their e-sports game online spectatorship. In the similar vein, subjective norms are defined as their perceived social pressure to watch e-sports games online.

Generally, the more favorable the attitude and subjective norm with respect to a behavior, the stronger should be an individual’s intention to perform the behavior under consideration (Ajzen, 1991). Over the past few decades, the predictive power of intention of actual behavior has been corroborated in many fields. Therefore, in the context of e-sports game online spectatorship, it seems plausible that those with strong watching intentions are more likely to view the e-sports tournaments through live-streaming services. In this vein, given the pivotal role of behavioral intention in the theory of reasoned action framework and its predictive power of actual behavior (Xiao, 2020), this study aims to explore the factors leading to behavioral intention.

In studies outside the domain of e-sports, Jeong et al. (2021) examined the process behind the decision of sports fans to attend sports matches at stadiums amid the pandemic. Their findings suggested that attitude and subjective norms would influence their attendance intention positively. A study conducted by Leung and Chen (2017) also showed that attitude and subjective norms could significantly influence consumers’ intentions to adopt mobile TV services.

Given the commonalities between watching mobile TV and watching e-sports games on an electronic device, we hypothesize that:

Tarkiainen and Sundqvist (2005) explored the relationships between subjective norms, attitudes, and intention to buy organic food by applying structural equation modeling. Not only did their results consolidate the predictive power of attitude in the consumers’ intention, but also confirmed the positive influence of subjective norms on the formation of attitude. Hence, the following hypothesis is posited.

Motivation is one of the most heavily studied constructs in sport-related research (Snelgrove et al., 2008). Due to the emergence of live-streaming platforms and e-sports games, an increasing number of studies have begun to call for work in the e-sports spectatorship motivations. Previous studies have primarily applied the uses and gratifications theory (UGT) to develop the motivation scale and examine the users’ motivations for selecting and consuming media in the context of traditional sports or e-sports (Trail and James, 2001; Hamari and Sjöblom, 2017; Sjöblom et al., 2017; Sjöblom and Hamari, 2017; Brown et al., 2018; Cabeza-Ramírez et al., 2020; Wulf et al., 2020). For example, Trail and James (2001) developed the Motivation Scale for Sport Consumption (MSSC) to measure the motivations behind sport spectator consumption behavior. The MSSC relies on a similar theoretical understanding as the UGT, in that it focuses on the gratification and experiences that sports consumption affords for spectators (Hamari and Sjöblom, 2017). The MSSC and UGT share many mutual aspects, such as an escapism from everyday life, acquiring information from the media content, being a fan, social interaction and so forth (Hamari and Sjöblom, 2017). In terms of the construct validity and reliability, the MSSC demonstrated the good psychometric properties (Trail and James, 2001), as illustrated in Tables 1, 2.

As one of the most widely applied scales for sports consumption, the MSSC has gone through rounds of revisions and has resulted in current variations, which commonly consist of 10 constructs, including vicarious achievement, aesthetics of sport, drama of sport, watching sports as a means to escape everyday life, knowledge acquisition related to the sport, admiring the skills of the athlete’s, social interaction with other spectators, physical attractiveness of the athletes, novelty of new players and teams, and the enjoyment of aggression and the aggressive behaviors the athletes exhibit (Trail and James, 2001; Hamari and Sjöblom, 2017). In a study that is more aligned with the goal of our current study, Xiao (2020) selected six motivators from the MSSC as behavioral beliefs to examine their correlation with the intention to watch e-sports. Also, Kim and Kim, (2020) applied the MSSC (Trail and James, 2001) in their e-sports spectatorship studies.

With the Internet has been a significant tool for sports consumption, professional sport teams’ Web sites became an important component of people’s sports consumption. Yet, little was know about the Internet users’ consumption motives for sport teams’ Web sites, as the established motivation scales (e.g., the MSSC) focused more on the traditional media. To this end, Seo and Green (2008) developed the Motivation Scale for Sport Online Consumption (MSSOC) to measure dimensions of motivation for traditional sport Internet users. Based on this study, Brown et al. (2018) applied the motivations from the MSSOC to compare the uses sought and gratifications obtained when consuming media related to e-sports and traditional sports.

Nevertheless, not every scholar agrees that traditional sports consumption motivations are the same as e-sports spectatorship motivations. To this end, Qian et al. (2020b) developed the Motivation Scale of Esports Spectatorship (MSES) by considering the innate links that connect esports with video gaming and traditional sport. The composite reliability (CR) and average variance extracted (AVE) values of the MSES are shown in Table 3, to validate the factor reliability, convergent validity, and discriminant validity of it. Although Qian et al. (2020b) did not mention the Cronbach’s alpha and correlation coefficient of the factors in the MSES in their study, they pointed out that its convergent validity and discriminant validity indices were all greater than the threshold values.

Although the MSSC was treated as an adequate measure for esports spectating by Macey et al. (2022), they also highlighted that additional aspects specific to esports require further investigation. Consequently, esports online spectator motivation should be considered a hybrid of traditional sports spectatorship motivations and new media consumption motivations (Seo and Jung, 2016). In this case, our study attempts to combine the motivation research on e-sports and traditional sports. Grounded in previous studies, the motivations can be divided into three categories: (a) streamer-oriented attributes (b) technology-oriented attributes (c) individual motivations. Please see Appendix 1 for the dimensions and references of three motivation categories.

Streamer-oriented attributes refer to the personality traits and characteristics of streamers perceived by the audience, such as perceived social attractiveness (Kim H.-M. et al., 2023), perceived similarity (Kim H.-M. et al., 2023) and streamer skills (Kim and Kim, 2023). Technology-oriented attributes refer to the attributes of live-streaming platforms, live-streaming apps and over-the-top (OTT) services, such as ease of use (Wang et al., 2021), viewing quality (Xu et al., 2023) and stream quality (Qian et al., 2020a). Although these external elements do have a huge impact on e-sports viewing, this study aims to focus on the individual motivations for two reasons. Firstly, individual motivations are the intrapersonal or interpersonal elements (hedonic or internal) that influence people’s decision to consume e-sports (Qian et al., 2020a). Since this study aims to study the e-sports spectatorship from the perspective of Chinese e-sports fans, individual motivations are more in aligned with the goal of this study. Tang et al.’s (2022) study also indicated that esports spectatorship was driven significantly more by individual factors. In terms of the technology-oriented factors, the apps of different live-streaming platforms in China have almost the same features and functions, showing a trend of homogeneity. Essentially, these platforms are all easy to use and full of enjoyment. In addition to the technology factors, streamer-oriented factors are not as important as they are in the context of live game streaming, since there is no fixed streamer in the official e-sports event live-streaming room at all. Therefore, we believe that the differences in the individual motivations would be more influential than streamer-oriented attributes and technology-oriented attributes in this study.

In this case, we integrated the MSSC (Trail and James, 2001; Trail et al., 2003) with the MSES (Qian et al., 2020b) to achieve a new list of individual e-sports spectatorship motivations. In accordance with the goal of this study, we aim to incorporate a set of internal and intangible factors that motivate individuals to seek specific experiences in sports or esports spectatorship through the lens of their socio-psychological needs. Therefore, we selected seven individual motivations from the list, consistent with the context of online spectatorship. The seven motivations include: (1) skill improvement, (2) vicarious achievement, (3) knowledge acquisition, (4) escapism, (5) entertaining nature, (6) socialization opportunity, and (7) friends bonding, as Table 4 shows.

Since motivations are at the core of e-sports spectatorship studies, a large body of research has directly related motivations with actual purposeful behaviors of watching e-sports. For example, Hamari and Sjöblom (2017) found that escapism, acquiring knowledge about the games being played positively predicted e-sport spectating frequency, and similarly, acquisition of knowledge, escapism, and friends were found to be positively associated with watching intention in another research (Macey et al., 2022). Brown et al. (2018) focused on comparing the motives when consuming media related to e-sports and traditional sports from a use sought and gratification perspective. Their results revealed that eSports participants sought out media for social sport, fanship, and Schwabism (a feeling of having superior knowledge) (Brown et al., 2018). In terms of a real-time strategy game, StarCraft II, motivations including vicarious achievement, skill of the athletes, and entertainment value were found to impact the game attendance frequency positively (Pizzo et al., 2018). Skill of the athletes might also attract viewers to watch e-sports games. As Qian et al. (2020b) confirmed before, skill improvement was the unique motive that emerged in the esports context. Given these existing studies and the predictive power of behavioral intention in actual behavior, we hypothesize that:

In Xiao’s research (2020), motivators seemed to reflect an individual’s behavioral outcome expectancies, and he applied the expectancy-value model to explain how motivators influence attitude formation. Their results uncovered a positive relationship between motivation (aesthetics, drama, and escapism) and attitude (Xiao, 2020). Likewise, a great deal of literature starts with the use and gratification theory (Katz et al., 1973), proving that in the context of e-sports spectatorship, those with different motives tend to meet their expectancies by watching e-sports online and subsequently fulfill both psychological and social needs (Hamari and Sjöblom, 2017; Pizzo et al., 2018; Kim and Kim, 2020). Thus, we assume that fans may develop a positive attitude toward watching e-sports games online if the behavior could meet their expectancies. A hypothesis is posited as follows:

Past behavior is typically the strongest predictor of future behavior (Ajzen, 1991; Wang and Zhang, 2016). Moreover, the significant effect of past behavior on future behavioral intentions has also been corroborated in many fields, such as physical activity (Wang and Zhang, 2016), travel (Huang and Hsu, 2009), hockey games (Cunningham and Kwon, 2003) and so on, while little is known about its role in the context of e-sports online spectatorship. When it comes to the measurement of past behavior, most researchers adopted a single item, such as “Including last season, how many _____ men’s hockey games have you attended (Cunningham and Kwon, 2003)?” This single-item measure is applicable in the context of travel and traditional sports spectatorship. However, it is not so appropriate for the context of e-sports game online spectatorship in this study. On one hand, e-sports is a broad notion that includes so many different types of games, and undoubtedly, there should be a big difference between these games. On the other hand, watching e-sports games online is quite different from offline spectatorship. The most obvious difference lies in that when watching e-sports games online, viewers are more likely to be influenced by other entertainment ways on their electronic devices. For example, Macey et al. (2022) found that e-sports spectating would positively influence gaming intention. That is to say, online viewers are likely to play games, use social media, or do other irrelevant things after watching a game for a while, which offers an inaccurate measurement in this construct.

Despite the operationalization of past experience being blurred and arguable, Huang and Hsu (2009) developed a new approach to solving the problem, with the measurement of overall satisfaction with past experience. In their studies, overall satisfaction refers to a subjective evaluation of all past travel experiences in the destination, and the results indicated that people are more likely to revisit a destination if they have had satisfactory past travel experiences with it (Huang and Hsu, 2009). Therefore, we decided to adopt the measurement of satisfaction with past experience to explain one’s past behavior (Huang and Hsu, 2009), instead of the previously used single item (Cunningham and Kwon, 2003; Kaplanidou and Gibson, 2010). To better align with the goal of this study, satisfaction with past experience was defined as one’s subjective evaluation of his past experiences in e-sports game online spectatorship. Given this, we hypothesize:

In the TRA framework, attitude is determined by behavioral beliefs (Fishbein and Ajzen, 1975). Behavioral beliefs are one’s outcome expectancy of a behavior (Ajzen and Driver, 1991). Xiao (2020) believed that this expectancy-value logic proposed by Fishbein and Ajzen (1975) for TRA is akin to the need-gratification rationality proposed in U&G theory (Blumler, 1979) to explain media consumption behaviors and attitudes. Thus, in this study, spectatorship motivations can be seen as the equivalent to the behavioral outcome expectation, just as behavioral beliefs function in the TRA framework (Xiao, 2020). Lin (2014) applied the cognitive–affective–conative framework (CAC) (Fishbein and Ajzen, 1975) to explain the mediating role of attitude in the relationship between the beliefs and behavioral intention. He believed that the central theme in the TRA is that intention or willingness to perform a behavior (i.e., conative) is influenced by one’s attitude towards the behavior (i.e., affection). And a positive attitude towards the behavior is shaped by one’s beliefs and evaluations about performing the behavior (cognitive) (Lin, 2014). In the same vein, he found that attitude mediates the path between the effects of all gratifications (pastime, entertainment, relaxation, escape, and surveillance motives for local news) and intention to read citizen journalism news. In the context of sports, Kaplanidou and Gibson (2010) also revealed that there were mediation effects of attitudes between satisfaction and intentions to participate in the event again among the elderly. Given these, the following hypotheses on the mediating effects of attitudes are proposed:

Team identification reflects the intensity of a spectator’s association, a collective identity, with his or her team (Tajfel, 1982; Lee et al., 2017). Within the sports context, identification significantly influences recurring behavior, such as watching games on TV and participating in second-screen conversations about their team (Cunningham and Eastin, 2017). Research conducted by Wann and Pierce (2003) indicated that a fan’s attachment to a team is an important predictor of numerous affective, cognitive, and behavioral reactions. Specifically, those with higher team identification are more likely to attend future games (Wann et al., 2001; Matsuoka et al., 2003). Furthermore, Lee et al. (2020) investigated the complexity of how spectators’ multiple identities influence their behavioral intention and posited that team identity directly affects their attendance intention. Notably, e-sports fans do not support every team. Instead, they are supporters of their favorite team during the spectatorship, and thus e-sports viewers are likely to own strong team identification and loyalty to specific teams. Given this, a hypothesis is proposed:

Within the existing literature on traditional sports, researchers always classified sports fans into different groups based on the level of team identification in order to compare differences in their behaviors and feelings. For example, Bristow and Sebastian (2001) found significant differences between two fan groups in behavior and attitude to the Chicago Cubs baseball team, depending on the commitment of them. Wann et al. (1994) were able to show that highly identifying sport spectators reported an increase in positive emotions after watching a win, whereas low identifying viewers showed almost no emotional change. Schramm and Knoll (2017) used social identity theory to explain these differences that highly identifying sports fans view their team as being part of their social identity and therefore take a defeat of their team as a personal defeat and a win of their team as a personal victory resulting in respective mood. Furthermore, Matsuoka et al. (2003) assessed the interaction effects of team identification and satisfaction with facets of a game on intentions to attend future games, finding that the intentions of highly identified fans relative to low identified fans were less influenced by satisfaction. Given that e-sports fans also have varying levels of identification with different teams, similar to traditional sports fans, we hypothesize a moderation model.

Based on the analysis, the full conceptual model is illustrated in Figure 1.

This study collected data from October to November in 2022. In the first stage, a snowball sampling method was applied. We posted the online survey link in the game communities throughout social media platforms and also asked our friends to share the link with those who like to watch the e-sport games online. During the second stage, we recruited the respondents via an online survey platform, Sojump.¹ All the participants were required to answer the screening question (whether they had watched e-sports games online before) on the first page of the online questionnaire, and 93 respondents who failed the screening questions were removed. Additionally, those who spent too much time (more than 500 s) or very little time (less than 100 s) completing the questionnaire were also removed. As a result, the final data set constituted 452 valid responses. The participants were rewarded with a monetary incentive (RMB12) after completing the survey.

Overall, the majority of participants were aged between 19 and 39 years (n = 411, 90.93%), mostly possessed a bachelor’s degree (n = 285, 63.05%), and were comprised of 183 (40.5%) females and 269 (59.5%) males. Compared to FPS (first-person shooter) games, participants obviously preferred watching e-sports tournaments based on MOBA (multiplayer online battle arena; n = 374, 82.74%) games. All the sample demographics are presented in Table 5.

After a screening question, qualified participants were asked to report their e-sports game online spectatorship motivation, attitude towards e-sports game online spectatorship, subjective norm, satisfaction with past experience, team identification and future online watching intention for the e-sports games. All measures were rated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree) unless otherwise stated.

The scale for e-sports game online spectatorship was adapted mainly from the MSSC (Trail and James, 2001) and the MSES (Qian et al., 2020b). The motivations include: skill improvement (α = 0.905, M = 5.65, SD = 1.02), vicarious achievement (α = 0.861, M = 5.88, SD = 1.12), knowledge acquisition (α = 0.759, M = 5.95, SD = 0.98), escapism (α = 0.849, M = 4.90, SD = 1.56), entertaining nature (α = 0.73, M = 6.00, SD = 0.91), socialization opportunity (α = 0.933, M = 4.63, SD = 1.60) and friends bonding (α = 0.899, M = 5.47, SD = 1.23). Twenty-five items were employed to measure seven motivations. Example items include “watching the e-sports game online helps me become a better player (skill improvement),” “I feel proud when my preferred team (or player) does well (vicarious achievement),” “I increase my knowledge about a game by watching e-sports game online (knowledge acquisition), “watching e-sports game online provides an escape for me from my day-to-day routine (escapism),” “I watch the e-sports game online because it is fun to watch (entertaining nature),” “I enjoy interacting with other fans online when watching the e-sports game online (socialization opportunity) and “watching the e-sports game online gives me a chance to bond with my friends (friends bonding).”

The measurement of attitude followed the design in Xiao’s study (2020). Xiao measured the attitude of the e-sports audience toward watching e-sports using these four items. Specifically, attitude was assessed using four semantic differential scales in response to the following item: “For me, watching e-sports game online is ….” The four scale items were “Extremely bad–Extremely good,” “Extremely worthless——Extremely valuable,” “Extremely unpleasant——Extremely pleasant” and “Extremely Boring——Extremely Interesting” (Xiao, 2020). Participants provided responses on an interval scale ranging from 1 to 7, where 1 = Extremely bad and 7 = Extremely good, for example. The internal consistency of this measurement is good (α = 0.883, M = 5.89 SD = 0.93).

Three adapted items from Xiao (2020) were used to measure subjective norm. Examples include “I want to watch e-sports game online because my friends do so, and I want to belong to the group.” The internal consistency of this measurement is good (α = 0.830, M = 4.42, SD = 1.52).

Satisfaction with past experience was measured with three adapted items from the scale developed by Huang and Hsu (2009). Example items include “my overall evaluation on the past experience of watching e-sports game online is positive.” The reliability of this measurement is satisfactory (α = 0.849, M = 5.86, SD = 0.92).

Team identification was measured with three adapted items from the scale developed by Jeong et al. (2021). Examples include “when someone criticizes my favorite e-sports team, it feels like a personal insult.” The internal consistency of this measurement demonstrates strong reliability (α = 0.832, M = 4.25, SD = 1.51).

Two items from Leung and Chen (2017) were used to measure future watching intention. Example items include “I intend to watch e-sports game online when it becomes available.” The internal consistency of this measurement demonstrates strong reliability (α = 0.811, M = 5.62, SD = 1.25). All the measurement items and their descriptive statistics are shown in the Appendix 2.

Demographics (i.e., age, gender, and educational background) and the favorite type of e-sport game to watch was controlled for in this study.

We performed the data analysis primarily using Mplus 8 software (version 8.3) and SPSS 24. Specifically, we utilized Mplus 8 for conducting structural equation modeling (SEM). To ensure the validity of the measures, we conducted confirmatory factor analyses (CFA) with maximum likelihood estimation using Mplus 8 software. We assessed the validity and reliability by examining Composite Reliability (CR), Average Variance Extracted (AVE), and the square root of AVE. In this study, we employed a customed calculator to calculate the CR and AVE values, which were obtained by calculating the standardized factor loadings of items. The functionality of this calculator has been verified by the authors.

The accepted threshold values for CR and AVE are 0.7 and 0.5, respectively. All item loadings should be higher than 0.5. Additionally, the square root of AVE must be greater than the correlation with any other constructs in the model to establish discriminant validity. As all the indicators were found to meet these standards, both validity and reliability were established (see Tables 6, 7 and Appendix 2).

We also examined the fit of our hypothesized model (Figure 2) in Mplus 8. It was assessed with the comparative fit index (CFI), the root mean square error of approximation (RMSEA), and the standardized root mean squared residual (SRMR) (see Table 8). CFI value of 0.95 or higher, an RMSEA value of close to 0.06 or less, and an SRMR value of close to or less than 0.08 are indicative of good model fit (Hu and Bentler, 1999). In addition, chi-square statistics and Tucker-Lewis index (TLI) were also reported to suggest a satisfactory fit.

SPSS 24 was applied to perform the hierarchical regression between the related factors and future watching intention. Furthermore, we used process 4.0 in SPSS to test the moderation and mediation effect of specific variables. By the calculation of SPSS 24, the Variance Inflation Factor (VIF) scores for all variables were between 1.063 and 3.626, which were well below the threshold of 5, indicating the absence of multicollinearity (Xiao, 2020).

In order to address RQ1 and H4, we conducted hierarchical regression analysis to confirm the direct relationship between the relevant factors and future watching intention The results are shown in Table 9. We found that skill improvement (β = 0.180, p = 0.001), entertaining nature (β = 0.225, p < 0.001), and friends bonding (β = 0.242, p < 0.001) were positively related to the intention to watch e-sports games online. Thus, RQ1 was addressed and H4 was partially supported. The results also indicated that friends bonding was the strongest spectatorship motivation among e-sports fans in China.

To investigate RQ3 and the related hypotheses, we examined the hypothesized relationships between constructs and reported standardized path coefficients, denoted as β. The model result is presented in Figure 3.

In our model, skill improvement (β = 0.255, p < 0.01), entertaining nature (β = 0.665, p < 0.001), friends bonding (β = 0.270, p < 0.001), and satisfaction with past experience (β = 0.231, p < 0.001) were positively associated with attitude towards e-sport game online spectatorship. Moreover, as expected, attitude (β = 0.622, p < 0.001) and satisfaction (β = 0.202, p < 0.001) could positively influence future watching intention.

Thus, H1, H6, and H7 were confirmed, and H5 was partially supported.

Furthermore, while the regression results showed a positive relationship between team identification (β = 0.155, p < 0.001) and watching intention, this path did not reach statistical significance in the structural equation model. Consequently, we concluded that H9 was not fully supported. The details would be discussed as below.

Exploring RQ2 and H8, we employed model 4 in Process 4.0 to examine the mediation effect of attitude. Based on our theoretical model, we hypothesized that attitude mediates the relationship between spectatorship motivators and watching intention, as well as the relationship between overall satisfaction and watching intention. The results confirmed the significant mediation role of attitude in the model (see Table 10). Thus, RQ2 was addressed, H8a and H8b were supported.

Model 15 in Process 4.0 was used to test the moderation effect of team identification. The results showed that team identification negatively moderated the relationship between satisfaction with past experience and future watching intention (β = −0.144, p < 0.01), while no significant moderation effect was found between attitude and intention (refer to Table 11). That is, high-identified individuals would persist in watching the games of their favorite team, despite their dissatisfaction with their past watching experience, while the low-identified individuals are influenced more by overall satisfaction with their past watching experience. Thus, H11 was supported.

This study contributes to e-sports spectatorship studies by revealing the significant motivations that are associated with attitude and watching intention. Previous studies were mainly conducted in the Western and ignored some defining characteristics of e-sports game spectatorship. This study goes beyond it by creating a new list of spectatorship motivations, consistent with the context of e-sports game online spectatorship. It indicates the powerful influence of skill improvement, entertaining nature and friends bonding on Chinese e-sports fans’ attitude and watching intention.

Skill improvement is the unique motive that emerges in the e-sports context (Qian et al., 2020b). Furthermore, e-sports gameplayers are always considered to be potential viewers of e-sports events (Tang et al., 2022), therefore it is plausible that these e-sports spectators tend to watch others play to improve their own gameplay performance (Seo and Jung, 2016; Jang and Byon, 2019; Qian et al., 2020b; Tang et al., 2022).

In addition, our study shows that the entertaining nature of e-sports game online spectatorship is associated with attitude and future watching intention, which is in line with previous work showing that esports is now considered a popular form of spectator entertainment (Brown et al., 2018; Qian et al., 2020b,c). One plausible explanation is that entertainment has a positive effect on creating pleasant emotions (Tan, 2008) and emotional gratification (Bartsch, 2012), which in turn elicits the favorable attitude and behavioral intentions (Eckler and Bolls, 2011; Foroughi et al., 2019).

Besides, it is also worth noting how esports might exert a positive impact on an existing social relationship or established friendship (Pizzo et al., 2018; Qian et al., 2020b), as our results show that friends bonding always positively influences watching intention, especially in China. On one hand, e-sports online spectatorship, as the digital counterpart of live events, provides bonding opportunities and social gratification similar to traditional sport spectatorship for e-sports fans (Qian et al., 2020b). On the other hand, because of the nature of e-sports and the relatively novelty of this phenomenon (Macey et al., 2022), the primary influencers when people first started watching esports were their close friends (Qian et al., 2020b). Therefore, it is common that sharing with close friends about e-sports games in China not only creates an environment that extends beyond mere socialization (Funk et al., 2004), but also treated as a means of relationship strengthening (Kissane and Winslow, 2016), particularly with friends in their social circle.

Prior work on this topic found that escapism (Hamari and Sjöblom, 2017) and knowledge acquisition (Hamari and Sjöblom, 2017; Tang et al., 2022) positively predict e-sports spectatorship. Besides, escapism was found to be positively related to attitude toward watching e-sports (Xiao, 2020). These findings are entirely different from the conclusions of this study. In our study, e-sports fans in China are motivated to watch games by their competence, entertainment and social motivations, as opposed to their informational motivations (Cabeza-Ramírez et al., 2020) or escapism (Hamari and Sjöblom, 2017). The significant differences in these motivations may perhaps explain why the e-sports industry has been able to rapidly grow in China, as watching and discussing e-sports games have become a part of people’s daily entertainment and social interaction.

In addition to extending the studies in spectatorship motivation, this study also seeks to understand the role of past behavior. Our study reveals a stable positive relationship between satisfaction and both attitude and behavioral intention.

Actually, rich literature in the broad leisure industry (such as tourism) has identified the critical role of past satisfaction (Petrick et al., 2001; Huang and Hsu, 2009; Huang et al., 2015), though it is understudied in the e-sports context. The underlying assumption of these studies is if an experience has a positive effect on an individual, they are more likely to repeat the activity (Petrick et al., 2001). Therefore, it is plausible that e-sports fans may develop positive emotions associated with e-sports online spectatorship based on their prior experience, as rewatching behavior resembles a rational planned behavior, similar to what is observed in tourism revisiting behavior (Huang and Hsu, 2009).

In general, the study contributed to the understanding of past watching experience’s role in future watching behavior by including satisfaction as a component of the construct.

This study also highlights attitude as a mediating mechanism through which spectating motivations and satisfaction with past experience are associated with future watching intention. Previous studies on e-sports spectatorship mainly focused on simply verifying the relationship between motivation, attitude and watching intention, while no one seeks to test the mediation effect of attitude and confirm the mediating role of it in the context of e-sports online spectatorship. Our study is one of the first studies to confirm the mediating role of attitude in the context of e-sports online spectatorship. Kim M. et al. (2023) explained the mediating power of attitude on spectator revisit intentions of Paralympics with the concept of sequential relationships among cognitive, affective, and behavioral components, which is common in the tourism literature (Kwon and Vogt, 2010; Wang et al., 2020). In the context of reading citizen journalism news, Lin (2014) also applied the cognitive–affective–conative framework (CAC) (Fishbein and Ajzen, 1975) to explain the mediating role of attitude in the relationship between the beliefs and behavioral intention. As motivation could be considered as a cognitive component (Kim M. et al., 2023), it is plausible that affective attitude works as a mediator on the effects of cognitive component (motivation) on behavioral intention (Wang et al., 2020).

In our study, we also found that the interaction of team identification and satisfaction with past watching experience is a significant predictor of future watching intention. As can be seen in Figure 4, both high-identified e-sports viewers and low-identified fans expressed lower levels of intention to watch e-sports games online in the future when they were less satisfied with their past watching experience. However, such reduction in future watching intention was much more pronounced in the case of low-identified viewers than high-identified viewers. Specifically, those low-identified individuals would be influenced heavily by satisfaction with past watching experience. One possible explanation is that team identification is always strongly associated with psychological factors such as trust (Wu et al., 2012), fan loyalty (Bodet and Bernache-Assollant, 2011), and a sense of belonging (Heere and James, 2007), thus influences the perception high-identified individuals have of their own behaviors. In this vein, it is reasonable to expect that they would persist in watching the games of their favorite team, despite their dissatisfaction with their past watching experience. As for these low-identified ones, overall satisfaction plays a more critical role in their future watching intention, as a result of a lack of perceived psychological connection with a specific team. This finding is consistent with previous studies conducted in the domain of traditional sports (Bristow and Sebastian, 2001; Matsuoka et al., 2003), underscoring the significant influence of team identification on sports or e-sports fans’ intentions and behaviors.

While previous research suggested that team identification positively influences future intention (Wakefield, 1995; Matsuoka et al., 2003; Trail et al., 2003), our study did not arrive at the same conclusion. This discrepancy may be attributed to differences in our measurement of future watching intention, which was not limited to watching only the games of the e-sports fan’s favorite team, unlike previous studies that measured it through the name of a specific team (Wakefield, 1995; Matsuoka et al., 2003). Additionally, the development of team identification in esports is unique and always in a more fluid and dynamic way (Qian et al., 2023). This implies that investigating its precise generative mechanisms within the context of e-sports could be a promising direction for future research.”

Our study also provides some practical implications. First, the motivations defined in our study suggest a need for live-streaming platforms and e-sports game organizers to devote more attention to the practical, entertaining, and social value of e-sports games. E-sports practitioners should implement specific mechanisms to help viewers better understand and learn from the gameplay of professional players, such as live replays and in-depth gameplay analysis during live-streaming. Given the popularity of e-sports games as a social activity among friends, e-sports game organizers should schedule more competitive games to encourage discussion among viewers. Second, the mediating role of attitude highlights the importance of cultivating a positive attitude towards e-sports game online spectatorship. E-sports practitioners should focus on creating a favorable attitude among spectators. Third, considering the moderating role of team identification, e-sports practitioners should aim to enhance overall satisfaction for low-identified viewers by improving live-streaming quality and reducing game pauses. On the other hand, e-sports clubs can foster loyal fans with high team identification by recruiting star players or enhancing their team’s performance.

However, this study has certain limitations. Firstly, the sample composition is limited, and the data analysis is cross-sectional, which only reveals the interrelationships between constructs. Future studies could employ longitudinal research to examine the causal relationships between these variables.

Secondly, this study did not impose restrictions on the types of e-sports games. However, each game has its defining characteristics, and preferences for specific game types would influence spectating behaviors (Tang et al., 2022). Therefore, future research could focus on a specific game, such as League of Legends (MOBA) and VALORANT (FPS).

Third, the identification scale in this study solely focuses on team identification, while player identification and e-sports identification may also be significant factors in e-sports game online spectatorship. Future research should consider including them in the context of e-sports and exploring the e-sports fandom culture further.

Lastly, gender differences exist in this male-dominated domain, particularly in terms of unique e-sports consumption motivations (Yu et al., 2022), but this study did not emphasize them. Therefore, scholars should make more effort to explore possible gender differences in the context of e-sports.