Factors Influencing eSports Viewership: An Approach Based on the Theory of Reasoned Action

Definition of eSports

Multiplayer online battle arena games, such as League of Legends, first-person shooter games, such as Counter Strikes, and real-time strategy games, such as StarCraft, are some of the typical examples of eSports games. Regarding the definition of eSports, academic scholars have not reached a consensus. In fact, some media practitioners even doubt the legitimacy of calling eSports as a genre of sports. For example, John Skipper, the president of ESPN, does not consider eSports as a real sport, but a competition just as chess and checkers (Tassi, 2014). Perhaps the lack of physical contact in eSports matches leads the president of ESPN to reach his conclusion (Witkowski, 2012). In contrast, some of the scholars disagree with the notion of dismissing eSports as a genre of sports. Based on the definition of traditional sports, Wagner (2006) defines eSports as “an area of sport activities in which people develop and train mental or physical abilities in the use of information and communication technology” (p. 3). Hamari and Sjöblom (2017) define eSports as “a form of sports where the primary aspects of the sport are facilitated by the electronic system; the input of players and teams, as well as the output of the eSports system, are mediated by human-computer interfaces” (p. 221). Their definition focuses on the difference between the outcome generated by traditional sports and eSports. Specifically, athletes who play traditional sports, such as football or basketball, create real-world outcomes (e.g., physical contact), while professional gamers create virtual outcomes (e.g., computer-mediated interactions).

eSports have become a popular topic among communication scholars in recent years. Bowman, Weber, Tamborini, and Sherry (2013) investigated how audiences of videos games may influence the performance of the player of the game in a laboratory experiment. Using social facilitation theory as the theoretical framework, Bowman et al. found a positive correlation between the presence of audiences and high-skilled players’ performance, while the presence of audiences negatively affected low-skilled players’ performance. Findings from their study suggest the presence of viewers has a significant influence on the performance of a gamer, while the performance of a gamer affects the enjoyableness of viewing the game in return. In a research study that is more aligned with the goal of the current study, Brown, Billings, Lewis, and Bissell (2018) found that competition, passing time, and Schwabism (i.e., a feeling of having superior knowledge) were factors motivating eSports fans to participate in eSports competitions, while fanship, camaraderie, and arousal were factors leading fans of eSports to be satisfied with the eSports-related media consumption experience. Their findings indicate that the reasons for one to participate in eSports consumption are similar to the reasons for one to select which media to consume (Brown, Billings, Lewis, & Bissell, 2018). Several extant studies of eSports have directly related motivators with actual behaviors of watching or playing eSports games (Brown et al., 2017; Brown et al., 2018; Hamari & Sjöblom, 2017). However, this may oversimplify the correlations between motivators and behaviors because the analytic process has missed essential elements (e.g., behavioral beliefs, attitude, and subjective norm), as well as the linkage between attitude, behavioral intentions, and behaviors. To advance our knowledge of eSports, the current research study examined a different set of eSports consumption motivators, situated the motivators in the theoretical framework of TRA as behavioral beliefs, and utilized structural equation modeling (SEM) as the data analytic approach to uncover correlations between variables.

TRA and the Intention to Watch eSports

TRA is often used to explain how different factors predict a person’s behavior (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975; Sheppard, Hartwick, & Warshaw, 1988). According to TRA (see Figure 1), behavioral intention is a reliable predictor of actual behavior; 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; additionally, one’s attitude and subjective norms are influenced by behavioral beliefs and normative beliefs (Ajzen & Fishbein, 1977 , 1980). Empirical research studies suggest that behavioral intention can be a reliable predictor of various behaviors such as coupon use (Bagozzi, Baumgartner, & Yi, 1992; Shimp & Kavas, 1984), condom use (Albarracín, Johnson, Fishbein, & Muellerleile, 2001), class attendance (Ajzen & Madden, 1986), and knowledge sharing (Bock, Zmud, Kim, & Lee, 2005). In the context of a research study about eSports viewership, it is likely that a person with a strong intention to watch eSports may actually watch a match. Given the pivotal role of behavioral intention in the TRA framework and the intention’s predictive power of actual behavior, the current study focuses on examining factors that relate to behavioral intention.

Sports Consumption Motivators and Behavioral Beliefs

Attitude, an essential part of the TRA framework, is determined by behavioral beliefs (Fishbein & Ajzen, 1975). Behavioral beliefs are one’s outcome expectancy of a behavior (Ajzen & Driver, 1991). If an individual deems the behavior of watching eSports can bring him or her the desirable outcome (e.g., relaxation), he or she is more likely to regard the behavior of watching eSports as a valuable behavior and develop a positive attitude toward the behavior. This expectancy-value logic proposed by Fishbein & 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.

U&G theory focuses on audiences of a medium and assumes that audiences’ expectation of using a medium to gratify their needs is the key to motivate them to use the medium (Haridakis, 2002; Papacharissi & Mendelson, 2007). In the U&G-related studies, motivators are often operationalized as latent constructs or variables, reflecting consumers’ desired outcomes that they wish to see or experience after using a medium (Rubin, 2009; Sundar & Limperos, 2013). Thus, motivators can be seen as the equivalent to the behavioral outcome expectation, just as behavioral beliefs function in the TRA framework. Given the similarity between motivators and behavioral beliefs as reflections of behavioral outcome expectancies, it is likely that expectancy-value model can be used to explain how motivators influence attitude formation (Ajzen & Driver, 1991; Fishbein & Ajzen, 1975).

In research studies, behavioral beliefs are operationalized by combining the strength of the beliefs and the evaluation of the outcomes (Ajzen, 2001 , 2002; Fishbein & Ajzen, 1975). For example, the strength of beliefs can be measured by an item in the questionnaire on a 7-point agreement scale such as “Using coupons help me save money.” The corresponding outcome evaluation item can be “Saving money to me is (1) bad…(7) good” (Shimp & Kavas, 1984). Then, the strength of beliefs will be multiplied by the evaluation of outcomes to form a composite item. If there are 3 composite items to measure the construct, all 3 composite items will be added up to form a behavioral belief construct. The mathematical expression of the process is ∑ i = 1 n b i e i . Empirical research has examined why college-aged consumers passed along online video advertisements using the TRA framework (Lee, Ham, & Kim, 2013). The researchers adapted Interpersonal Communication Motives Scale in the study and used motivators revealed in the scale as behavioral beliefs. According to the study, behavioral beliefs of gaining pleasure and escaping from reality influenced consumers’ attitude toward passing along interesting video ads on the Internet. In a similar vein, identifying one's motivation to consume eSports would help the researcher uncover factors related to behavioral beliefs of watching eSports.

The behavior of watching eSports shares common traits with the behavior of watching traditional sports. For example, both eSports and traditional sports matches can be very competitive and suspenseful (Brown et al., 2017; Hamari & Sjöblom, 2017). Sometimes, the result of a game cannot be determined until the very last second. Given the similarity between eSports and traditional sports, it would be helpful for the researcher to first identify motivations of sports consumptions. When these motivations are identified, they can be used as behavioral belief constructs in the study of eSports under the guidance of the TRA.

Yoshida, Gordon, Nakazawa, and Biscaia (2014) discussed sports consumption from the perspective of fan engagement. In Yoshida et al., fan engagement positively associated with intentions of buying tickets and team merchandise, while basking in reflected glory (BIRGing) and team identification were more influential factors than positive affect in influencing fan engagement. Findings from this study emphasize the sequential correlation between team identification, performance tolerance, and purchase intention. From a sports media consumption perspective, Pritchard and Funk (2006) uncovered a media-dominant patronage model: Heavy sports media consumers, not frequent live event attendees, are more involved with the sports, more likely to purchase team-related merchandise, and more tolerant to view sports-related advertising. Findings from this study reveal the significant role played by the media in influencing a team’s fan base and fandom behaviors.

In research studies about traditional sports consumption, Trail and James (2001) have developed the Motivation Scale for Sport Consumption (MSSC) based on existing research studies of sport sociology. The scale is an integration and expansion of two scales developed by Wann (1995) and Milne and McDonald (1999) about motivations of being sports fans. Wann’s (1995) scale is based on the conceptual literature of sports sociology, and Milne and McDonald (1999) developed their scale based on Sloan (1985) and Maslow (1943). The MSSC contains constructs that are similar to U&G theory but specifically designed for studying gratification of needs in sports consumption. As discussed earlier, both motivators and behavioral beliefs are reflections of one’s behavioral outcome expectations. Thus, motivators can be utilized as behavioral beliefs in TRA to study the formation of attitude. An individual can hold an infinite number of behavioral beliefs, but they can only attend to a limited amount of them at a moment for a specific action (Miller, 1956). This study identified six relevant traditional sports consumption motivators from the MSSC and used them as behavioral beliefs to explore how these factors may correlate with the formation of the attitude toward watching eSports.

Achievement is a viewer’s sense of relatedness to a team or player when the team or player wins. Empirical research suggests that sports spectators experience a sense of achievement through BIRGing (Wann, Schrader, & Adamson, 1998). Sports fans vicariously related themselves with the victory of a sports team/athlete for a purpose of maintaining or enhancing their self-images (McDonald, Milne, & Hong, 2002). In the study conducted by McDonald, Milne, and Hong (2002) about sports spectators and participant markets, achievement was positively associated with the viewership of all types of sports (e.g., baseball, basketball, bowling, football, tennis). The finding suggests that sports spectatorship is inherently achievement oriented (McDonald et al., 2002). A fan of eSports may watch eSports games for the same reason. The person may believe that one of the outcomes of watching eSports games is to experience a sense of achievement vicariously. The stronger the belief, the more likely the person will have a positive attitude toward watching eSports.

Hypothesis 1: Achievement is positively correlated with attitude toward watching eSports.

Aesthetics is related to the beauty or gracefulness inherent in sports games. Stylistic sports, such as figure skating and gymnastics, are often the ones that attract viewership through artistic beauty in an athlete’s movements. Wann, Grieve, Zapalac, and Pease (2008) compared motivators of sports consumption between different types of sports. Their findings revealed that aesthetics is a key motivator differentiating stylistic sports consumption motivations from the others. In their research study, the measurement scores for aesthetics motivation to watch figure skating and gymnastics were much higher than the scores of the same constructs for nonstylistic sports. However, one should note that aesthetics is not an exclusive motivator for stylistic sports consumption as those who watch nonstylistic sports, such as golf, often talk about their appreciation for the aesthetic execution of movements (Sargent, Zillmanm & Weaver, 1998). In eSports, a game with well-designed graphics is likely to provide viewers with an aesthetically appealing experience. A professional eSports player who controls the in-game character and executes a movement can trigger visually appealing special effects. Thus, the viewers of eSports games who have seen such visual effects may find the experience aesthetically satisfying, and the viewers may perceive eSports games that provide them such experience in a positive way.

Hypothesis 2: Aesthetics is positively correlated with attitude toward watching eSports.

Acquisition of knowledge is the degree to which media consumption enables one to acquire information about a specific type of sports. Learning the information of a specific team or player is cognitive motivation for some spectators to watch traditional sports (Gantz & Wenner, 1995). James, Kolbe, and Trail (2002) surveyed 507 season ticket holders of a new Major League Baseball franchise to examine what motivated this group of sports fans to watch baseball matches. The results of the study suggest that acquisition of knowledge is an important driver for fans who have a strong psychological connection with the team and who purchase season tickets. Hamari and Sjöblom (2017) conducted an online survey and sampled eSports viewers from the platforms of Reddit, Facebook, Twitter, and other game-related forums. The acquisition of knowledge was found to be one of the positive predictors of eSports viewership in their study. This indicates that, for eSports viewers, watching eSports games is a way to acquire information about teams/players and learn their ways of play (Hamari & Sjöblom, 2017). Therefore, eSports viewers may regard the behavior of watching eSports positively, if they believe that watching eSports helps them learn useful knowledge about the game, player, and player’s style.

Hypothesis 3: Acquisition of knowledge is positively correlated with attitude toward watching eSports.

Drama is about the enjoyment of uncertainty and suspense that happened in sports games. Trail and Kim (2011) surveyed club members of National Collegiate Athletic Association (NCAA) basketball teams (i.e., fans of collegiate basketball teams in the United States) to investigate why they attend women’s basketball games at the university. Drama, together with achievement, escape, social interaction, aesthetics, role model, and support for women’s sports, was found to be an internal motivator that explained 5–22% variance in club members’ intention to attend a women’s basketball game. Knobloch-Westerwick, David, Eastin, Tamborini, and Greenwood (2009) used suspense theory to investigate the formation of perceived suspense in college football games. They found that positive affect (e.g., excitement and joyfulness) and negative affect (e.g., frustration and angriness) influenced the level of suspense a spectator of the game experienced when the person’s favorite team was winning or losing. Their findings are meaningful in that they investigated and revealed psychological affects as sources of perceived suspense in a sports game. In another study conducted by Trail, Fink, and Anderson (2003), the researchers surveyed university students and employed MSSC as the survey instrument. The study by Trail et al. was among one of the few early endeavors that used MSSC motivators to investigate sports spectatorship. They found that drama, as well as other MSSC motivators, is an indicator of potential future sports consumption behaviors. Similar to traditional sports, the results of many eSports matches cannot be determined until the last few seconds of a match. Thus, the belief that watching eSports games will bring an individual the thrill of suspense may not only motivate the individual to watch eSports but also influence one’s attitude toward watching eSports in a positive way.

Hypothesis 4: Drama is positively correlated with attitude toward watching eSports.

Escapism stands for the escape of an individual from one’s stress and other bothersome daily activities in life. Haugh and Watkins (2016) examined motivators of social media sports consumption. According to the findings from Haugh and Watkins (2016), affective motivations, such as escapism, were related to frequent usage of Instagram and Snapchat for sports, while informational motivators were related to Facebook and Twitter usage for sports. Weiner and Dwyer (2017) investigated motivators of fantasy sports consumption. The study categorized fantasy sports consumers in three categories: daily users, season-long users, and traditional users (i.e., traditional demographics of fantasy sports users such as males). Only daily fantasy sports users (i.e., heavy users) considered escapism as an important motivator behind their usage behaviors. In eSports, Weiss and Schiele (2013) used semistructured interviews and survey to examine hedonic (e.g., escapism, social relationship, fun) and competitive motivators (competition and challenge) that influence eSports usage. The finding revealed escapism as the only hedonic determinant of eSports usage. Hamari and Sjöblom (2017) also found escapism as a positive predictor of eSports viewership in their study. In contrast, Brown, Billings, Murphy, and Puesan (2017) did not discover a similar pattern between escapism and fandom in the context of both eSports and traditional sports. However, one should note that fandom and viewership are conceptually different constructs. In fact, the survey respondents rated escapism with a higher score in the eSports condition than in the traditional sports condition (Brown et al., 2017). It is likely that viewers of eSports may consider the behavior of watching eSports games as a way to help divert their attention from problems in their life. Those who value the sense of escapism provided by watching eSports are likely to develop a positive attitude toward the behavior.

Hypothesis 5: Escapism is positively correlated with attitude toward watching eSports.

Social interaction represents the need of sports fans to interact with other fans when they are consuming sports. Individuals are likely to regard sports viewership as a social activity because watching sports provides conversational topics and helps them bond with others (Gantz & Wenner, 1995). Empirical studies using MSSC suggest that sports fans are likely to interact with other fans when they consume sports-related content or attend live events (Trail, Fink, & Anderson, 2003 ; Trail & James, 2001 ; Trail & Kim, 2011). McDonald et al. (2002) found social interaction as one of the factors motivating sports fans to watch sports on television or to attend live events. In eSports, Brown et al. (2017) compared eSports consumption motivators with traditional sports consumption motivators and reported similarities and differences between motivators of eSports and traditional sports consumption. Brown et al. (2017) discovered that social interaction is a significant motivator for both eSports and traditional sports consumption, indicating eSports consumption and traditional sports consumption are quite similar in the social–motivational aspect. Hence, the person who believes that watching eSports is a good social opportunity will perceive the behavior in a positive way.

Hypothesis 6: Social interaction is positively correlated with attitude toward watching eSports.

Two original MSSC constructs, physical attractiveness and physical skills, were not examined in the current study. The physique of the athlete and the athlete’s physical ability to perform an action are at the center of the traditional sports broadcasts. In contrast, eSports spectators’ focus is on the video game but not the professional eSports player. As a reflection of viewers’ preference, the broadcast of eSports is usually game centered. Although the television or online broadcast may show the face of the player occasionally in full screen or in a small corner on the screen, viewers’ foci may remain on the game. For a similar reason, it is not common to see a broadcast of eSports with cameras focusing on the hand of a professional eSports player to display how this person controls the in-game characters.

Attitude Toward eSports Viewership

TRA postulates behavior as a function of perceptions relevant to the behavior (Ajzen, 1991; Ajzen & Fishbein, 1977). Ajzen and Fishbein (1980) proposed that the more positive one’s attitude toward the behavior, the more likely one behaves in a certain way; however, a less favorable attitude might reduce the likelihood of one to perform a given behavior (Fishbein & Ajzen, 1975). Kim and Hunter (1993) conducted a meta-analysis on attitude–behavior research and found a strong and positive correlation between attitude and behavioral intention. In traditional sports, a sports fan’s attitude toward sports teams has an impact on his or her decision to watch sports games (Mahony & Moorman, 2000). Respondents in the study of Mahony and Moorman (2000) expressed their preferences for watching their favorite teams’ games, as well as their preference to watch the game featuring their liked and disliked athletes. The findings suggest that sports practitioners can market an athlete by eliciting either positive or negative attitudes from fans since both attitudes lead to an increased likelihood to watch a game. Drayer, Shapiro, Dwyer, Morse, and White (2010) interviewed 13 fantasy sports fans. They found a reciprocal relationship between fantasy sports participation and fans’ attitude toward the National Football League (NFL). Fantasy sports participation positively correlates with fans’ attitude toward the NFL, while the attitude toward the NFL further influences future fantasy sports participation. The results of the study suggest that fantasy sports experience enhances one’s NFL-related consumption experience. Funk and James (2001) categorized sports fans’ psychological connection with sports in four dimensions from a low level of awareness to higher levels of attraction, attachment, and allegiance. A positive attitude toward sports is at the attachment and allegiance level, representing a unique psychological value attached to the sports in the eyes of sports fans. If such positive attitude would persist, the fan might be attitudinally loyal to the sports, which is likely to influence fans’ behavioral intention to keep watching the sports (Funk & James, 2001). In a similar vein, if a person has a positive attitude toward eSports games, he or she may intend to watch more games in the long term. Therefore, the following hypothesis is proposed.

Hypothesis 7: Attitudes toward watching eSports positively correlate with behavioral intention to watch eSports.

Subjective Norms, Normative Beliefs, and Attitude

In addition to attitude, the subjective norm is another factor that correlates with behavioral intentions under the TRA framework. Ajzen and Fishbein (1980) defined subjective norms as an individual perceived social pressure of performing or not to performing a behavior. The correlation between subjective norms and behavioral intention has been extensively discussed in research studies (Ajzen, 1991; Hansen Jensen, & Solgaard, 2004; Ryan & Bonfield, 1980; Vermeir & Verbreke, 2006). Ryan and Bonfield (1980) found a positive correlation between family, friends, and neighbors’ influence on one’s purchase intention of a product. Empirical research also reveals that perceived social norms (social pressures) directly related to consumers’ intention to buy sustainable food (Vermeir & Verbreke, 2006). In the domain of online grocery shopping, researchers find a positive and direct association between subjective norms and purchase intentions (Hansen et al., 2004). Regarding sports merchandise consumption, group norms are believed to be a significant predictor of purchase intentions among college football fans (Madrigal, 2000). The findings from Madrigal’s (2000) study implicate an association between social pressure and one’s group identity, and how the pressure and identification positively influence in-group behaviors such as buying team-related merchandises. For a behavior such as watching eSports, social pressures or subjective norms may influence one’s decision on which eSports game he or she needs to watch because the individual wishes to assimilate to the social group and avoid of being left out in the group. Thus, the subjective norms’ correlation with behavioral intention may be positive. Hence, the following hypothesis is posited.

Hypothesis 8: Subjective norms positively correlate with behavioral intention to watch eSports.

On the other hand, normative beliefs are about one’s beliefs in whether the important referent individuals would support or disapprove the behaviors that they intend to perform (Ajzen & Driver, 1991). Similar to how behavioral beliefs affect one’s attitude, normative beliefs are determinants of one’s subjective norms (Ajzen & Fishbein, 1977). In operationalization, normative beliefs are the products of the multiplication between the strength of beliefs and motivation to comply ∑ i = 1 n n i m i (Ajzen & Fishbein, 1977, 1980). The process is similar to the demonstration of how behavioral belief is operationalized in the previous section. Existing studies reveal that one’s normative beliefs affect subjective norms about adopting new brands (Ryan, 1982), improving school performance (Park, 2000), and making a moral judgment (Vallerand, Deshaies, Cuerrier, Pelletier, & Mongeau, 1992). It is possible that the stronger one’s beliefs that his or her friend would approve/disapprove of the behavior of watching eSports, the more subjective norms (or social pressures) one perceives.

Hypothesis 9: Normative beliefs of watching eSports positively correlate with subjective norms of watching eSports.

Empirical research findings suggest that subjective norms relate to attitude formation (Bock et al., 2005; Chang, 1998; Ryan, 1982; Shepherd & O’keefe, 1984). For example, if one’s social connections, such as friends, hold a positive attitude toward buying organic food, the person may develop a similar positive attitude toward the behavior (Tarkiainen & Sundqvist, 2005). Peer-pressure has proven to be one of the predictors of opinion formation (Asch, 1955; Zimbardo & Leippe, 1991). These significant others in one’s life may exert a greater influence on the formation of one’s attitude or opinion than strangers examined in some of the laboratory experiments (Asch, 1955). Furthermore, individuals obtain some of the behavioral beliefs from their significant others that these beliefs also affect the formation of one’s attitude (Chang, 1998). In the context of eSports, an individual may learn from their friends about an interesting eSports game that he or she should play and watch. Group members’ opinions are likely to influence this individual. If everyone in this person’s social circle considers the game is good, the person may as well develop a similar positive attitude toward this eSports game. Likewise, if all group members consider that the behavior of watching eSports is socially acceptable, the individual will also hold a positive attitude toward the behavior of watching eSports.

Hypothesis 10: Subjective norms positively correlate with attitude toward watching eSports.

Data Collection

The researcher administered an online survey on Amazon’s Mechanical Turk (MTurk) to investigate factors influencing the behavior of watching eSports. A pilot study (n = 41) was first carried out to ensure the validity and reliability of constructs in the survey. The wording of some of the questions was refined based on the feedback of pretest. A total of 341 respondents were reached in the survey. Forty respondents who failed the screening question (whether they have watched eSports before) were removed. Additionally, those who were over 60 years old (n = 6) were removed as they are not falling in the typical demographics of eSports viewers (GlobalWebIndex, 2018; The Nielsen Company, 2017 , 2016). The final sample size was n = 295. The participants were rewarded with a monetary incentive (US$1) after the completion of the survey. All participants were from the United States since the researcher selected the geographic limitation as one of the requirements before dispatching the MTurk request. This is helpful to prevent bots and other suspicious MTurkers from completing the survey because U.S. MTurkers have to input their social security numbers when creating the account, and their financial gains on MTurk are monitored by the Internal Revenue Services of the United States. The requirement for the Human Intelligence Task (HIT) approval rate was greater than 95%, and the requirement for the number of HIT approved was greater than 1,000.

Sample

Male respondents (79%, n = 233) outnumbered female respondents (21%, n = 62). The respondents were mostly White (64.1%, n = 189), following by Asian (16.6%, n = 49), Black (9.5%, n = 28), Hispanic/Latino (7.1%, n = 21), and Native American (1.4%, n = 4). About 56.8% of respondents (n = 167) held a bachelor’s degree or above, and 31.9% of respondents (n = 94) had some college experience or held a 2-year associate degree in college. The average annual household income of respondents fell in the range from US$ 40,000 to US$ 49, 999. The age of respondents ranged from 18 to 54 (M = 31, SD = 7.62) with 77.6% of respondents 34 years old or younger. The sample of the survey is representative of the eSports viewer demographics, as eSports viewers are predominantly male (71%), under 34 years old (73%), and mostly with mid or below mid-level income (71%; GlobalWebIndex, 2018; The Nielsen Company, 2017 , 2016).

The survey also asked respondents to rate how much do they like to use different media devices or platforms to play or watch eSports on a 7-point scale. The result suggested that computer (M = 5.04, SD = 1.76) was respondents’ favorite device to play or watch eSports games. YouTube (M = 5.05, SD = 1.73) was respondents’ favorite media platform to watch eSports matches, and the Call of Duty series (M = 3.89, SD = 2.19) were among the favorite games of respondents.

Measures

Intention to watch eSports was measured on 3 items adapted from Ajzen and Fishbein (1980). All items were measured on a 7-point scale, ranging from extremely unlikely (1) to extremely likely (7; α = .905). Attitude toward watching eSports was measured on 4 semantic differential items adapted from Ajzen (2002; α = .867). In total, 4 items adapted from a study of smartphone app usage and technology adoption were used to measure digital communication related subjective norms (Verkasalo, López-Nicolás, Molina-Castillo, & Bouwman, 2010). The items were measured on a 7-point scale ranging from strongly disagree (1) to strongly agree (7; α = .840).

The behavioral belief items were developed based on Trails and James’s (2001) MSSC. The strength of individuals’ perception of each belief was measured on a 7-point scale from strongly disagree (1) to strongly agree (7). Then, items measuring the strength of behavioral beliefs were revised to measure outcome evaluations (Vallerand et al., 1992). The outcome evaluation items were measured on a 7-point scale ranging from very bad (1) to very good (7). As demonstrated earlier in the literature review, the strength of each behavioral belief (b) was multiplied by its corresponding outcome evaluation (e), following the operationalization guidance of TRA (Ajzen, 1991; Fishbein & Ajzen, 1975). The resulting products were summed over n items that measure a behavioral belief. The mathematical expression of the process is ∑ i = 1 n b i e i (e.g., (4 × 5) + (5 × 6) + (6 × 7) = 92; Ajzen & Driver, 1991; Ajzen & Fishbein, 1980; Wigfield & Ecceles, 2000).

The strength of normative beliefs was measured on items adapted from Ajzen and Fishbein (1980) and Lee, et al. (2013). Motivation to comply items were revised based on the items measuring the strength of normative beliefs. All items were measured on a scale from 1 to 7 (strongly disagree to strongly agree; Ajzen & Fishbein, 1980; Vellerand et al., 1992). The composite normative belief variable was processed in a similar way as how the researcher processed behavioral beliefs (Ajzen & Driver, 1991; Ajzen & Fishbein, 1980; Wigfield & Ecceles, 2000). Specifically, the strength of each normative belief item (n) was multiplied by the motivation to comply item (m), and the resulting products for each reference group (friends/family/people whose opinion I value) were summed to create a normative belief construct ( ∑ i = 1 n n i m i ; see Table 1 for item descriptions and descriptive statistics).

OPEN IN VIEWER

Table 1.

Descriptive Statistics Measurement Items.

	Mean	SD
Behavioral Belief Strengths
Achievement
Feel like I have won when the eSports teams/players I support wins	4.48	1.923
Feel a personal sense of achievement when the eSports teams/players I support does well	4.26	1.881
Feel proud when the eSports teams/players I support play well	4.77	1.792
Knowledge acquisition
I regularly track the statistics of specific eSports teams/players	3.90	1.807
I usually know the eSports player’s or team’s win/loss record	3.87	1.820
I read the eSports player’s or team’s statistics regularly	3.76	1.832
Aesthetics
I appreciate the inherent beauty in eSports	4.88	1.610
There is a certain natural beauty to eSports	4.77	1.628
I enjoy the gracefulness associated with eSports	4.70	1.682
Drama
I enjoy the drama of an eSport match with fierce competition	5.45	1.425
I prefer a close eSport match rather than a “one-sided” match	5.54	1.551
An eSport match is more enjoyable to me when the outcome is not decided until the very end	5.54	1.456
Escapism
Watching eSports represent an escape for me from my day-to-day activities	4.85	1.548
Watching eSports are a great change of pace from what I regularly do	4.89	1.582
I look forward to watch eSports because they are something different to do	5.00	1.482
Social interaction
Interacting with other fans is a very important part of watching eSports matches	4.12	1.845
I like to talk to other people sitting near me when I’m attending live eSports events	4.09	1.837
Watching eSports are great opportunities to socialize with others	4.56	1.650
Behavioral outcome evaluations
Achievement
Feeling like I have won when the eSports teams/players I support wins would be for me (1) very bad…(7) very good	4.75	1.597
Feeling personal sense of achievement when the eSports teams/players I support does well would be for me…	4.63	1.697
Feeling proud when the eSports teams/players I support play well would be for me…	4.82	1.595
Knowledge acquisition
Tracking statistics of specific eSports teams/players would be for me…	4.20	1.584
Knowing the eSports teams’ or players’ win/loss record would be for me…	4.27	1.584
Reading eSports teams’ or players’ statistics regularly would be for me…	4.17	1.623
Aesthetics
Appreciating the inherent beauty in eSports would be for me…	4.92	1.545
Knowing a certain natural beauty in eSports would be for me…	4.86	1.475
Enjoying the gracefulness of eSports would be for me…	4.97	1.484
Drama
Enjoying the drama provided by the fierce competition in an eSport match would be for me…	5.39	1.333
Watching a “close” eSport match rather than a “one-sided” eSport match would be for me…	5.53	1.479
Enjoying the unpredictable outcome of an eSport match would be for me…	5.48	1.484
Escapism
Escaping from my day to day activity by watching eSports would be for me…	5.05	1.497
Changing pace from what I regularly do by watching eSports would be for me…	5.02	1.457
Having something different to do by watching eSports would be for me…	5.05	1.504
Social interactions
Interacting with other fans when watching eSports matches would be for me…	4.45	1.729
Talking with others sitting near me during live eSports events would be for me…	4.45	1.729
Socializing with others when watching eSports would be for me…	4.65	1.675
Normative belief strengths
My close friends think that I need to watch eSports	3.52	1.855
My family thinks that I need to watch eSports	2.83	1.827
The person, whose opinion I value, think that I need to watch eSports	3.50	1.851
Motivation to comply with normative belief
I want to do what my close friends think I should do	3.52	1.771
I want to do what my family thinks I should do	3.69	1.821
I want to do things that people, whose opinion I value, think I should do	4.29	1.768
Subjective norms
I want to watch eSports because my friends do so, and I want to belong to the group	3.29	1.892
Watching eSports reflects my personality to other people	4.11	1.776
According to people who are important to me, I should watch eSports	3.51	1.909
Attitude: For me, watching eSports is…
Extremely bad…Extremely good	5.59	1.186
Extremely worthless…Extremely valuable	5.00	1.362
Extremely unpleasant…Extremely pleasant	5.66	1.251
Extremely Boring…Extremely Interesting	5.70	1.203
Behavioral intention
I plan to watch eSports	5.56	1.403
I will make an effort to watch eSports	5.28	1.509
I intend to watch eSports	5.54	1.392

Note. SD = standard deviation.

Data Analysis

MPlus 7 (version 7.4) was used to process the data by implementing linear SEM. After the removal of six outliers who were over 60 years old, no new outliers were identified. The Variance Inflation Factor (VIF) scores for all variables were between 1.95 and 3.26, which were well below the threshold of 5, indicating the absence of multicollinearity (Hair, Black, Babin, & Anderson, 2014). Parameters of models were estimated using maximum likelihood estimation with robust standard errors (MLR) method. This method is robust to data nonnormality; thus, the impact of data nonnormality on analytic results should be mitigated (Muthén & Muthén, 2018). No missing data was reported. The χ² test, root mean square error of approximation (RMSEA), comparative fit index (CFI), and standardized root mean square residual (SRMR) were tested and inspected as model fit indices. According to Kline (2016), these model fit indices are the ones that reported by most studies because this group of indices offers researchers a comprehensive view on model fit. The researcher followed an absolute-fit data analytic philosophy. Hence, the researcher’s decision of retaining or rejecting a model was largely based on two absolute-fit indices, RMSEA and SRMR, while other fit indices were also considered during the evaluation of model fit. Hu and Bentler (1999) proposed the two-index evaluation rules and suggested that RMSEA with a value of .06 or lower and the SRMR with a value of .08 or lower were signs of very good model fit.

Measurement Model

The overall model fit indices suggested a good fit of the measurement model. The χ² test value was χ² = 704.845, p < .001, df = 389, while the RMSEA = .052 with a 90% CI [.046, .059], which indicated a good fit of the model (Hooper, Coughlan, & Mullen, 2008). The CFI value was .948 and the SRMR = .068, which suggested a good fit of the model (Hu & Bentler, 1999). Confirmatory Factor Analysis (CFA) results showed that behavioral beliefs and other TRA-related dimensions were well identified. All indicators were confirmed as the measurement of corresponding latent variables, and all standardized factor loadings were significant and sizable (average λ _standardized = .848; see Table 2). Furthermore, the analysis showed that all measured constructs were reliable with Cronbach’s α values ranging from .803 to .927. In general, the measurement model was confirmed and could be used in the test of the model.

OPEN IN VIEWER

Table 2.

CFA Result of Measurement Model for Watching eSports.

Construct and Indicators	Standardized Factor Loading	Standard Error	α
Behavioral beliefs ∑ i = 1 n b i e i
Achievement
Ach 1	.909***	.018
Ach 2	.893***	.019	.927
Ach 3	.898***	.023
Knowledge acquisition
Knw 1	.902***	.021
Knw 2	.857***	.027	.915
Knw 3	.895***	.019
Aesthetics
Aesth 1	.918***	.024
Aesth 2	.902***	.017	.925
Aesth 3	.873***	.022
Drama
Dram 1	.767***	.063
Dram 2	.718***	.068	.803
Dram 3	.787***	.063
Escapism
Esc 1	.804***	.031
Esc 2	.860***	.025	.884
Esc 3	.879***	.023
Social interaction
Soc 1	.862***	.025
Soc 2	.890***	.020	.905
Soc 3	.864***	.026
Normative beliefs ∑ i = 1 n n i m i
Nom 1	.901***	.014
Nom 2	.807***	.022	.903
Nom 3	.901***	.014
Subjective norms
SN 1	.860***	.022
SN 2	.677***	.039	.840
SN 3	.870***	.026
Attitude toward watching eSports
Att 1	.791***	.031
Att 2	.678***	.043	.867
Att 3	.837***	.028
Att 4	.868***	.022
Intention to watch eSports
Int 1	.853***	.030
Int 2	.888***	.021	.905
Int 3	.879***	.027
Goodness of fit statistics
χ2(df)	704.845 (389)
RMSEA	.052
CFI	.948
SRMR	.068

Note. b _i = strength of each behavioral belief; e _i = outcome evaluation; n _i = strength of each normative belief; m _i = motivation to comply with each reference; b _i e _i = combined statement of strength of each behavioral belief and its outcome evaluation; n _i m _i = combined statement of strength of each normative belief and motivation to comply with each reference. RMSEA = root means square error of approximation; CFI = comparative fit index; SRMR = standardized root means square residual.

***p < .001

Structural Model

For the SEM model, the fit indices showed a good model fit (χ² = 756.502, p < .001, df = 403; RMSEA = .055, 90% CI [.049, .060]; CFI = .942; SRMR = .071). Hypotheses were then tested. The results were reported in Figure 2, and construct correlations were reported in Table 3. Hypotheses 1–6 posited that beliefs of achievement, aesthetics, acquisition of knowledge, drama, escapism, and social interaction have a positive correlation with attitude toward watching eSports. Results demonstrated that aesthetics (γ = .163, SE = .068, p < .05), drama (γ = .469, SE = .099, p < .001), and escapism (γ = .242, SE = .091, p < .01) had a statistically positive relationship with attitude. Thus, Hypotheses 2, 4, and 5 were supported. Hypothesis 7 postulated that the attitude toward watching eSports has a positive correlation with the intention to watch eSports. The result indicated a statistically significant correlation between the attitude toward watching eSports and intention to watch eSports (β = .832, SE = .031, p < .001). Hypothesis 9 proposed that normative beliefs positively correlate with subjective norms. The result supported the hypothesis (γ =.921, SE = .022, p < .001). Hypothesis 10 postulated a positive correlation between subjective norms and attitude. The analytic result (β = .242, SE = .063, p < .001) confirmed the hypothesized correlation in Hypothesis 10. Hypothesis 8 predicted that subjective norms positively correlate with intention to watch eSports. However, the result did not support Hypothesis 8 (β = .073, SE = .047, p > .05). No other significant correlations were discovered (see Figure 2).

OPEN IN VIEWER

Figure 2.

The results of the hypothesized model. *p < .05 **p < .01 ***p < .001; χ² = 756.502, p < .001, df = 403; RMSEA = .055; CFI = .942; SRMR = .071

OPEN IN VIEWER

Table 3.

Means, Standard Deviations, and Correlation Coefficient of All Constructs.

	1	2	3	4	5	6	7	8	9	10
1. Achievement	1
2. Knowledge acquisition	.633**	1
3. Aesthetics	.632**	.542**	1
4. Drama	.403**	.246**	.469**	1
5. Escape	.654**	.478**	.578**	.554**	1
6. Social interaction	.561**	.572**	.513**	.199**	.471**	1
7. Normative beliefs	.320**	.524**	.223**	−.077	.120*	.470**	1
8. Subjective norms	.339**	.517**	.278**	−.095	.195**	.558**	.793**	1
9. Intention	.602**	.540**	.567**	.558**	.633**	.450**	.264**	.317**	1
10. Attitude	.580**	.492**	.612**	.572**	.627**	.443**	.280**	.316**	.731**	1
Mean	70.67	54.50	75.33	94.27	78.75	64.26	42.42	3.64	5.46	5.49
SD	40.11	35.84	37.19	33.38	34.86	37.16	32.87	1.62	1.32	1.05

Note. The means of constructs are listed in the table.