The influence of social TV on viewer loyalty of National Hockey League (NHL) viewers: ice hockey’s second screen and social media engagement behaviour

By Dr. Sebastiano Mereu on October 15, 2021 • ( 1 Comment )

Engaging with others on social media while watching television (TV) or live-streaming any sort of entertainment has become a common phenomenon. It is occasionally referred to as second-screening or social TV, because the main viewing activity of a programme is combined with interaction with other users on a second device, such as a mobile phone, tablet, or laptop. Such engagement has been found to positively affect ratings of certain programmes (Lim et al., 2015). This study examines whether social media engagement activities of viewers of the National Hockey League (NHL) positively influence their viewing intentions and loyalty towards the sports product, namely following the championship or their favourite team. It further offers recommendations on how to foster and reinforce viewing intentions for the NHL using social second screen tactics. A random sample of users that follow the NHL on social media (n=216), mainly from North America, were surveyed during the Stanley Cup playoffs 2021. A self-administered online questionnaire was adapted from Lim et al. (2015) with the aim to assess the influence of functional, emotional, and communal engagement on viewer loyalty towards the NHL product, as well as the mediating roles presented by social presence of viewers and their commitment to the league. Results from the structural equation model (SEM) show that only communal engagement directly influences viewers’ loyalty. The other independent variables, functional and emotional engagement, lead to loyalty through mediation of social presence and commitment. Social presence was not found to directly affect viewers’ loyalty.

Section 1: Sports viewers and second screens (background to the research)

Watching a programme on television or via over-the-top (OTT) streaming services is generally considered a passive experience (Pine and Gilmore, 1999). A social aspect can be added by inviting companions to watch and discuss the programme; receiving acknowledgement from other viewers for emotions expressed in response to the happenings on the screen is another benefit of socialising while watching the offered programme (Rubin, 1983). This can be applied to a variety of broadcasted programme types, such as reality TV, the news, and sports.

Rein, Kotler and Shields (2006) emphasize, “Sports are a universal subject of communication […] and serve as a common bond between groups of friends, communities, business and professional relationships across all geographical regions and economic classes” (p. 59). Before social media, conversations between sports spectators and fans were limited to living rooms, sports bars, or the watercooler, unless these spectators and fans were attending a game in the stadium. However, the technological landscape and respective user behaviours have started to change with the advent of mobile phones and the evolution of networked media in the 1990s and 2000s (Jenkins, 2006). This led to the emergence of conversational channels like Twitter, Instagram, or WhatsApp, which then enabled spectators and fans to get closer to a sporting event and discuss the happenings virtually with others watching the same event in the venue or remotely via a broadcast (Sutera, 2013). Such virtual conversations are now fuelled by a so-called second screen, usually a mobile phone, tablet, or laptop, and result in social television (TV). Guo (2019) defines social TV as “the degree of intensity or types of connections that audiences develop with television content through online or mobile apps via second screen platforms” (p. 140). Second-screening, on the other hand, can be defined as “the act of coupling a TV viewing activity with second screen interaction” (Doughty, Rowland, and Lawson, 2021, p. 80). The difference between the two activities is found in that social TV implies a social element to the viewing experience, whereas second-screening would be considered the umbrella term and does not necessarily include the element of socialisation; for example, looking up descriptive information about a game online on a mobile phone, while watching a sporting event does not include a social component.

According to a global survey undertaken in 2019 (Statista, 2021a), 95 percent of online users aged 16 to 21 were very likely to use a second digital media device while watching TV; 72 percent of respondents would use social media while viewing a programme and 71 percent would chat with friends online, making these the most common second screen activities. But second-screening is not only popular among younger generations. AdColony (2021) reviewed data from a global audience collected by Global Web Index and found that “65% of internet users aged 35-44 said they used their mobile phones often while watching television. More than half of users (53%) aged 45-54 also said they second screened with their smartphones. A significant number of Baby Boomers (41%) are still using their mobile devices as second screens and we expect this share to grow as they become more accustomed to multitasking with media.” Hence, second-screening has become a habitual activity among all age groups, also the one’s relevant to this study, namely sports viewers with access to digital communication devices.

In 2021, spectator sports were mostly consumed live or near-live (PwC Sports Survey, 2021). Their real-time advantage forces fans to either go to the stadium in order to maximise the experience or watch that specific sporting event exactly when it takes place through a live broadcast (Rein et al., 2006). This, in contrast to highlights and social media clips, attracts viewers to a screen at the same time, which offers the opportunity for a wholesome social TV experience and heightens fans’ functional engagement on social media. Additionally, live sports carry meaning and values for spectators and fans through their collective social experience that can be intensified by communal rituals (Hutchins, Li and Rowe, 2019). Moreover, sharing emotions on social media, for example, about a team’s performance or referees’ decisions, has become common practice and an established outlet for viewers to connect over (Sanderson, 2011).

From the discussion above, it can be deduced that live sports and virtual conversations among viewers of spectator sports, in this case the NHL, may rely on the social presence of other viewers and the commitment to the sports product, in order to establish and foster viewer loyalty (cf. Lim et al., 2015). Consequently, the leading research question for this research is defined as follows: How does social TV influence viewer loyalty of NHL viewers?

Section 2: Definitions and conceptual framework

The following section introduces the subject of this study and its product, namely the NHL and its season-long championship, then describes the variables evaluated in the research, and defines the hypotheses to be tested. The applied conceptual framework is visualised in Figure 1 at the end of this section.

The NHL and its product

The NHL was founded on 26 November 1917 in Montreal, Canada, as its predecessor, the National Hockey Association of Canada Limited (NHA), suspended its operations due to disputes between franchises executives; the founding members of the NHL included the Montréal Canadiens, Montréal Wanderers, Ottawa Senators and Quebec Bulldogs, who were later joined by the Toronto Arenas as fifth team in the league (NHL.com, 2021a). The NHL expanded to the United States in 1924 with the Boston Bruins becoming the first American team to join the league (NHL.com, 2021b). In 1925, The New York Americans and the Pittsburgh Pirates became the second and third American team respectively to join the NHL (Holzman and Nieforth, 2002). Over the next century various franchises from both Canada and the USA would come and go or move city and change their names. By the 2018-19 NHL season, the league’s last regularly operated season before the forced COVID-19 break and playoffs behind closed doors, the NHL included 31 teams, 7 from Canada and 24 from the USA (NHL.com, 2021a), and it recorded a total attendance of 23.7 million spectators (NHL.com, 2021c). During season 2018-19, viewership in North America across NBC, NBCSN, NBCSports.com and the NBC Sports app grew 2 percent from the previous season to an averaged Total Audience Delivery (TAD) of 424,000 viewers per game; live minutes for the regular season reached a record 248.8 million, which is an increase of 54 percent compared to season 2017-18 (NBC Sports Group, 2019). The Stanley Cup Finals 2019, the final series that crowned the winning team of the NHL championship 2018-19, were viewed by an average 5.33 million viewers in the USA (Statista, 2021b). The Canadian sports specialty channel Sportsnet stated that the Stanley Cup Finals 2021, that featured a Canadian team for the first time since 2011 and were played between the Montréal Canadiens and the Tampa Bay Lightning, became “the most-watched Stanley Cup final in its history” with 4.1 million viewers watching game five and 3.6 million viewers in average across all five games (SportsPro, 2021). These numbers illustrate the popularity of the NHL in North America in recent years.

Although the core product of spectator sports could be defined as an individual game played between two opposing teams (Bühler and Nufer, 2013), a more longitudinal perspective may be applied to a sports league such as the NHL. Hence, the sports product offered by the NHL can be considered the season-long championship that ranks competing teams according to the points they accumulate with wins over other teams or draws, with the continuous goal to improve the standing on the league table and ultimately win the showdown in the final of the playoffs; in the case at hand, it means winning the Stanley Cup (Mastromartino, Wann, and Zhang, 2019).

The characteristics and statistics discussed above in regard to the NHL, exemplify the great interest by spectators and fans in watching or streaming the league’s product live on TV or OTT, which consequently provides an opportunity for social TV. The question remains, however, which type of second screen engagement has an effect on the recurring intention to watch the offered sports product?

Functional engagement

Lim et al. (2015) define functional social media engagement behaviour as “a social media user’s interactions with other users in the process of co-creating, conversing and sharing the content” (p. 159). This behaviour goes beyond a simple online purchase and results from motivational influences (Dolan et al., 2016; Vale and Fernandes, 2018). Functional engagement is often considered the key performance indicator for measuring effectiveness of marketing and communications activities online (Sajjad and Zaman, 2020). For this study, it can be considered that NHL viewers may reply to others’ social media posts (i.e. retweet or share others’ posts), share game-related videos or photos, bring up things about the game in conversations, or use trending words such as hashtags in their social media communication.

Emotional engagement

Emotional social media engagement behaviour can be defined as a viewer’s emotional state, while browsing through social media posts related to the programme the user is watching, and sharing emotions about the programme or conversations around the programme that are happening in real-time (Lim et al., 2015). Schmitt (1999) explains, “People deliver feelings at point of consumption [when experiencing a brand] and use communication to “frame” the consumption experience” (p. 219). For the case at hand, the point of consumption can be considered the live-broadcast of the NHL product. It can therefore be interpreted that the consumption of the entertainment product will evoke feelings in real-time, which viewers may want to express in their communication with others on social media. Furthermore, if users experience emotional attachment to the offered social media content, such as the above-mentioned emotionally framed comments, they are more likely to share the content with their online friends and followers (Sajjad and Zaman, 2020).

Communal engagement

To reiterate the discussion above, watching live sports on TV, via OTT, or in the stadium offers viewers the opportunity to engage through social TV with others and foster a sense of belonging to a community (Lim et al., 2015). This sense of belonging is built upon an immersive brand experience that requires a certain degree of active participation by spectators (Pine and Gilmore, 1999). For this research, active participation involves sharing opinions about the game with other viewers, sharing tweets/posts from NHL teams, players or the league, contributing to online discussions by adding useful information, and interacting with other viewers using the hashtags related to NHL games. Encouraging viewers to engage with community members who watch the same game, can create and strengthen long-term patterns of behaviour, as well as recurring rituals around the viewing experience (Schmitt, 1999).

Lim et al. (2015) define social presence as “the degree to which users perceive one or many others as being present via [a] mediated interface” (p. 160) and add that social presence can be considered a convincing, although reduced, simulation of face-to-face communication. The concepts of immediacy and intimacy play a role in social presence in that immediacy strengthens the relationship of the psychological distance from one person to another and intimacy assists the interpretation of interpersonal interactions (Rettie, 2003). The synchronous just-in-time nature of social media offers users, e.g. sports spectators watching a live NHL game, the occasion to connect with other viewers and to perceive their conversations as “real”; this would enhance the social presence of the community watching the game (Dunlap and Lowenthal, 2009). For this study, it means assessing if second-screening sports spectators feel like they are physically communicating with others, if they feel like they are watching the game with friends, and if they feel like many people are watching at the same time. The discussion above leads to the following hypotheses:

H1: Functional engagement through social TV has a positive effect on perceived social presence.
H2: Emotional engagement through social TV has a positive effect on perceived social presence.
H3: Communal engagement through social TV has a positive effect on perceived social presence.

Commitment

A person’s commitment to a brand or an organisation can be viewed from three perspectives: affective commitment refers to an emotional attachment to, identification with, and involvement in a brand or organization; continuance commitment refers to the cost or effort associated with changing from one brand or organisation to another; and normative commitment refers to peoples’ feelings of obligation towards a brand or organisation to stay with them (Allen and Meyer, 1990). In regard to sports spectators, Guttmann (1986) sees a fan as “the emotionally committed consumer of sports events” (p. 6), which leads us to adopt the definition of affective commitment from Allen and Meyer (1990), as stated above.

Affective commitment exerts a positive influence on brand loyalty (Turri, Smith, and Kemp, 2013). It is therefore of great importance for broadcasting rights owners of sports programmes and their respective advertisers to build and maintain viewer commitment for their product (Hutchins, and Rowe, 2012). This further includes reinforcing viewers’ emotional attachment and connection to, in this case, their favourite NHL team competing for the Stanley Cup or the NHL competition as a whole (Lim et al., 2015). Consequently, further hypotheses to be tested include:

H4: Functional engagement through social TV has a positive effect on the commitment to the NHL or a viewer’s favourite team.
H5: Emotional engagement through social TV has a positive effect on the commitment to the NHL or a viewer’s favourite team.
H6: Communal engagement through social TV has a positive effect on the commitment to the NHL or a viewer’s favourite team.
H10: Social presence has a positive effect on the commitment to the NHL or a viewer’s favourite team.

Viewing intention and viewer loyalty

The desired behaviour assessed in this research is the intention to view the sports product offered by the NHL, i.e. the ice hockey competition that crowns the winners of the Stanley Cup, and the loyalty towards that product established via social TV. In continuation of the discussion above, the following hypotheses to be tested complete the conceptual framework in Figure 1:

H7: Functional engagement through social TV has a positive effect on viewer loyalty.
H8: Emotional engagement through social TV has a positive effect on viewer loyalty.
H9: Communal engagement through social TV has a positive effect on viewer loyalty.
H11: Social presence has a positive effect on viewer loyalty.
H12: Commitment to the NHL or a viewer’s favourite team has a positive effect on viewer loyalty.

Detailed research question

After having reviewed the variables relevant to this study, the research question can be defined in more detail as follows: How does social TV influence viewer loyalty of NHL viewers, where functional, emotional, and communal social media engagement behaviour is mediated through social presence and commitment to the sports product?

Figure 1 provides the conceptual framework applied to answer the research question.

Section 3: Methodology and sample

This study adopted a positivist research philosophy and collected data through a self-administered quantitative online questionnaire. The questionnaire-items to assess social media engagement behaviour and viewer loyalty were adapted from Lim et al. (2015) and include 22 items; see Figure 3 for a list of items and their respective descriptive statistics. Additionally, four questions assessed the demography of participants and their team affiliation. The items were assessed with a 7-point Likert-scale ranging from 1 (disagree very strongly / never) to 7 (agree very strongly / all the time). Possible participants were asked to complete the questionnaire via posts published on Facebook, Twitter, and LinkedIn. The data was collected from 216 respondents during the playoff stage of the NHL season 2020-2021, from Thursday, 3 June to Sunday, 6 June 2021. All questionnaires were valid and, hence, accepted.

An overview of the sample is depicted in Figure 2. The questionnaire was completed by 81% male and 19% female NHL viewers. Respondents are evenly distributed across age groups with 55 to 64-year-olds including 24.5% of the sample, followed by 45 to 54-year-olds with 20.8%, then 35 to 44-year-olds counting 16.2%, 18 to 24-year-olds amounting to 15.7%, and 65+-year-olds with 8.3%. Respondents under 17 years of age amount for 1.4% of the sample. Although the questionnaire was spread across various countries around the world that are known to follow the NHL or express interest in ice hockey in general, most of the respondents live in North America with 68.5% living in Canada and 7.4% living in the United States. These two groups amount for more than three-quarter of the sample. The large number of Canadian respondents is reflected in the team affiliation that sees the largest support for the Toronto Maple Leafs (19.4%) and the Montreal Canadiens (19.0%), followed by support for Edmonton Oilers (6.5%) and Winnipeg Jets (6.5%). The American teams with the largest support from this sample are Colorado Avalanche (5.1%) and the Boston Bruins (4.2%).

In summary, the sample of this research can be considered to mainly represent the views of a more masculine audience from Canada, spread across viewers from different age groups, and that mainly supports Canadian NHL teams.

Figure 2: Demography of the sample (n=216)

Section 4: Results

The question-items from the questionnaire are presented in Figure 3, which also displays their mean, standard deviation (STDEV), factor loading, and the scales’ Cronbach’s Alpha (α), composite reliability (CR), and average variance extracted (AVE). The following section reviews the descriptive findings based on the perception of the responding NHL viewers, i.e. mean and STDEV. The internal consistency and reliability, i.e. α, CR, and AVE, will be addressed with the model fit in Section 6.

Functional engagement (FNC)

In reference to NHL viewers’ functional engagement, respondents show a neutral attitude towards three of the four items and a negative attitude towards one item. Results in Figure 3 show that respondents sometimes do and sometimes do not reply to others’ social media posts (FNC1, M=3.44, STDEV=2.248). Similarly, NHL viewers sometimes share game-related videos or photos (FNC2, M=3.03, STDEV=2.215) or bring up things about the game in conversations on social media (FNC3, M=3.91, STDEV=2.227). Nonetheless, all three items recorded high standard deviation, indicating disagreement among respondents. The last item in this scale shows that respondents do not use trending words such as hashtags while using social media during an NHL game (FNC4, M=2.60, STDEV=1.930). The moderate standard deviation for this item suggests relative agreement among respondents. It can be inferred that respondents do not necessarily make use of the social functionalities offered by social media, hence, this may be improved with dedicated marketing and communication activities.

Emotional engagement (EMO)

All four items in this scale recorded a neutral stance in reference to emotional engagement, however, with high standard deviation for all of them. Respondents sometimes post and sometimes they do not post their feelings in real-time online conversations (EMO1, M=3.03, STDEV=2.181), which also applies to posts about their feelings when they like or dislike the NHL game they are watching (EMO2, M=3.13, STDEV=2.171). Similarly, they sometimes quote or retweet from others when it is good or witty (EMO3, M=3.67, STDEV=2.397), and sometimes they may express their feelings about the NHL (EMO4, M=3.76, STDEV=2.292). These results show that emotional engagement of NHL viewers through social TV can be considerably improved, since respondents do not necessarily engage with affective messages on social media while watching the NHL. Yet, all four items recorded among the highest standard deviation across the questionnaire, implying that there is considerable disagreement among participants of this survey.

Communal engagement (CME)

Respondents recorded a neutral stance for two of the four question-items and a negative stance with the other two items of this scale. The collected data shows that NHL viewers may sometimes share their opinions about the game with other viewers on social media (CME1, M=3.89, STDEV=2.245) and they may sometimes share tweets or posts from NHL teams, players or the league (CME2, M=3.22, STDEV=2.211); yet, both items recorded a high standard deviation. The other two items, however, show that viewers do rather not contribute to the online discussion by adding useful information (CME3, M=2.99, STDEV=2.053) and they do not interact with other viewers using hashtags related to NHL games (CME4, M=2.22, STDEV=1.862). Both items recorded a moderate standard deviation denoting a certain degree of agreement among viewers. These results underline the lack of interactions with the NHL community viewing games at the same time.

Similar to the previous three variables, respondents showed a neutral attitude for the three items of the scale and high standard deviation for all three items. They may feel like they are physically communicating with others (SP1, M=3.21, STDEV=2.134) and they may feel like they are watching the game with friends (SP2, M=3.29, STDEV=2.206). Nevertheless, respondents recorded a considerably higher mean when asked, if they feel like many people are watching at the same time (SP3, M=4.43, STDEV=2.306). These findings indicate that viewers feel the virtual presence of others, but may not feel strongly attached to them.

Commitment (COM)

In contrast to the variables discussed above, the variable on viewers’ commitment to the NHL or a viewer’s favourite team portrays positive attitudes. Respondents are emotionally attached to the NHL and/or the team they support (COM1, M=5.57, STDEV=1.824). They are also emotionally connected to the NHL and/or their team (COM2, M=5.44, STDEV=1.867), and they are emotionally committed to the NHL and/or their team (COM3, M=5.42, STDEV=1.936). Low or rather low standard deviation across the three items means that participants mostly agree on their stance. It can therefore be deduced that the responding NHL viewers are committed to the NHL or their favourite team even without considerable social TV engagement.

Viewing intention and loyalty (LOY)

Continuing from the commitment of NHL viewers to the NHL and their favourite team, the collected data shows that respondents highly agree with the statement ‘I will continue watching the NHL’ (LOY1, M=6.09, STDEV=1.579). Furthermore, respondents will recommend the NHL to others (LOY3, M=5.09, STDEV=2.073), although the standard deviation relativizes the recorded attitude. When asked if they will watch the NHL for more than only their team’s games, some respondents are inclined to do so (LOY2, M=4.69, STDEV=2.098), and they are also willing to expand viewing NHL programs other than games, e.g. behind-the-scenes and other documentaries (LOY4, M=4.40, STDEV=2.157). However, the high and moderately high standard deviation on both items shows some disagreement among respondents.

Figure 3: Overview of question-items including mean, standard deviation, factor loading, Cronbach’s Alpha (α), composite reliability (CR) and average variance extracted (AVE)

In summary, the three variables that result from social media engagement behaviour, i.e. functional, emotional, and communal engagement, show neutral to negative attitudes with high standard deviation. In regard to functional engagement, respondents do not seem keen on using social media functionalities to engage with other viewers. Similarly, respondents do only moderately express emotions with others on social media while watching NHL games. Furthermore, they do not necessarily contribute to online conversations or interact with other viewers. Respondents feel that other NHL viewers are virtually present, but social attachment is rather low. Yet, the commitment of NHL viewers to the NHL and their favourite team is high, as well as the intention to keep watching the competition.

This section examines how social TV influences viewer loyalty of NHL viewers based upon statistical results computed through structural equation modelling (SME); see Section 5.1. Recommendations on how to foster and reinforce viewing intentions for the NHL using social second screen tactics for social TV are offered in Section 5.2. Figure 4 provides a visualisation of the results from the SEM by including the standardised regression coefficients (β), R-squared measures (R²), and significance (p) within the applied conceptual framework.

Section 5.1: Interpreting the SEM

Figure 4 reveals that functional engagement has a moderately low and significant effect on social presence (H1, β=0.206, p<0.05); no significant direct effect is found between functional engagement and commitment to the NHL (H4), nor between functional engagement and viewer loyalty (H7). Similarly, emotional engagement shows a moderately low and highly significant effect on social presence (H2, β=0.331, p<0.001), but no significant direct effect on commitment to the NHL (H5), nor on viewer loyalty (H8). Communal engagement, however, records a moderately strong and significant effect on social presence (H3, β=0.563, p<0.01) and a moderate and significant effect on viewer loyalty (H9, β=0.383, p<0.001). No significant effect is found between communal engagement and commitment to the NHL (H6). Social presence directly affects commitment with high significance (H10, β=0.438, p<0.001), but has no significant effect on viewer loyalty (H11). Lastly, commitment depicts a strong and highly significant influence on viewer loyalty (H12, β=0.663, p<0.001).

This study recorded moderate-to-high squared multiple correlation scores (R²), which indicates an acceptable predictability regarding viewer loyalty of people watching the NHL indirectly being influenced by social media engagement behaviour via second screen activities (see Figure 4).

These findings parallel the findings from Lim et al. (2015), except for H2 and H5; their research examined how viewers of the 2014 Sochi Olympic Games in South Korea would engage in social TV and found that, contrary to this study, emotional engagement has no significant effect on social presence, however, it has a significant effect on commitment. In this study, the above-discussed significant relationships seem plausible, as all three social media engagement behaviours are mediated through social presence and commitment to influence viewing intentions. This implies that the sensation of being with others on social media while watching the NHL is built upon functional features offered by social media platforms, the opportunity to share feelings with others, and immediate access to others. The sensation then deepens the commitment that viewers feel towards the NHL experience, which, ultimately, creates and strengthens loyalty towards the entertainment product. Nevertheless, building a sense of belonging to the NHL community for viewers and fans through communal and immersive rituals that require active participation on social media can directly and effectively foster loyalty.

Figure 4: Conceptual framework. Standardised regression coefficients (β), R-squared (R²), significance (p)

Figure 5 presents the final model of this research based on the significant relationships within the conceptual framework. As discussed above, the model depicts how social media engagement behaviours positively influence social presence, which then leads to viewer commitment and, finally, to viewer loyalty. This section proposes tactics for the creation of social media content and communications to bolster these effects via social TV, specifically through the three social media engagement behaviours ‘functional, emotional, and communal’.

Figure 5: Final model. Standardised regression coefficients (β), R-squared (R²), significance (p)