Musical Engagement and Social Media
Engagement
in the Music Industry:
The Mediating Role of Brand Loyalty and
the Moderating Role of
Active Social Media Usage
by
KWAME AGYEMANG
University of Groningen
Faculty of Economics and Business
2
ABSTRACT
The digitalization and globalization of the music industry have made social media critical success factors in the industry nowadays and social media engagement (SME) plays an important role in the social media domain. Next to the engagement with artists on social media, the engagement with the music of artists has become a relevant topic, since Musical Engagement (ME) is a recently theorized phenomenon, affecting music consumption behavior of customers and the effect of ME on customer engagement behavior on social media is investigated. ME consists of three sub-dimensions: Social-Identity ME, Transportive ME and Affect-Inducing ME. The relation of ME and SME seems to be affected by brand loyalty and active social media usage (ASMU), as research suggest that these concepts are interrelated. This study explores these relationships with the following research
questions: What effect does musical engagement have on brand loyalty and social media engagement? And what effect does active social media usage have on these relationships? A quantitative,
4
1. INTRODUCTION
The music industry is a promising, dynamic industry that has been growing globally for the fourth consecutive year in 2018 (IFPI, 2019). The global music market has grown with 9.7 percent in 2018, which is one of the highest growth rates since 1997 (IFPI, 2019). Almost 50 percent of the global revenues were generated through online music streaming. In the U.S., online music streaming accounted for 80 percent of the music industry revenues in the first half of 2019 (Friedlander, 2019). According to Friedlander (2019), the total digital value of the music industry in the U.S has been growing with over 20 percent in the first half of 2019 since last year. These developments show that online music streaming has become the standard way of music consumption over the last few years (IFPI, 2019). The music industry is constantly evolving with new trends and artists from different cultures and countries that are distributing their music all over the world, reaching a global audience in a digital way (IFPI, 2019).
Digitalization is not only happening in the music industry, but the entire world is growing to become more digital every day. Nowadays, more than half of the world’s population is an internet user (Kemp, 2019a) and social media networks have become important factors in business practice and consumer’s daily lives. Zahoor and Qureshi (2017) describe social media as: “… the usage of web-based and mobile technologies to create, share and consume information and knowledge without any geographical, social, political or demographical boundaries through public interaction in a
participatory and collaborative way” (p. 47). To illustrate their importance, the social networks amass large numbers of active users spread out over the entire world. Currently, there are over 3.484 billion active social media users in the whole world (Kemp, 2019b). This accounts for about 45 percent of the world’s population (Kemp, 2019b). Therefore, social media plays a role in many customer’s lives.
Besides the growing importance of social media for consumers, social media have become relevant tools in business practice in general. Presently, 25 percent of the Facebook pages use paid media to reach their target audiences on social media (Kemp, 2019a). According to research by Hootsuite (2019) among over 3.200 business customers, 78 percent of the respondents stated that they have invested in social advertising or plan to do so in the next year, which indicates the importance of social media in general business practice.
Since social media have become such relevant phenomena in consumer’s daily lives and business practice, it is interesting to explore how social media engagement (SME) is related to musical
engagement and brand loyalty. There are several academic studies dedicating their attention to SME. Jahn and Kunz (2012) found that SME plays an essential role for successful social media management. This is supported by Van Asperen, De Rooij and Dijkmans (2018), who found that SME is
strategically important for companies, since it enhances their performance. SME plays an important role in the social media domain and is therefore a promising concept to study. To contribute to the current academic knowledge on social media, this research aims to study SME in the music industry.
5 The present study investigates SME in the music industry. Social media is relevant in the music business, apart from the globalization and digitalization of the industry, for two main reasons. First, many people use social media for music related activities. For instance, a study by MusicWatch found that 9 out of 10 social media users engage in music related behavior on their social media apps and that music artists are the most popular category of celebrities or public figures followed on social media (Crupnick, 2018). More research found that access to content is one of the key motivators for customers to use social media in the music industry (Salo, Lankinen, & Mäntymäki, 2013). This implies that artists can reach people that are interested in their content through social media. Second, social media offer interesting online opportunities and the constantly evolving networks make it more attractive to share music through them. A study by Erdoğmuş and Cicek, (2012) supports this notion. They found that social media users prefer to share positive, entertaining content, and that music is one of the most preferred types of content to share on social media. When users share the music of artists on social media, it helps artists reaching a larger audience and gain more awareness and exposure for themselves and their music. Moreover, the integration of social media and music streaming services also provide several interesting online opportunities for users and artist. Users can now share links to songs on Spotify in their stories on both Instagram (“We’ve Made it Easier to Share Spotify to Instagram Stories”, 2018) and Facebook (“Share Your Songs on Facebook Stories — Now with Music”, 2019). Instagram also allows users to share stickers with high quality music in their stories via Instagram Music (“Introducing Music in Stories”, 2018). These features make sharing music on social media more accessible and attractive for both users and artists. To conclude, distributing music and content through social media networks is important for artists and users. As previously mentioned, the digitalization and globalization of the music industry has made social media important factors in the industry since they offer accessible, digital ways for artists and fans to connect, interact and reach each other across the globe (Crupnick, 2018; Salo, Lankinen, & Mäntymäki, 2013; Erdoğmuş & Cicek, 2012). Therefore, this research studies SME in the music industry.
In the music industry, not only SME seems to be important, but specific engagement with the product (i.e. music) as well, referring to musical engagement (ME). ME is defined as: “A second-order construct comprising four types of consumer experiences with music (i.e. identity-, social-,
transportive-, and affect-inducing experience) that, collectively, comprise ME” (Hollebeek et al., 2016, p. 418). Hollebeek and colleagues (2016) are one of the first to investigate how ME affects music consumption behavior. Indeed, they found that the higher the ME of individuals, the higher their music consumption will be. As music consumption behavior seems to be related to ME, the present study argues that ME is important in relation to more general customer behavior as well, namely SME.
The findings of the effect of ME on music consumption behavior by Hollebeek et al. (2016) seems to be especially relevant in relation to SME because ME affects one’s brand loyalty. In the context of the present study, brand loyalty represents the behavioral and attitudinal attachment of customers towards an artist (brand) (Fang, Jianyao, Mizerski, & Huangting, 2012). It consists of two dimensions, namely attitudinal and behavioral brand loyalty (Oliver, 1999). The attitudinal dimension entails the attitudinal preference and commitment towards the brand (artist) (Jacoby & Chesnut, as cited in Bennett & Rundle-Thiele, 2002), while the behavioral dimension is mainly focused on purchasing behavior (Zhao, Chen, & Wang, 2016). Studies argue that brand loyalty is a complex,
multidimensional concept that cannot be completely measured by merely studying customer behavior (Punniyamoorthy & Raj, 2007). Therefore, to generate meaningful insights, both attitudinal- and behavioral loyalty are investigated in this study.
The effect between ME and SME seems to be partially affected by brand loyalty considering that the current study hypothesizes that ME affects brand loyalty (Hollebeek et al., 2016) and brand loyalty is related to SME (Chiang et al., 2017). ME affects brand loyalty, because specific
6 SME can have a positive effect on brand loyalty and is important for gaining and maintaining loyalty amongst customers (Van Asperen, de Rooij, & Dijkmans, 2018; Chiang et al., 2017; Leckie et al., 2016). The reverse direction of the relationship of SME and brand loyalty is studied, since it is argued that SME is a form of affirmation of brand loyalty (Keller, 2013). Thus, the present study argues that brand loyalty partially mediates the effect of ME on SME.
Apart from the relations of SME with ME and brand loyalty, SME seems to be closely related to active social media usage (ASMU) of individuals (Van Asperen et al., 2018), although they are not the same concept. ASMU is a general measure of the social media usage of customers, defined as the range of activities individuals participate in on social media (Smith & Gallicano, 2015), while SME stands for the specific engagement behavior of customers towards artists on social media. ASMU is described in line with Van Asperen et al. (2018) as contributing to social media content in general, and not only towards content of artists. Hence, ASMU is different from SME because ASMU is a concept that measures the activities of individuals on social media in general, while SME focuses on the specific engagement behavior on social media towards artists. ASMU is expected to play a moderating role in the relationships between brand loyalty-SME and ME-SME, because research found that individuals who are in general using social media actively, are more likely to engage with brands on social media and thus, it is assumed that ASMU strengthens the translations of brand loyalty and ME into SME (Van Asperen et al., 2018).
In conclusion, this research investigates the effect of ME on SME, how brand loyalty could play a mediating role in this relationship and it explores the moderating role of ASMU in these relationships. This unfolds into the following research questions: What effect does musical engagement have on brand loyalty and social media engagement? And what effect does active social media usage have on these relationships?? Since social media have become crucial factors for artists to become successful in the music industry nowadays and ME is a recently theorized phenomenon that is significantly related to customer behavior, insight in the effect of ME and brand loyalty on SME could be fruitful for artists, labels and other parties operating in the music industry (Hollebeek et al., 2016). The moderating role of ASMU will be studied to gain insights into behavior of customers on social media. Facebook and Instagram are the social media networks that this study focusses on, since these are the biggest social networks with most active users in the world together with YouTube (Kemp, 2019b).
The present study aims to make four academic contributions by: (1) exploring the effect of ME on brand loyalty and SME, (2) investigating to what extent brand loyalty affects SME, this study will be one of the first to explore the extent to which this relationship could work the other way around (other than Leckie et al., 2016 e.g.), (3) studying the mediating role of brand loyalty in the relationship between ME and SME, this study will be one of the first to investigate to what extent the theories associating brand loyalty with ME and SME complement each other, and finally (4) exploring the moderating role of ASMU in the relationships of ME and brand loyalty with SME.
7
2.
LITERATURE
Social Media Engagement as a Specific Type of Customer Engagement
The present research focuses on the effect of musical engagement (ME) on behavioral activities of individuals on social media towards artists, namely, social media engagement (SME). As previously argued, due to the digitalization and globalization of the music industry, social media networks have become important factors in the music business and therefore, they form the main context for this study regarding customer engagement behavior. Thus, the present study does not investigate the entire multidimensional concept of customer engagement but focuses on the behavioral dimension of customer engagement in the social media context: SME.
SME is a form of customer engagement and this concept has been an interesting topic in academic research. Studies argue that customer engagement is a relevant concept to study, since it impacts many different stakeholders, ranging from the brand and other customers to society in general (Van Doorn et al., 2010). Customer engagement encompasses multiple objects and has a complex nature (Unal, Schivinski, & Brzozowska-Woś, 2017). The concept knows several academic
definitions. Higgins and Scholer (2009), for example, define customer engagement as “… a state of being involved, occupied, fully absorbed, or engrossed in something - sustained” (p. 102). Other studies describe engagement as a commitment to an active relationship with a specific market offering and distinguish different dimensions of engagement: utilitarian, hedonic and social (Abdul-Ghani, Hyde & Marshall, 2011). There is no clear consensus in the literature about the exact definition of engagement. However, according to a review of online customer engagement literature by Unal et al. (2017), many academic studies adopted the view of Brodie, Hollebeek, Jurić, and Ilić (2011), by distinguishing three dimensions of customer engagement: behavioral, emotional and cognitive. Brodie et al. (2011) define customer engagement as: “… a psychological state that occurs by virtue of
interactive, cocreative customer experiences with a focal agent/object (e.g., a brand) in focal service relationships…” (p. 9). This definition was further developed by Hollebeek, Glynn and Brodie (2014) as: “A consumer's positively valenced brand-related cognitive, emotional and behavioral activity during or related to focal consumer/brand interactions” (p. 154). This indicates the complex nature of customer engagement. To conclude, customer engagement is an interesting concept with a multi-dimensional nature and in the present study, the focus lies on the behavioral dimension of customer engagement: customer engagement behavior.
In this study, customer engagement behavior will be investigated on social media. Customer engagement behavior is defined as: “the customer’s behavioral manifestation toward a brand or firm, beyond purchase, resulting from motivational drivers” (Van Doorn et al., 2010, p. 253). The focus lies on the behavioral aspect of customer engagement on social media to gain more insights into when and how fans engage with artists on social media. Besides, social media engagement behaviors are tangible measures for artists and could generate meaningful insights into the artists’ performance on social media. Furthermore, behavioral engagement rates are important performance indicators on social media in business practice, since they are used as a performance benchmark (Kemp, 2019b). Additionally, according to Van Doorn et al. (2010), customer engagement behavior is: “… an important research topic for marketing scholars who want to take a comprehensive and integrated approach to understanding customers” (p. 262). Therefore, the conceptualization of SME in the present study refers to behavioral dimensions of customer engagement only.
8 commenting, interacting, sharing, recommending and uploading content, whereas passive engagement is described as the consumption of content, viewing pictures and videos (Van Asperen et al., 2018). In the present study, these active and passive engagement behaviors represent the behaviors of
individuals towards artists and artists’ content on social media. Both forms of behavioral engagement on social media are incorporated in the present research to gain more complete insights in the effects of ME and brand loyalty on SME compared to only focusing on either passive or active behaviors, because both types of engagement are considered important social media engagement behaviors (Van Asperen et al., 2018).
The Relationship Between Musical Engagement and Social Media Engagement
The present study revolves around engagement behavior of consumers on social media: SME, but research indicates that musical engagement (ME) can play a significant role in affecting consumer behavior in the music industry as well (Hollebeek et al., 2016).
In the current research, the effect of ME on SME will be studied. ME includes several consumer experiences with music and it is defined as a second-order construct that consists of social-identity, transportive-, and affect-inducing experiences (Hollebeek et al., 2016). According to the study of Hollebeek et al. (2016), the social-identity customer-music experience focuses on the desire of the projection of an individual’s identity through interactions with objects and on the desire to connect with peers, friends and family through the music they are listening to. The transportive experience entails the desire to escape, relax and virtue development through interaction with specific music. And finally, the affect-inducing experiences reflect the particular feelings music evokes in an individual (Hollebeek, et al., 2016).
The study by Hollebeek and colleagues (2016) is one of the first, pioneering studies
investigating ME of customers. According to Hollebeek and colleagues (2016), they were the first to investigate customer engagement with music from a uses and gratifications theory perspective, a theory that assumes that customers: “proactively integrate specific media into their lives” (Hollebeek et al., 2016, p. 418). Hollebeek and her colleagues (2016) developed a scale of measurement for ME and tested this scale with regression and mediation modeling. The results of their study show that especially the social-identity and transportive experiences with music were significantly affecting music consumption, while the affect-inducing dimension was found to be less relevant. However, all three dimensions of ME are incorporated in the present study to explore to what extent the previous findings of Hollebeek and colleagues are consistent across different contexts, as the concept of ME is relatively new. Thus, the Social-Identity-, Transportive- and Affect-Inducing dimensions of ME are therefore all incorporated in the present study.
9 Therefore, this research aims to investigate the relationship between ME and SME by
generating the following hypothesis:
Hypothesis 1. Musical Engagement is positively related to Social Media Engagement. To study the relation between the several dimensions of ME on SME, the following sub-hypotheses were created:
Hypothesis 1a. Social- Identity Musical Engagement is positively related to Social Media Engagement.
Hypothesis 1b. Transportive Musical Engagement is positively related to Social Media Engagement.
Hypothesis 1c. Affect-Inducing Musical Engagement is positively related to Social Media Engagement.
Musical Engagement and Social Media Engagement:
The Mediating Role of Brand Loyalty
The concept of customer engagement proves to be an important topic for academic studies, but research on the antecedents of the phenomenon is also important, since it is unclear how antecedents of customer engagement interact with each other to affect customer engagement behavior (Van Doorn et al., 2010). In the present study, the mediating role of brand loyalty in the relationship between ME and SME is investigated.
Academic studies view brand loyalty as a two-dimensional construct, consisting of a
behavioral- and an attitudinal dimension (Gommans, Krishman, & Scheffold, 2001). The behavioral dimension often incorporates frequent or repeated purchase behavior, while the attitudinal dimension focuses on the cognitive aspects of customer loyalty towards brands. Oliver (1999) displays this view by defining loyalty as: “A deeply held commitment to rebuy or repatronize a preferred product/service consistently in the future, thereby causing repetitive same-brand or same brand-set purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior” (p.34). This definition highlights the two dimensions of brand loyalty. The behavioral part is stressed in the repetitive purchasing part, while the attitudinal dimension is displayed through the deeply held commitment. In line with the two-dimensional view of loyalty, attitudinal loyalty is defined as: “The consumer’s predisposition towards a brand as a function of psychological processes. This includes attitudinal preference and commitment towards the brand” (Jacoby & Chesnut, as cited in Bennett & Rundle-Thiele, 2002, p. 6). And, behavioral loyalty refers to “…the tendency of a consumer to purchase a particular brand repetitively over time” (Liu, Li, Mizerski, & Soh, 2012, p. 925). As both attitudinal and behavioral dimensions have been regarded as important components of brand loyalty, the present study focuses on both these dimensions. In the specific context of the current research, brand loyalty embodies the behavioral and attitudinal attachment of customers towards an artist (brand) (Liu et al., 2012), since the mediating role of brand loyalty is investigated in the relationship between ME and SME of customers with artists in the music industry. This study incorporates both behavioral- and attitudinal loyalty dimensions to facilitate a comprehensive measurement of brand loyalty and a more complete view of the construct (Punniyamoorthy & Raj, 2007).
10 Nenycz-Thiel, 2013; Zhao, Chen, & Wang, 2016), this proposes a relation between ME and the behavioral part of brand loyalty. Hollebeek and colleagues (2016) purely investigated the relation of ME with music consumption behavior of individuals. In the current study, the aim is to extend their research by investigating the relationship of ME and brand loyalty altogether, since both behavioral and attitudinal dimensions of brand loyalty are incorporated in the present research.
The extension of the research of Hollebeek et al. (2016) is important, since brand loyalty is a multidimensional construct consisting of behavioral and attitudinal dimensions, and the exploration of the effect of ME on the attitudinal dimension of brand loyalty is important to analyze the full effect of ME on brand loyalty (Punniyamoorthy & Raj, 2007). It is important to investigate the attitudinal dimension of brand loyalty to measure true loyalty, as brand loyalty does not only exist of certain behavior customers. It also includes the attitudes about brands, which have different implications for customer-brand relationships and can affect other factors such as brand image and customer-based brand equity for example Keller (2003). Besides, the theory of planned behavior by Ajzen (1991) argues that attitudes are important elements to explain behavior. This view is supported by several empirical academic studies, showing a significant impact of attitudes on future behavior (Kraus, 1995). This shows that attitudes are related to the behavioral dimensions of brand loyalty. Therefore, the attitudinal dimension of brand loyalty should be incorporated in the measurement of the full construct. This study investigates the effect of ME on both brand loyalty dimensions by producing the following hypothesis:
Hypothesis 2. Musical Engagement is positively related to Brand Loyalty.
This hypothesis is also divided into three sub-hypotheses, as the construct of ME consists of three sub-dimensions:
Hypothesis 2a. Social-Identity Musical Engagement is positively related to Brand Loyalty. Hypothesis 2b. Transportive Musical Engagement is positively related to Brand Loyalty. Hypothesis 2c. Affect-Inducing Musical Engagement is positively related to Brand Loyalty.
The establishment of a mediating role for brand loyalty starts with the hypothesis that ME positively influences brand loyalty. To proceed, this study argues that brand loyalty is positively related to SME, which unfolds into a mediating effect of brand loyalty between ME and SME.
The present research incorporates studies from the brand loyalty and customer loyalty
academic fields, since brand loyalty seems to be closely related to customer loyalty. Customer loyalty refers to the loyalty of customers related to brands, firms or products and services (Van Asperen et al., 2018). Studies view customer loyalty as a concept that encompasses loyal behavior towards brands and companies (Van Asperen, et al., 2018). Therefore, the concept of customer loyalty is intertwined with brand loyalty, since brand loyalty displays the loyalty of customers towards brands. Thus, this study builds on both customer loyalty and brand loyalty literature.
As previously mentioned, various studies prove the explanatory power of SME regarding brand loyalty and customer loyalty (Yoshida et al., 2018; Van Asperen et al., 2018; Islam, Rahman, & Hollebeek, 2018; Chiang et al., 2017; Leckie et al., 2016). These academics assume that engagement results in more loyalty. To illustrate, the study by Van Asperen et al. (2018) studied the effect of social media engagement of customers with travel agencies on customer loyalty. They found that SME has a positive effect on customer loyalty. However, as a limitation of their study, Van Asperen and
11 affection or higher purchase intention. This suggests that it could be interesting to explore the direction of the relationship between SME and brand loyalty, since it could work in a different direction.
The customer-based brand equity pyramid by Keller (2003) argues how customer loyalty could result in more SME rather than reversed. According to this theory, engaging, interacting behavior with a brand on social media is a consequence of a loyal, close customer-brand relationship. Keller (2003) argues that the most valuable brand building block is brand resonance. Brand resonance represents a relationship between customers and the brand that is in complete harmony. Keller (2003) states: “With true brand resonance, customers have a high degree of loyalty marked by a close relationship with the brand, such that customers actively seek means to interact with the brand and share their experiences with others” (p. 14). More recently, Keller (2013) argues that the strongest affirmation of brand loyalty occurs when customers are engaged, and he mentions that the customers are willing to invest energy, time, money or other resources in the brand, besides the consumption or purchase expenses of the brand. Customer engagement behavior on social media is an example of an investment in the brand beyond consumption or purchase expenses. Thus, it is argued that SME is an extra investment in the brand, as a consequence of brand loyalty.
The present study will examine the extent to which this theory can be validated by exploring the effect of brand loyalty on SME. This contributes to the current literature, as most studies assume that engagement affects loyalty (e.g., Yoshida et al., 2018; Van Asperen et al., 2018; Islam, Rahman, & Hollebeek, 2018; Chiang et al., 2017; Leckie et al., 2016), and the reverse direction remains to be explored as far as known. The present study explores this relationship with the following hypothesis:
Hypothesis 3. Brand Loyalty positively affects Social Media engagement.
Next to the theoretical contribution, insights in the direction of the relationship between brand loyalty and SME can be valuable for business practice, since the relationship could work in another direction than current studies are suggesting (Van Asperen, de Rooij, & Dijkmans, 2018; Islam, Rahman, & Hollebeek, 2018; Chiang et al., 2017; Leckie et al., 2016). These insights could help the business practice in the music industry, since it provides more clarity in the direction of the
relationship, which could have implications for the strategies of parties operating in the music industry. For example, artists or companies could deploy different marketing strategies when they want to focus on establishing brand loyalty first instead of SME. Therefore, it is important to investigate the relation between brand loyalty and SME in a reverse direction.
Derived from findings of the previously mentioned studies, the present study argues that the relationship between ME and SME is partially mediated by brand loyalty, since it seems that
customers who experience ME with the music of an artist, will be more loyal to that artist (hypothesis 2), and this brand loyalty affects SME, as a confirmation of brand loyalty (hypothesis 3). Investigating this mediating effect based on hypothesis 2 and 3 contributes to the literature by exploring the extent to which different theories complement each other in the context of the present study. Thus, the current research explores the mediating role of brand loyalty by investigating the following hypotheses:
Hypothesis 4. Brand Loyalty partially mediates the effect of Musical Engagement on Social Media Engagement.
Hypothesis 4a. Brand Loyalty partially mediates the effect of Social-Identity Musical Engagement on Social Media Engagement.
12 Hypothesis 4c. Brand Loyalty partially mediates the effect of Affect-Inducing Musical Engagement on Social Media Engagement.
Musical Engagement and Social Media Engagement:
The Moderating Role of Active Social Media Usage
The engagement with the music of artists and engagement behavior on social media of consumers towards artists are relevant research areas in the digitalized music industry, but general behavior of customers on social media seems to be of importance regarding these types of
engagements as well. The present study argues that active social media usage (ASMU) plays a moderating role in the relationships between ME-SME, and brand loyalty-SME.
In the current research, ASMU will be a measure of the general activities on social media of consumers. Social media usage refers to the range of activities individuals participate in on social media (Smith & Gallicano, 2015). The concept of ASMU is further defined based on the active engagement dimension in the study of Van Asperen et al. (2018). It refers to active, contributing behavior towards content on social media in general. ASMU is different from SME, because SME focuses on the specific engagement behavior of consumers towards artists on social media, while ASMU is as a measure of general active behavior of consumers on social media.
Findings of a study by Van Asperen and colleagues (2018) imply that customers who are in general more active and more willing to share on social media, contribute more actively to brand-related content on social media. This suggests that active social media usage in general could affect the relation between ME and SME, since it is assumed that if individuals actively use social media in general, they are more likely to actively engage with artists on social media. Thus, ASMU will strengthen the relationship of ME and SME. Therefore, the present study hypothesizes the following:
Hypothesis 5. The higher the active social media usage, the stronger the relation between musical engagement and social media engagement.
Hypothesis 5a. The higher the Active Social Media Usage, the stronger the relation between Social-Identity Musical Engagement and Social Media Engagement. Hypothesis 5b. The higher the Active Social Media Usage, the stronger the relation between Transportive Musical Engagement and Social Media Engagement.
Hypothesis 5c. The higher the Active Social Media Usage, the stronger the relation between Affect-Inducing Musical Engagement and Social Media Engagement.
Furthermore, research found that customers who actively contribute more on social media towards brands, do not necessarily have a higher degree of loyalty towards those brands (Van Asperen et al., 2018). This suggests that the amount of ASMU in general can help explain and have an
13 behavior or decision. This action-based model could be applied to engagement behavior of customers in the music industry as well. To illustrate, when customers frequently actively engage with content on social media in general (attitudes are often adjusted based on behavior), and they perceive (high) brand loyalty towards an artist, they could find themselves in a state of cognitive dissonance when they do not engage with the content of these artists on social media, because they do not engage with the artist they are loyal and committed to, while they frequently actively engage with other content on social media. According to the action-based model of dissonance, to reduce this state of dissonance between their general engagement behavior and their behavior towards artists, they could engage more actively with artists on social media. Therefore, ASMU could affect the relationship between brand loyalty and SME, since ASMU in general affects the cognitions of customers, and according to the action-based model of dissonance, it indirectly affects SME towards artists. Thus, the following is hypothesized:
Hypothesis 6. The higher the active social media usage, the stronger the relation between brand loyalty and social media engagement.
The ASMU concept is incorporated in the study with the aim to prevent the issues found by Van Asperen and colleagues (2018), since the results of their study imply that the general SME behavior of individuals could affect their brand-related SME behavior as well. By incorporating ASMU, insights into the general SME behavior of the customers are generated and the effect of ASMU on the relations between ME-SME and brand loyalty-SME is investigated. This could lead to clearer and more accurate observations of these relationships and could thus generate more useful findings. Furthermore, it contributes to the academic knowledge of social media usage by
investigating the extent to which ASMU can affect the relationships of brand loyalty and ME with SME (Van Asperen, et al. 2018). Therefore, the present study incorporates ASMU as a measure of customers’ active engagement behavior in general.
Conceptual Model and Hypotheses
The evolved hypotheses will be studied in the context of consumers and artists in the music industry with the following research questions: What effect does musical engagement have on brand loyalty and social media engagement? And what effect does active social media usage have on these relationships? The hypotheses as previously stated are visually represented in the conceptual model below (Figure 1).
The proposed study aims to make four important contributions. First, the relationship between ME and SME in the music industry will be explored. This is important because it could offer relevant insights for the music industry, and it contributes to the exploration of ME in the academic discipline by extending current research towards different types of customer behavior and investigating its explanatory power regarding other brand loyalty dimensions than only behavioral loyalty (Hollebeek et al., 2016). Second, rather than focusing on how SME affects brand loyalty, this research investigates to what extent the reverse direction holds. Since SME could be a confirmation of brand loyalty
14
H5 H6
Figure 1: Conceptual Model Musical Engagement • Social-Identity • Transportive • Affect-Inducing Brand Loyalty Active Social Media
Usage
15
3.
METHODOLOGY
Research Design: Data collection and sampling strategy
The quantitative correlational research design was employed through a digital web-based survey. Through a digital survey, data could easily be gathered and directly imported to statistical programs used to analyze the dataset. This was an appropriate way to execute the research, since this type of survey was convenient to distribute to the target population and it enabled efficient data gathering and analyses in line with the explorative aim of the present research. Qualtrics software was used to design the survey. Qualtrics offer a suitable way to create surveys through their online research platform and allows the dataset to be exported directly to SPSS (“Online Survey Software – Trusted by +5.5M Survey Creators | Qualtrics UK”, 2019).
To gather meaningful data through the digital web-based survey, the target population of the present study consisted of individuals with social media profiles on Facebook and/-or Instagram, who listen to music on Spotify, Apple Music or YouTube, because (1) these social media platforms are the most popular in the world (Kemp, 2019b) and (2) these streaming services amass the most music listeners in the world (Richter, 2017). The study addressed this target group to facilitate meaningful data gathering, since the research revolved around social media engagement (SME) in the music industry.
The target group was reached by spreading the survey through the Internet on Facebook and Instagram. Spreading the survey through social media offers a convenient way to reach many people, at low costs (Blumberg, Cooper, & Schindler, 2008). Additionally, as the study investigated social media behavior, the highly relevant target group could be reached by spreading the survey through the social media platforms incorporated in the study (Facebook and Instagram). Thus, convenience sampling was the main sampling strategy used in this study, since this strategy enabled the targeting of the broad target population and it offered the most convenient way to collect data in the situation of restricted resources (Blumberg, Cooper, & Schindler, 2008). Furthermore, friends and acquaintances were asked to share the survey on Facebook and amongst friends as well, creating a form of snowball sampling, which led to increased dissemination of the survey (Blumberg, Cooper, & Schindler, 2008).
To improve the power of the research, which according to Van Voorhis and Morgan (2007) represents: “the probability of correctly rejecting a false null hypothesis” (p. 44), a certain sample size was needed. Following the rules of thumb of Van Voorhis and Morgan (2007), a good sample size should be around 300 respondents, since a factor analysis will be conducted in this study to measure the constructs. This size was needed because it affects the power of the research and therefore, the goal of 300 respondents was set to achieve in this study (Van Voorhis & Morgan, 2007).
Measures of Variables, Procedure and Questionnaire
16 students consisted of a Communication Sciences Master student, a Marketing Master student and a Business Administration Bachelor student.
To measure the construct of musical engagement (ME), the scale of Hollebeek et al. (2016) was used since they conceptualized and tested the ME concept. Brand loyalty was measured by measuring both behavioral and attitudinal brand loyalty based on scales from Watson, Beck,
Henderson and Palmatier (2015) and Brady, Voorhees, and Brusco (2012). The scale of Van Asperen et al. (2018) was used to measure SME and active social media usage (ASMU). The adapted scales regarding the several constructs in this study can be found in the survey, enclosed in Appendix A.
Age and Gender were incorporated as control variables in the current research, since the study by Hollebeek and colleagues (2016) found that these control variables significantly affected the ME of customers. These variables are important in the current research, as the target population was a very broad group of individuals of all ages, and were thus likely to influence the ME of the respondents (Hollebeek et al., 2016).
The digital survey designed through Qualtrics consisted of several matrix tables, measuring the constructs in this study. Respondents were guided through the survey through the utilization of skip logic and display logic systems. First, respondents were required to own social media profiles on Facebook or Instagram and to use the Spotify, Apple Music or YouTube platforms to listen to music. Then, the respondents were asked to enter the name of an artist, musician or band they have listened to. This name would then be displayed amongst the questions of the several constructs in this study, to enable the measurement of their musical engagement (ME), brand loyalty and SME towards this artist. All constructs in the study were measured using seven-point Likert scales, ranging from strongly disagree (1) to strongly agree (7). The constructs and their subdimensions were randomized throughout the survey to prevent artificial correlations between the subdimensions within the theoretical constructs. The survey with all the questions can be found in Appendix A.
Data Analyses
The data gathered with the survey was analyzed through IBM SPSS 25. First, descriptive statistics were produced to generate a general overview of the dataset. Second, the data was cleaned by exploring outliers and participants who did not complete the survey or were unable to meet the
participation requirements (i.e. owning a profile on Facebook or Instagram, or streaming music on Spotify, Apple Music or YouTube). Furthermore, the validity and reliability of the constructs were tested by conducting a factor analysis and conducting a reliability analysis, where Cronbach’s α was consulted for each factor.
Before conducting the factor analyses, the Kaiser-Meyer-Olkin (KMO) statistic and Bartlett’s Test of Sphericity were consulted to see whether the factor analyses could be properly executed. According to Malhotra (as cited in Ramnarain & Govender, 2013), the KMO statistic measures the sampling adequacy, ranges from 0 to 1 and indicates that factor analysis is an appropriate technique for KMO values above .5. The Bartlett’s Test of Sphericity tests the null hypothesis that the
correlation matrix has an identity matrix, which means that the items are unrelated and unsuitable for a factor analysis (Bartlett, 1950). Thus, a significant p-value for the Bartlett’s Test of Sphericity
indicates that the items are related and thus suitable for a factor analysis. Therefore, the KMO and Bartlett’s Test of Sphericity were performed to ensure the appropriateness of a factor analysis in the present study.
17 criteria is a superior method to determine a factor solution (Fabrigar, Wegener, MacCallum, &
Strahan, 1999; Reio & Shuck, 2015).
Three different methods to select factor solutions were used in the current research. Firstly, factor selection based on the criterium that over 50 percent variance should be explained by a factor to prevent problems with the structure of the factor was used (Farrell & Rudd, 2009). Secondly, factor selection based on the examination of the scree plot was used, where the eigenvalues are visualized in a graph, plotted against the number of factors, and the number of factors on the left of the ‘elbow’ in the scree plot are selected (Reio & Shuck, 2015). Finally, factor selection based on the eigenvalues of the factor solutions was used. Kaiser (as cited in Reio & Shuck, 2015) argued that the eigenvalues of the factor solutions should be larger than 1 in order to be interpretable as a factor.
Besides the different methods for selecting factor solutions, factor loadings of items with the specific factors were examined to analyze the structure and strength of the factors. A cutoff criterium of a factor loading of .70 was used to analyze factor loadings of the items with the chosen factors regarding the factor structure and reliability of the factor (Farrell & Rudd, 2009).
The thresholds for the factor- and reliability analysis were used as guidelines in the present study to enable reasoned decision making in the process of the statistical analyses. However, there are multiple possibilities regarding these thresholds as different scientists argue for and use different criteria and methods in their research (e.g., Fabrigar et al., 1999; Reio & Shuck, 2015; Taber 2018). Therefore, the methods and criteria were not always decisive in the present study and certain reasonable decisions were made.
Finally, the hypotheses of the present study were tested with moderation- and mediation analysis using Hayes Process Macro model 1 and model 4 (Hayes, 2017). To ensure that linear regression analysis was appropriate, the OLS assumptions were consulted (Winship & Radbill, 1994). Most of the assumptions were met, except for the assumption of normality and homoscedasticity. However, Nimon (2012, p. 4) assumes that linear regression models are robust enough with minor violations of the homoscedasticity principle and according to Anglim (2011) the sample size of the current study is large enough to correct for the violation of the normality principle (as presented in Appendix D). Hence, linear regression analyses were used to test the hypotheses of the current study. Please view appendix D for an overview of these assumptions and the conclusions based on the dataset of the current study.
Data preparation
Factor Analyses & Reliability Analyses
18 TABLE 1
Musical Engagement: Factor & Reliability Analyses
First, musical engagement (ME) was analyzed. An interesting finding was discovered in the correlation matrix of the factor analysis of ME (Appendix B1). Several correlations below .30 of the items in the Transportive dimension with the other dimensions that comprise ME (Social-Identity- and Affect-Inducing ME) were found. This suggest that the Transportive dimension of ME measures a different type of dimension of the latent variable (ME) than the other two dimensions. Furthermore, the correlation table (Appendix B1) shows that the correlation is insignificant of two Affect-Inducing ME items: ‘I sometimes feel like crying after listening to certain songs of [artist]’ and ‘Music of [artist] sometimes touches me deep down’ with a Transportive ME item: ‘I like to wear t-shirts or other clothing with the logo or name of [artist]’.
The low correlation could be explained by examining the nature of the dimensions of ME. The Transportive dimension of ME is a dimension that focuses on actual customer behavior, which is quite different from the Social-Identity and Affect-Inducing dimensions, as they focus on the customer experiences with the actual music of artists, hence the low correlation between these different types of dimensions of ME.
The factor analysis of ME passed the KMO criterium of .50 (Malhotra, as cited in Ramnarain & Govender, 2013) (KMO = .83) and passed the Bartlett’s test of Sphericity (p < .001) and therefore, the factor analysis could be conducted (Table 1). The scree plot and the table of total variance explained clearly indicated a three-factor solution, since the three-factor solution explains 65.20 percent of the total variance of the latent construct ME with an eigenvalue of 1.31 (Farrell & Rudd, 2009; Kaiser, as cited in Reio & Shuck, 2015).
Construct and items Factor
Loadings Cronbach’s α KMO Bartlett’s Test of Sphericity
Social-Identity Musical Engagement (SI_ME) .81 .83 p < .001
ME_1_1 - I often unwind and relax by listening to music of [artist].
.71
ME_1_2 - Listening to music of [artist] is an escape. .82
ME_1_3 - I feel less stress after listening to music of [artist]. .86
ME_1_4 - I lose myself in the pleasure of listening to my favorite music of [artist].
.78
Transportive Musical Engagement (T_ME) .76 .83 p < .001
ME_2_1 - I “like” or follow [artist] on Facebook or Instagram. .71
ME_2_2 - I like to discuss [artist] and music of [artist] on social media sites (Facebook or Instagram).
.85
ME_2_3 - I like to wear t-shirts or other clothing with the logo or name of [artist].
.80
ME_2_4 - Part of my role among friends is to keep them informed about new music or about live shows of [artist].
.72
Affect-Inducing Musical Engagement (AI_ME) .78 .83 p < .001
ME_3_1 - Some songs of [artist] send shivers up my spine or give me goose bumps.
.70
ME_3_2 - I sometimes feel like crying after listening to certain songs of [artist].
.90
19 Furthermore, the Oblimin rotation technique was used to gain better insights into the factor solutions, since it is assumed that these items all measure the same latent variable: ME, and are
dependent of each other (Hollebeek et al., 2016). The pattern matrix indicates that the items measuring the Social-Identity ME dimension fit into one factor as the factor loadings all live up to the cutoff criterium of .70 (Farrell & Rudd, 2009). The items measuring the Transportive ME dimension all fit into the second factor. And finally, the items related to the Affect-Inducing ME dimension fit into the third and final factor. The three ME factors and the factor loadings are visualized in Table 1. The component correlation matrix (Table 2) shows low correlation (r < .40) between the three new factors, which prevents issues of multicollinearity.
TABLE 2
Musical Engagement: Component Correlation Matrix
Component Correlation Matrix
Component 1 2 3
1 1.00 .39 .37
2 .39 1.00 .22
3 .37 .22 1.00
"Extraction Method: Principal Component Analysis. Rotation Method: Oblimin with Kaiser Normalization."
To test the reliability of the new factors, Cronbach’s α was computed for Social-Identity Musical Engagement (Cronbach’s α = .81), Transportive ME (Cronbach’s α = .76), and Affect-Inducing Musical Engagement (Cronbach’s α = .78) (see Table 1). These Cronbach’s α values exceed the value of .70 and thus indicate that these factors are internally consistent enough to measure the latent constructs of the factors (Taber, 2018). Therefore, the new factors were created in SPSS by adding up the scores of all the items in each of the factors and dividing the total by the number of items included in one particular factor, which created mean scores for all the respondents for each individual factor.
Secondly, social media engagement (SME) was analyzed. The correlation table (Appendix B2) shows a high correlation of .93 between: ‘I watch videos on the Facebook or Instagram page of [artist]’ and ‘I view pictures on the Facebook or Instagram page of [artist]’. The high correlation indicates that these items are measuring almost the same, which could cause multicollinearity issues. However, as these items measure the same construct, the multicollinearity of these items does not present any issues for the further analyses in the study, since discriminant validity is not required amongst the items within a theoretical construct as it is theoretically assumed that these items are related to each other. Therefore, both items were incorporated in the factor analysis of SME.
20 TABLE 3
Social Media Engagement: Factor & Reliability analysis
Next, brand loyalty was treated in the factor analysis (Table 4). The items: ‘I only listen to music from [artist]’ and: ‘The last time I listened to this type of music, I listened to [artist]’ showed several low correlations with the other variables measuring brand loyalty (Appendix B3). Furthermore, ‘I listen most to [artist]’ showed some low correlations under .30 (Appendix B3), and the nature of this question seemed to be related to the nature of the other items with low correlations, since these questions all undermine the opportunity for respondents to listen to other artists.
The factor analysis passed the KMO criterium of .50 (Malhotra, as cited in Ramnarain & Govender, 2013) (KMO = .85) and Bartlett’s test of Sphericity (p < .001). The one-factor solution explains 47.47 percent of the total variance and lies under the cutoff value of 50 percent (Farrell & Rudd, 2009) and the eigenvalues of the second factor are higher than 1 (Kaiser, as cited in Reio & Shuck, 2015). However, the scree-plot clearly suggested a one-factor solution and the two-factor solution did not clearly distinguish two different factors. Some scholars argue that eigenvalues tend to extract too many factors and that the scree plot method can be more effective, despite the subjective judgment of the plot (Henson, Capraro, & Capraro, 2001). Therefore, the options for a one-factor solution were further considered.
Two items showed low factor loadings for a one-factor solution: ‘I only listen to music from [artist]’ had a factor loading of .39 and ‘The last time I listened to this type of music, I listened to [artist]’ had a factor loading of .36 (Table 4). As mentioned before, these questions seem to be of a different nature than the other questions measuring brand loyalty. However, to prevent data loss, a reliability analysis of the factor of brand loyalty was conducted to see if the items with the low factor loadings strongly affected the reliability of the construct.
The Cronbach’s α of the brand loyalty factor shows that the items are internally consistent (Cronbach’s α = .86). To examine if the items with the low factor loadings and correlations significantly affect the reliability of the brand loyalty construct, the item-total statistics table was consulted (Table 5). The table shows that the Cronbach’s α of brand loyalty could be improved by .01 if ‘The last time I listened to this type of music, I listened to [artist]’ would be left out of the analysis. However, this increase is relatively small, and the costs of data loss would outweigh the benefits of the improved Cronbach’s α. Therefore, it was decided to include all the items in the brand loyalty factor.
Construct and items Factor
Loadings Cronbach’s α KMO Bartlett’s Test of Sphericity
Social Media Engagement (SME_F) .89 .86 p < .001
SME_1_1 I engage in conversations on the Facebook or Instagram page of [artist].
.71
SME_1_2I share posts of [artist on my own Facebook or Instagram page.
.77
SME_1_3I recommend the Facebook or Instagram page of [artist] to my contacts.
.74
SME_1_4I upload [artist]- related video, audio, pictures or images.
.69
SME_2_1 I watch videos on the Facebook or Instagram page of [artist].
.87
SME_2_2I view pictures on the Facebook or Instagram page of [artist].
.87
SME_2_3I read posts of [artist] or user comments (or reviews).
21 To conclude, based on the reliability analysis and to prevent data loss, the two items with factor loadings below the cutoff value were incorporated in the brand loyalty construct, as deleting those items would not have substantially improved the reliability of the construct (see Table 4). Finally, the construct for brand loyalty was created by computing the mean of all the items measuring brand loyalty.
TABLE 4
Brand Loyalty: Factor & Reliability Analysis
TABLE 5
Brand Loyalty: Reliability Analysis, Item-Total Statistics
Item-Total Statistics Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted
BL_1_1 I prefer [artist] over other artists. .65 .84 BL_1_2 I enjoy listening to [artist]. .52 .85 BL_1_3 I consider [artist] as my first preference when I listen to
this type of music.
.66 .84 BL_1_4 I have a positive attitude toward [artist]. .56 .85 BL_1_5 I really like [artist]. .63 .84 BL_2_1 I often listen to music from [artist]. .75 .83 BL_2_2 I only listen to music from [artist]. .40 .86 BL_2_3 The last time I listened to this type of music, I listened
to [artist].
.31 .87 BL_2_4 I frequently listen to [artist]. .76 .83 BL_2_5 I listen most to [artist]. .65 .84
Finally, active social media usage (ASMU) was analyzed with a factor analysis (Table 6). The factor analysis of ASMU passed the KMO criterium (Malhotra, as cited in Ramnarain & Govender, 2013) (KMO = .70), and Bartlett’s Test of Sphericity was passed (p < .001). The results of the factor
Construct and items Factor
Loadings Cronbach’s α KMO Bartlett’s Test of Sphericity Brand Loyalty (BL_F) .86 .85 p < .001
BL_1_1 I prefer [artist] over other artists. .73 BL_1_2 I enjoy listening to [artist]. .64 BL_1_3 I consider [artist] as my first preference when I listen to this
type of music.
.73 BL_1_4 I have a positive attitude toward [artist]. .70 BL_1_5 I really like [artist]. .76 BL_2_1 I often listen to music from [artist]. .83 BL_2_2 I only listen to music from [artist]. .39 BL_2_3 The last time I listened to this type of music, I listened to
[artist].
22 analysis of ASMU show that the total variance explained by a one factor solution is 49.13 percent, which is below the cutoff value, but still lays in proximity of the cutoff value of 50 percent (Farrell & Rudd, 2009). Furthermore, the scree plot clearly shows a one-factor solution and it has an eigenvalue of 1.97 (Kaiser, as cited in Reio & Shuck, 2015). Therefore, the variables explaining ASMU were merged into one factor. Two items: ‘I engage in conversations on Facebook or Instagram’ and ‘I upload content on social media that is related to others’ showed factor loadings (.63 and .69 respectively, see Table 6) beneath the cutoff value of .70 (Farrell & Rudd, 2009). However, these loadings were still in the proximity of the cutoff value and therefore, the items were included in the reliability analysis to see if they could still help to form a reliable ASMU construct.
The Cronbach’s α of this factor was .65, which indicates that the scale is less internally consistent, and the items correlate less strongly with each other to measure the same underlying construct (Taber, 2018). Therefore, the item-total statistics table was consulted to see if the Cronbach’s α could improve when certain items would be deleted. The table showed that the Cronbach’s α could not be improved by deleting certain items (Table 7). Thus, the final Cronbach’s α of ASMU was .65. The study incorporated this factor, since the value of ASMU Cronbach’s α was close to the threshold of .70 (Taber, 2018) and the new factor for ASMU was computed.
TABLE 6
Factor & Reliability analysis Active Social Media Usage
TABLE 7
Active Social Media Usage: Reliability Analysis, Item-Total Statistics
Construct and items Factor
Loadings Cronbach’s α KMO Bartlett’s Test of Sphericity
Active Social Media Usage (ASMU_F) .65 .70 p < .001
ASMU_1_1 I engage in conversations on Facebook or Instagram.
.63 ASMU_1_2 I share other people’s posts on my own Facebook
or Instagram page.
.76 ASMU_1_3 I recommend other people’s Facebook or
Instagram page to my contacts.
.71 ASMU_1_4 I upload content on social media that is related to
others. .69 Item-Total Statistics Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted
ASMU_1_1 - I engage in conversations on Facebook or Instagram.
.37 .63 ASMU_1_2 - I share other people’s posts on my own
Facebook or Instagram page.
.50 .54 ASMU_1_3 - I recommend other people’s Facebook or
Instagram page to my contacts.
.44 .58 ASMU_1_4 - I upload content on social media that is
related to others.
23
4. RESULTS
Descriptive Statistics
The goal of 300 respondents for the study was reached. In total, 308 respondents filled out the survey. However, some respondents did not live up to the criteria of the study and were therefore deleted from the dataset. Specifically, out of the 308 respondents, 18 respondents indicated that they did not use any of these platforms to listen to music, and 40 respondents indicated that they did not use Instagram or Facebook, which reduced the total sample to 250. Therefore, the total sample used for the final analyses was 250.
Out of the total sample (N=250), most respondents were listening to music on Spotify (80.7%), followed by YouTube (16.6%) and Apple Music (2.8%). For the graphic representation, please view Appendix C1. Furthermore, most respondents mostly used Instagram (72%), followed by Facebook (28%) (see Appendix C2).
The sample consisted mostly of females (70.4%) and the age of the respondents was ranging from 14 years to 77 years with a mean age of 27 (M = 27, SD = 12). The sample mostly consisted of females, which is not completely representative for the population. However, as previously argued, gender is incorporated in the study as covariate, and therefore, the regression analysis will control for the possible effect of gender on the SME of the respondents.
Respondents were asked to enter the name of the artist, musician or band they listened to the last time they used Spotify, Apple Music or YouTube. The most entered artist is Coldplay with seven entries, followed by Avicii and The Script with both six entries. Other frequently named artists were Drake, Ed Sheeran, Post Malone and Snelle with five entries each and Adele, Billie Eilish and Maroon 5 with four entries each. For a graphic overview of these artists, please see Appendix C3.
The highest average form of Musical Engagement (ME) was Social-Identity ME (M = 4.87, SD = 1.17), followed by Affect-Inducing ME (M = 4.79, SD = 1.39) and Transportive ME (M = 2.57, SD = 1.50). Further, the respondents on average scored a 4.61 on average on brand loyalty (M = 4.61, SD = 0.91) and a 3.53 on Active Social Media Usage (ASMU) (M = 3.53, SD = 1.27). Finally, the average level of Social Media Engagement (SME) with artists was 2.46 (M = 2.46, SD = 1.42).
Relationships between Musical Engagement, Brand Loyalty and
Social Media Engagement
The first four hypotheses were tested using the mediation model 4 of Hayes’ PROCESS macro, version 3.4 (Hayes, 2017). This study includes three separate independent variables for ME, namely: Social-Identity ME (X1), Transportive ME (X2) and Affect-Inducing ME (X3). Therefore, model 4 of Hayes’ PROCESS macro was executed three times, with a different variable of ME included in the model as independent variable each time. Furthermore, the mediator brand loyalty (M) was included and SME (Y) was included as dependent variable. Additionally, the other two variables measuring ME were added as covariates, as well as the moderator in the theoretical model: ASMU (C1). Further, the
other control variables: Age (C2) and Gender (C3) were added as covariates, as discussed in the
methodology section. The overview of the results can be found in Table 8.
24 ME (a1 = .24), Transportive ME (a2 = .24) or Affect-Inducing ME (a3 = .10). This finding provides support for Hypothesis 2 (and all three sub-hypotheses):
Hypothesis 2. Musical Engagement is positively related to Brand Loyalty.
In the right column of Table 8 the results regarding the dependent variable SME (Y) can be found. It was found that when customers experience Social-Identity- and Transportive ME, their SME with the artists is higher too (Social- Identity c1 = .11 and Transportive c2 = .73, respectively). However, no
significant relationship between Affect-Inducing ME and SME was found. These findings partially support hypothesis 1, by confirming the following sub-hypotheses:
Hypothesis 1a. Social- Identity Musical Engagement is positively related to Social Media Engagement.
Hypothesis 1b. Transportive Musical Engagement is positively related to Social Media Engagement.
Furthermore, the results in the right column of Table 8 show that brand loyalty does not
significantly influence SME (p = .123). Thus, the results do not provide support for the hypothesis that brand loyalty is positively related to SME (hypothesis 3) and for the hypothesis that brand loyalty mediates the relationship between ME and SME (hypothesis 4), since a significant effect of brand loyalty on SME is required to establish a mediating effect. However, it was found that the direct effects of Social-Identity- and Transportive ME (c'1 and c'2, respectively) were lower (Δ = .03 for both
variables) than the total effects (c1 and c2). Besides, the direct effect of Social-Identity ME (c'1 = .08)
25 TABLE 8
Mediation effect of Brand Loyalty between Musical Engagement and Social Media Engagement
Active Social Media Usage as moderator
To investigate the moderating role of ASMU in the theory of the present study, Hayes’ PROCESS macro model 1 for simple moderation was used (Hayes, 2017). In this process, the moderating role of ASMU in the relationships of ME and brand loyalty with SME was studied. The PROCESS macro was run once for each independent variable (four times in total), while the control variables Age and Gender and the other independent variables were incorporated as covariates (please see Appendix E for the detailed reports of the different analyses).
The model incorporating all the independent variables (Social-Identity-, Transportive- Affect-Inducing ME and brand loyalty), covariates (Age and Gender) and the moderator (ASMU) overall explains 73 percent of variance in the dependent variable SME. The results of the moderation analyses are summarized in Table 9. As the main effect section shows, a significant positive effect of
Transportive ME on SME was found (β = .69). This means that the higher the Transportive ME of customers, the more they engage on social media with these artists. Besides, the results show that the moderator variable ASMU had a significant direct effect on SME (β = .12), which means that the more customers actively use social media, the more they engage with artists on social media. Furthermore, the results revealed that brand loyalty, Social-Identity ME and Affect-Inducing ME were not
significantly influencing the SME of customers with artists.
Variable Brand Loyalty (M) Social Media Engagement (Y)
β SE p 95% CI β SE p 95% CI
26 In contrast with theoretical expectations, the interaction effects were found not to be significant (Table 9). Thus, hypothesis 5 and hypothesis 6 were rejected. However, the analysis of the simple slopes (Table 10) of the interaction of brand loyalty and ASMU shows a significant conditional interaction effect (β = .20, t(241) = 2.02, p < .05) of ASMU on the relationship between brand loyalty and SME at the level of one SD above the mean of ASMU (SD = 1.27). This finding suggests that the brand loyalty of customers could have a stronger positive effect on their SME with artists when these customers are using social media more actively than average. This reveals signs of a possible
moderating role of ASMU in the relationship between brand loyalty and SME. However, these signs do not provide sufficient significant statistical support to confirm hypothesis 6.
TABLE 9
Moderation effect of Active Social Media Usage on the relationships of Musical Engagement and Brand Loyalty with Social Media Engagement
TABLE 10
Conditional interaction effects of Active Social Media Usage on the relation between Brand Loyalty and Social Media Engagement
Variable β t p 95% CI Model Main Effects Social-Identity ME .08 1.51 .133 -.02; .18 F(8, 241) = 81.07, p < .001; R2 = .73 Transportive ME .69 16.13 < .001 .61; .77 F(8, 241) = 81.71, p < .001; R2 = .73 Affect-Inducing ME .02 .46 .649 -.06; .10 F(8, 241) = 81.34, p < .001; R2 = .73 Brand Loyalty .13 1.75 .081 -.02; .28 F(8, 241) = 81.85, p < .001; R2 = .73
Active Social Media Usage .12 2.93 < .01 .04; .20 F(8, 241) = 81.85, p < .001; R2 = .73
Interaction Effects Model Change
Social-Identity * Active Social Media Usage .00 .01 .990 -.06; .06 ΔF(1, 241) = .00, p = .990; ΔR2 = .00
Transportive * Active Social Media Usage .03 1.17 .243 -.02; .09 ΔF(1, 241) = 1.37, p = .243; ΔR2 = .00
Affect-Inducing * Active Social Media Usage .02 .76 .448 -.03; .07 ΔF(1, 241) = .58, p = .448; ΔR2 = .00
Brand Loyalty * Active Social Media Usage .05 1.30 .195 -.03; .13 ΔF(1, 241) = 1.69, p = .195; ΔR2 = .00
Conditional effects of the focal predictor at values of the moderator(s):
Active Social Media Usage Effect se t p 95% CI
