MASTER THESIS
Business studies, UvA
23 jan. 2014
Involvement and ownership:
a consumer study into music consumption
Yuvraj Kumar
0352950
1st Supervisor: Dr. J.J. Ebbers
2nd Supervisor: Prof. Dr. N. Wijnberg
Table of contents
1. INTRODUCTION…...…...…...…...…...1
1.1 The context of the changing music industry…...…...1
1.2 Access VS ownership…...…...…...…...5
1.3 Involvement…...…...…...…...7
2. THEORY…...…...…...…...9
2.1 Music as a product…...…...…...…...9
2.1.1 Cultural goods…...…...…...…...9
2.1.2 Music's distinctive attributes…...…...…...11
2.1.3 Hedonic goods…...…...…...…...12 2.2 Involvement…...…...…...…...13 2.2.1 Product involvement…...…...…...16 2.2.2 Purchase-decision involvement…...…...19 2.2.3 Consumption involvement…...…...…...20 3. METHOD…...…...…...…...…...23 3.1 Sample…...…...…...…...…...23 3.2 Measurement…...…...…...…...…...23 3.3 Variables…...…...…...…...…...24 4 RESULTS…...…...…...…...…...…...30 4.1 Descriptive statistics...…...…...…...…...29 4.2 Collinearity analysis...…...…...…...…...…..32 4.3 Hypotheses testing...…...…...…...…...34 4.3.1 Regression 1...34 4.3.2 Regression 2 …...37 4.3.2 Regression 3 …...42 5. DISCUSSION…...…...…...…...…...…...46
5.1 Support for the hypotheses...…...…...…...…...47
5.2 Limitations and future research…...…...…...49
5.3 Conclusion…...…...……...…...…...53
1. INTRODUCTION
Within the consumer behavior literature the strand of hedonic consumption research investigates aesthetic goods such as music which have the capacity to generate unusually strong emotional involvement (Holbrook 1980). It has been noted that especially music - which 'readily pervades virtually every culture and every level of society' and as such 'may be considered as the most fundamental of the entertainment businesses' (Vogel, in Dolfsma, 2000)- should be an important topic of study for both consumer behavior and marketing (Lacher and Mizerski, 1994).
Yet research on music remains rather scant (Larsen et al., 2010). Similarly involvement, while being a crucial concept in consumer behavior literature (Park and Moon, 2003), is under-researched in the hedonic literature (Hightower et al., 2002;
Styven, 2010). This thesis aims to fill these gaps by exploring the recent changes in music consumption, due to the digitization of the music record, through the theory of
involvement.
1.1 THE CONTEXT OF THE CHANGING MUSIC INDUSTRY
The twenty-first century ‘crisis’ of digital reproduction (Lessig, 1999) of various copyright industries - film, photographic, publishing, videogame and software - has proven to be a fertile ground for academic research from various disciplines. It was however the music industry that had to face this digital crisis of reproduction first (Leyshon et al., 2005). At the heart of this crisis lies the alleged annihilation of the industry by the mp3 (Fox, 2004; Liebowitz, 2004; Sirois and Wasko, 2011), spurring the debate among economists on whether this is 'plain destruction or creative destruction' (Liebowitz, 2006; Handke, 2006, 2010).
Most economic studies tend towards the former view with the mp3 format serving as the 'provision' of a public good1 and the failure of copyright protection
1
(Handke, 2010)2. The digitization of music thus created a host of economic problems.
Music files lost their economic shell of protection given by copyright and became – analytically speaking – pure 'information goods'3 that are infinitely reproducible at zero
marginal cost (Shapiro and Varian, 1998).
This was disastrous in the industry's eyes, whose main market device is the creation and maintenance of 'artificial scarcity' (Hesmondhalgh, 2007). The industry started publishing figures of an increasing decline in sales and attributed this to the widespread phenomenon of mp3 file-sharing. The argument was that music now became free to consume with no necessity (or incentive) to purchase anymore. This resulted in a powerful lobby which enforced stricter legislation that greatly increased the scope of copyright. This was arranged in such manner that 'benefits corporate
interests at the expense of those of both artists and consumers' (Frith and Marshall, 2004, p.4). This then lead to numerous cases of the industry suing it's own customers (Hughes, 2005).
These legislative acts suggest that the stated purpose of copyright takes second note to offering protection to corporate interests (Hesmondhalgh, 2009)4.
The musician Janis Jan wrote in her now famous article, responding to the industry crisis:
Public goods express two key characteristics. Firstly, they are non-rivalrous, meaing that one person's consumption or use of one unit does not impose limits on another's use of it. So there is no scarcity
(consumption of the mp3 does not destroy or deteriorate it). Secondly, they are non-excludable, meaning that once they are produced it is not possible to prevent other people from using the goods (mp3's circulate without control over copying and usage).
2 The reasoning and rhetoric behind the institution of copyright is that it is a way to arrive at an optimal level of supply of cultural goods with innovation and creativity being key to let this diversity in supply prosper. And this innovation and creativity can only be maintained by maintaining an efficient market protection mechanism that ensures remuneration for the creators and thereby solving the problem of the free-rider. This reasoning has become the standard justification of copyright since the 1970s (Handke, 2010) and it's based on the theories of public goods (Samuelson, 1954) and information goods (Arrow, 1962).
3 Information goods differ from (most) tangible goods because, while fixed costs of producing the first copy tend to be very high, the variable costs of reproduction and marketing are very low (Landes and Posner, 1989).
4 Although copyright is a rather complex topic and lies outside the purpose of my paper it is worth noting that the one group that is actually not at all surprised by the record industry's draconian ways of pursuing copyright legislation and enforcement are historians of copyright (Johns, 2009, QuestionCopyright.org, 2006). They know, what the general population is slowly finding out: copyright was never enacted to to ensure payment for artist, but strictly for protecting distributors (QuestionCopyright.org, 2006). Adrian Johns’ rigorous historical analysis in fact shows how piracy has often even stimulated technological, intellectual and economic innovation rather than having inhibited it (Johns 2009).So copyright's is a story of conflict between producer and consumer which this thesis does not address.
'...let me remind you of something: the music industry had exactly the same response to the advent of reel-to-reel home tape recorders, cassettes, DATs, minidiscs, VHS, BETA, music videos ("Why buy the record when you can tape it?"), MTV, and a host of other technological advances designed to make the consumer's life easier and better. I know, because I was there.'5
Many authors note this industry response simply makes no business sense (Bakker, 2005; Sinha et al., 2010) with no thought being given to a basic marketing rule: provide what customers value (Condry, 2004; Fox, 2004).
Leyshon et al. (2005) on the other hand make the case – drawing upon a wider research project on e-commerce as well as findings informed by interviews with music industry informants, conducted both within Europe and the United States – that the problems of the music industry had actually accumulated over time as a result of bigger cultural forces that had altered the role of music for consumers, and that the rise of file-sharing was 'a decisive tipping point' (p. 183).
This pending crisis manifested firstly in the rising popularity of dance music and the club scene which resisted commercial exploitation because the independent dj was literally stealing the show. Attempts had been made to monetize on this market through branded cd's, but ultimately the real cultural (and economic) value of the music was created by ‘being there’. Secondly, popular music no longer commanded the attention of consumers on that same level as it once did. Popular music used to be young people's central cultural interest (Willis, 1990), but with the rise of other media and consumer electronics youngsters' cultural habits changed, resulting in less disposable income for music.
To strengthen their point further the authors argue that the industry had actually been close to crisis since perhaps the 1980s, but that it was able to stall this with a series reproduction technology innovations:
“..without the introduction of CDs, without classic hits radio, and without new markets in Asia and Latin America, popular music as we know it would no longer exist. The evidence is clear that popular music as a commodity is passing through a rapid transformation, assisted by the introduction of video games from Sega and Nintendo and related new technologies, which has
5 The Internet Debacle: An Alternative View at www.janisian.com/reading/internet.php (Originally published in Performing Songwriter Magazine, May 2002)
heightened the anxiety circulating among some in the music promotion business.” (Breen, 1995, p. 497–8; in Leyshon, 2005, p. 183)
This might be an explanation for the decline in music purchasing where file-sharing simply delivered the final blow to the commercial power of records.
At the same time we have been witnessing a steady growth in the live
performance sector. Popular music scholars Williamson and Cloonan report that in 2008 'for the first time in living memory, it was reported that British consumers spent more on live music than they did on recorded music...they did the same thing in 2009 and 2010, with the gap between the value of live music and that of recorded growing'
(Williamson and Cloonan, 2012,p. 13). Others note that there are indications that people don't spend less on music at all, just less on CD's (Bakker, 2005).
It has now become 'common parlance' to speak of a shift in revenues from recording to the live sector (Marshall, 2012; Williamson and Cloonan, 2007) or even the end of the traditional recording industry (Sandall, 2003; Williamson and Cloonan, 2012). But even though the live sector may be thriving as record sales go down, to speak of an actual case of substitution seems to me to disregard the very distinction between a live concert experience and the consumption of recorded music, which is repeated listening (Lacher and Mizerski, 1994). These are two different industries and products (Katz, 2010).
“A recording is one thing, a concert is another, and never the twain shall meet.”
(John Pfeiffer, music producer, composer and engineer)
Whatever the cause(s) of the decline in sales, the apparent lack of interest in actually owning music as an original (Rifkin, 2000b) combined with the sort of
'unspoken norm' that music is a public good that can be downloaded for free (Fischer, 2004; Depoorter et al., 2005) has eventually resulted in a belief that the digital music format for many consumers has no economic value. The music industry itself has started believing that popular music by itself is decreasing in value (McLeod, 2005; Sirois and
Wasko, 2011) and that it is not good enough anymore as a standalone product6.
Consequently we have witnessed experimentation with new business models that incorporate this widely held belief of free.
1.2 ACCESS VS OWNERSHIP
The prime example of this business model experimentation is the digital music service Spotify, which is now the second largest source of digital music revenue in Europe (IFPI, 2013). Here consumers have access to a nearly unlimited database of songs and albums. They can choose to either use the free advertising-based version or to pay for a monthly subscription which is ad-free. This service does not offer music to own and is exemplary of 'the declining importance of owning a piece of music' (Rifkin, 2000b, McGuinness, 2008) which seems to be in line with the transformation of modern markets from ownership to access (Humphreys and Giesler, 2007a; Rifkin, 2000a; Vargo and Lusch, 2004). Rifkin (2000b) noted that Napster – which spear led the disruption of the music industry – was the first of a new genre of businesses to work with a network mentality (i.e. access), instead of a traditional market-logic.
“At the heart of the controversy [downloading of music] is a clash that goes far beyond the music industry. What we're witnessing is the beginning of an epic struggle between two great economic systems. On one side lies the old market-based economy, made up of sellers and buyers. On the other side lies the new network-based economy, made up of servers and clients. In markets, the parties exchange property. In networks, the parties share access to services and experiences”. (Rifkin, 2000b)
So the notion of substituting unlimited access to music (through subscriptions) for sales of the individual music product lies at the heart of what is supposedly a global market transformation where access over a fixed time is being purchased.
“The access model is now a reality for millions of people across the globe. This is a model that works; it is no longer an experiment. We are here to stay.” (Ken Parks, chief content officer Spotify, quoted in
IFPI, 2013)
6 This is a widely held belief now both in the music industry at large and among independents. (Music in the Digital Age - What is its value?, Anth Gaskill, founder of Forbidden Fruit Records, jan. 2012
www.musicindie.com/news/1175)
It seems that what Rifkin predicted 13 years ago is indeed where the market is heading. However, the ongoing popularity of digital stores such as iTunes and Amazon and upcoming stores like Bandcamp -while not making up for the declining cd sales - shows many consumers still value the classic 'ownership' of music, even in the digital domain. Ten years ago Steve Jobs said at the iTunes launch event:
“We think people want to buy their music on the Internet by buying downloads, just like they bought LPs, just like they bought cassettes, just like they bought CDs. They’re used to buying their music and they’re used to getting a broad set of rights with it. When you own your music it never goes away.”(Jobs, iTunes
lauch Event 2003)
Since the iTunes launch digital music sales have been increasing every year. More recently in the physical domain vinyl has seen a sort of a resurrection. Sales are steadily increasing and not just among the older music fans (who were brought up with this format) but among younger groups as well (Felten, 2012; Kozinn, 2013). Hence, amidst all the changes in music consumption there is ample reason to assume that many consumers still value ownership. There is an affective relationship underlying
ownership of music (Holbrook, 1987), which digital music subscription services may not provide (Burkart, 2008; Gillespie, 2007; Styven, 2010).
So, although the shift towards access is widely accepted in both industry and academic circles (as the popularity of the service-dominant logic school7 in marketing
attests to) and many argue that the concert sector is replacing the recording industry, this thesis argues that consumers' interest in ownership of recorded music can be explained by their level of involvement in/with music.
1.3 INVOLVEMENT
“the greatest thing since sliced bread”
(Rothschild, 1984, p. 216)
7 The service-dominant logic (Vargo and Lusch, 2004) states that tangible goods serve as appliances for a provision of the service, which is what the customer ultimately values. The good does not in itself possess value. As such marketing is about delivering a service, rather than delivering goods.
Involvement has become a major area of research in consumer studies and marketing for well over 50 years now and it is recognized as the most important construct in the
consumer behavior literature (Brennan and Mavondo, 2000; O'Cass, 2000; Park and Moon, 2003). Although the concept of involvement has been used diversely (Cohen, 1983) and numerous definitions have been coined in the literature, involvement basically describes “a person’s perceived relevance of the object based on inherent needs, values and interests” (Zaichkowsky, 1985, 342). This ‘object’ can be: physical (as in a product), an activity (purchasing/shopping or consuming something), an idea (this may be a goal or objective) or an issue (like an environmental or social cause) (Hunt, Keaveney and Lee, 1995).
It can also be said that involvement lies at the heart of the person-object relationship (O'Cass, 2000). It is the relational variable that most strongly predicts product purchase behavior (Barber et al., 2009; Evrard and Aurier, 1996; Martin, 1998). Initially it was developed for traditional products (Slama and Tashchian, 1985;
Zaichkowsky, 1985), but it has been tested with hedonic goods such as music (Dixon, 1980; Puceley et al., 1988; North and Oishi, 2006). Walsh et al. (2003) argued in their research paper on the effects of digitization on music consumption that there is strong reason to believe that the consumer's level of music involvement affects the choice of music format, such as cd's over digital downloads. More recently Styven (2010) found that highly involved consumers still prefer the physical music format.
My aim is to further investigate whether consumers' involvement does explain their preference for ownership in the face of alleged shift towards access in modern markets. This paradigm shift from ownership to access (Humphreys and Giesler, 2007a) represents, what I believe to be, a robust theoretical lens through which to analyze the consumption of music. Only a few years ago numerous researchers had tried to
formulate music consumers´ decision making process as a dichotomous choice model of ´pirate or purchase´ (Coyle et al., 2009; Ouellet, 2007; Sinha et al., 2010; Taylor et al., 2009). My study moves away from the piracy and behavioral focus to emphasize the relationship between involvement, which is a mind-set (Mittal, 1989), and preference for ownership of cultural goods in the wake of the ongoing digitization.
This paper also distinguishes itself from past music involvement research by looking at multiple forms of involvement, rather than using only one form. Furthermore I integrate research from other disciplines to argue, as others have as well (Bolin, 2011; Hesmondhalgh, 2009), that academically speaking the changes we see in music
consumption are not so radical. Rather, they only bring to the surface more profoundly the public good characteristic of music because today - more than ever - listening to music is easy, convenient and free (McCourt, 2005) and does not require purchase anymore.
Some have gone as far as saying music will soon become a utility, delivered through pipes like our water (Kusek and Leonhard, 2005). What this only serves to illustrate however is a low involvement attitude towards a product/commodity which, while perhaps important, is not of personal relevance to the person (Zaichkowsky, 1985). Whether a consumer does actually want to purchase and own a widely available and accessible good such as music might very well be explained by their level of involvement.
The research will be conducted through an online survey. Since both access and ownership as theoretical concepts are dependent on their empirical counterparts in the digital and physical domain, access will be captured by the streaming service (Spotify), the free (music) video sharing site Youtube and the radio; ownership by physical records (CD and vinyl) as well as digital music (mp3's).
The main research question is:
'Do highly involved music consumers prefer ownership rather than strictly access?' Sub research question:
- 'Do highly involved consumers also value digital ownership?'
The paper is structured as follows: chapter two will cover the theory and
hypotheses; chapter three will discuss the method and data collection; chapter four will cover the results of the empirical research; the final chapter will present the discussion.
2. THEORY
2.1 MUSIC AS A PRODUCT
Lacher (1989) argued in her research into 'Music as a Product' for a blending of marketing and music literature to define what characteristics of consumers and what characteristics of music interact to produce consumption and purchase. Her work is now slightly out of date so I aim to provide a review of some more recent work to properly differentiate music from other products.
2.1.1 Cultural goods
Academics have for a long time grappled with the question of what differentiates cultural goods (such as music) from other products on the market (Straw, 2002). By isolating the unique characteristics of cultural goods academics could understand the structures and processes of it's production, distribution and consumption. Although a lot of contributions came from sociologists of cultural production (i.e. Hirsch, 1972; Miege, 1979) – which stressed the high level of 'demand uncertainty' of these goods – for this paper the interest lies in consumption.
The main contribution to the understanding of cultural goods came from the economic paper 'Information and Consumer Behavior' By Richard Nelson (1970), where the theory of experience goods was developed. Cultural goods and services are now often categorized as ‘experience goods’ (Towse, 2010), a concept introduced by Nelson (1970) to produce a theoretical framework for analyzing the relationship between the industry structure, it's product characteristics and consumer behavior. Nelson
contrasted experience goods with 'search goods':
• search goods are goods for which the key characteristics and value are known
prior to purchase (i.e. most utilitarian goods)
• experience goods are goods whose key characteristics are not known prior to
For experience goods8 there are two aspects that are differentiated in the literature (e.g.
Handke, 2010, Towse, 2010):
1. individual learning; that shapes the development of personal preferences. Past experiences with cultural goods (exposure to an artist/creator or a genre) shapes the preferences, taste and knowledge of the consumer.
2. quality uncertainty; the personal experience of consuming a specific cultural good is the most reliable way to assess utility for a consumer. So the consumer needs to experience the good in order to know it's value.
The first implies not only that taste and knowledge develops over time but also that in general consumer's willingness to pay increases with age and experience (Towse, 2010). This characteristic does not lend itself so easily for temporal/situational analysis however. For the present case of music, past experience shapes the preference of an individual indeed. Some authors have noted that music from a person´s youth often has the most impact on that person, even later in life it conjures up feelings that other genres are not capable of for the person (Holbrook and Schindler, 1989). However recent
research is not only lacking, but music consumers are getting older. While in the past the primary age of being heavily into music was from the teens to mid twenties (Dixon, 1980), today many music consumers are older than 30 (IFPI, 2005).
The second characteristic of cultural goods describes how a cultural good needs to be experienced/consumed in order to know it's value. So the 'use value' lies in the 'raw experience' itself, and cannot be predetermined from specific product
characteristics and can only be evaluated and reflected upon post-consumption.
This may lead to – what is commonly known as in the information good literature – the ‘paradox of disclosure’ (Arrow, 1962), which refers to the fact that for consumers to adequately assess their willingness to pay for (information) goods disclosure of the good is required. But that very disclosure then actually reduces consumer's willingness to pay (wtp as the maximum amount of money the consumer wants to pay to obtain the good).
8 It should be mentioned that both Nelson himself as well as much of the literature after wards stress that there are almost no practical examples of either pure search goods or experience goods (Handke, 2010). But ofcourse categorization is never applied for the sake of categorization.
This results in a lower market value or worse a lack to even want to buy the good (Shackle, 1952).
This economic problem has often been mentioned in regards to the digital music product (Liebowitz and Watt, 2006) and it is hard to argue that for a large segment of the music consumer population (who just want to hear music without any hassle), this does apply.
2.1.2 Music's distinctive attributes
The theory of information goods provides us with some insight into music as a good, but the consumption and purchasing of music records differs from other cultural goods such as films, books and single-player video games in three seperate but directly related ways that has implications how it is marketed and consumed (Straw, 2002; O´Reilly et al., 2013):
• repeated consumption; Generally music does not employ such a narrative structure
that culminates in an (often climaxing) end or revelation (Straw, 2002). Music is meant to be heard repeatedly. This is also what differentiates the two main forms of music
consumption: live concerts and recorded music (Lacher and Mizerski, 1994). Music, as most of us know it, is experienced through repeat listens and this makes the record a very durable item as consumption does not decrease it´s usability.
Tied to this is another distinguishing feature, which also relates back to quality uncertainty characteristic mentioned above:
• consumption prior to or without purchase; As consumers we try to limit uncertainty
of purchase through recommendations, reviews, trailers, former experience with the author/director/actor etc. Movies and books (like most consumer goods) are marketed in such a way that ultimate consumption occurs after purchase and consumers generally accept this quality uncertainty. Music however is promoted and sold in a way that you first experience the song (or entire album if you wish), what is commonly called
'sampling' and then decide whether you wish to purchase it. Radio programmers and marketers of course know that prior exposure is a major factor in the desire to purchase music (Straw, 2002).
Traditionally this held more strongly for popular music than for niche genres which receive little airtime and discussion in the media or promotion through tours (Lacher and Mizerski, 1994). Also the purchase was not required unless the consumer wanted to experience the music more than one time and desired to be in control of the music (Lacher 1989).
• active or passive; People often listen passively to music (Bradshaw and Holbrook,
2008) because it is a) so all-pervasive in our daily lifes and b) it is free and can be listened to easily and for free via numerous sources (Mizerski et al., 1988). Often music then serves as a distraction. Active consumption is the type of consumption where people are more involved with music (Shankar, 2000), choose to attentively listen to a specific artist or song and engage in concert attendance, record purchasing and music information search.
These three characteristics sketch traditional profiles of a music purchaser: he/she desires repeat listens, wishes to be in control of the music and listens actively and attentively. The first two of these characteristics however do not hold up anymore in the digital landscape where access provides for the possibility to re-experience and be in control. This means that depending on how active a person consumes music (i.e. how involved he/she is), purchasing may or may not occur. So the level of involvement could be a key predictor of a consumer's preference for ownership of music.
2.1.3 Hedonic goods
Economic literature distinguishes, as mentioned, between search and experience goods. Marketing and consumer behavior literature have developed a distinction between utilitarian and hedonic goods. Utilitarian goods are goods that are purchased for their utilitarian function. Traditional consumer research has tended to focus on utilitarian goods such as toothpaste, cigarettes, food etc. (Holbrook 1980). In contrast, hedonic consumer research investigates the performing arts (ballet, dance, theater), the craft arts (painting, photography, sculpture) and popular culture (movies, pop music, fashion clothing).
The focus of hedonic research is on 'the multisensory, fantasy and emotive
(still) dominant information-processing model with it's overly cognitive focus (Marsden and Littler, 1998), that typically shackles our perspectives (Tadajewski and Hewer, 2011). Aesthetic products like music records, novels, plays and movies play an
important role in the life of the consumer and they provide him/ her with entertainment and function as an outlet for emotion and fantasy (Hirschman and Holbrook 1982).
Although all products can of course be hedonically experienced by consumers, hedonic goods have the capacity to generate unusually strong emotional involvement (Holbrook 1980). Especially aesthetic products such as music, movies, concerts and novels can evoke complex and rich fantasies and fulfill deep and salient emotional needs (Hirschman and Holbrook 1982). Their consumption can represent an act of importance to the consumer.
It is somewhat suprising then that despite the seminal work of Holbrook and colleagues and their emphasis on the high involvement that encompasses hedonic goods, Hightower et al. (2002) note that involvement in hedonic literature remains under-researched.
2.2 INVOLVEMENT
A vast amount of empirical and theoretical work has been devoted to the concept of involvement. This effort has been driven by researchers' desires to better understand the various ways in which consumers become involved with products, to find the effect of involvement on purchasing and consuming behavior (Browne & Kaldenberg, 1997; Traylor & Joseph, 1984) and behavior related to possessions (Laurent & Kapferer, 1985; Mittal & Lee,1989; Slama & Tashchian, 1985; Zaichkowsky, 1986). It has proven to be the relational variable that most strongly predicts product purchase behavior (Barber et al., 2009; Evrard and Aurier, 1996; Martin, 1998). It has also been researched as a predictor of other behavioral outcomes such as the willingness to pay a premium price (O'Cass and Choy, 2008), repurchase loyalty (Olsen, 2007) and frequency of product usage (Mittal and Lee, 1989).
The concept of involvement has been used diversely however (Cohen, 1983) and numerous definitions have been coined in the literature since it was first introduced in marketing by Krugman in 1965 (O'Cass, 2000). In the literature studies for a long time
focused on only one form of involvement: either product involvement (Bloch, 1981; Traylor and Joseph, 1984), purchase (decision) involvement (Mittal, 1989) or situational involvement (Clarke and Belk, 1978). This eventually resulted in two dichotomies: situational/enduring and purchase/product involvement (Mittal and Lee, 1989).
Situational involvement is temporary and in a particular situation (i.e. when buying a gift) whereas enduring involvement is about an ongoing interest with a product or activity (Rothschild, 1979; Huang, 2006), that transcends situational influences (Houston and Rothschild, 1978; Laurent and Kapferer,1985; Richins and Bloch, 1986). Purchase involvement refers to a consumer's interest in shopping where-as purchase-decision involvement indicates how important the decision to purchase is to the consumer.
However product involvement and enduring involvement were often one and the same thing (Bei and Heslin, 1997; Bloch and Bruce, 1984). The construct overall has thus been met with criticism because of the confusion in terminology (Mittal and Lee 1989; Poiesz and de Bont, 1995). O'Cass (2000) addressed this confusion and lack of distinction between types of involvement (Batra and Ray, 1983) and proposed a model of four forms of involvement, which are relevant to the consumer's environment, and argued to 'maintain involvement as an enduring relationship between a consumer and an object' (p. 553) instead of a temporary or situational one.
Using O'Cass' (2000) conceptualization, involvement refers to the extent to which the consumer views the object as a central part in their life; meaningful, engaging and important. This focus on enduring reflects past researchers calls to focus on enduring involvement as this reflects a deep stable trait that cannot be manipulated and will not change easily (Richins and Bloch, 1986). The adoption of this enduring nature
involvement can be seen in more recent studies on involvement (Barber et al., 2007; Kolyesnikova et al., 2009)
High levels of involvement imply a relatively complete engagement of core aspects of the self in the object. Low or even no involvement implies apathy or
automatic behavior towards the object. This definition refers to all four forms, they just refer to a different context. Together they cover the whole environment of interest to the consumer (and consequently to consumer researchers):
1) product involvement: involvement in the product on a general level 2) purchase-decision involvement: involvement in purchasing and purchase
decision making of the product
3) consumption involvement: involvement in the consumption experience of the product
4) advertisement involvement9: involvement in advertisements related to the
product
2.2.1 Product involvement
Product involvement is a goal-oriented state of enthusiasm, perceived relevance,
interest, and excitement that consumers exhibit towards a product (Barber et al., 2009). It describes the potential of the product to deliver personal relevance (Higie and Feick, 1989). Others note that it refers to the pleasure obtained from thinking about or using the product/engaging in the activity (Richins & Bloch,1986). High product involvement implies a stronger sense of possession of the object and low price sensitivity (Beatty, Kahle, and Homer, 1988; Lichtenstein, Bloch and Black, 1988). Consumers exhibiting high levels of product involvement are willing to pay more for the product.
Furthermore a very important feature is that this form of involvement describes the degree that the product/activity is related to consumer's self-image (Higie & Feick,1989). In the context of hedonic goods, very high levels of involvement plays a central role in shaping a person's identity (Holbrook, 1987). In studies of Popular Music and Media and Culture music has long been recognized for it´s identity creating and strengthening power (Bennett, 2000; Frith, 1995). Consumers had for long expressed their identity through their record collections (Holbrook, 1987). As Russell Belk wrote in his seminal paper 'Possessions and the Self' (1988): we cannot understand consumer behavior without gaining first some understanding of the meanings that a consumer 9 While the research for traditional consumer goods stresses the importance of advertisement involvement (as can be witnesssed from the vast amount of literature coming from this corner of research), I decided to exclude this factor for two reasons: firstly, the marketing of music does not rely on advertisements so much as a communication channel (word-of-mouth plays a bigger role here). Secondly, advertising involvement has been noted to be a situation-specific construct, related more to behavioral outcomes such as actual purchases made (Batra and Ray, 1983). O'Cass himself later also noted that this form is not enduring compared to the other constructs of involvement (O'Cass and Choy, 2008).
attaches to possessions. Possessions are a part of ourselves. Acquiring, possessing, and collecting are the fundamental processes in consumer behavior (Belk, 1982).
While record collecting lies at the (extreme) high end of involvement (Bloch, 1986; Holbrook, 1987), the core theme of investing one's self in owning possessions (Belk, 1988) is the reason why consumer researchers often invoke the construct of involvement (O'Cass, 2000). Furthermore, the sense of acquiring and owning are worldwide still deeply embedded in consumers (Belk, 2009).
Ownership is a term that carries traditional economic connotations and is directly related to value and utility. But while possessing is a term more frequently used in consumer literature both refer to the processes of acquiring and possessing music. But even in economic terms ownership is not simply about the price users pay for music, or the utility they assign to it, or even the means through which they acquire it (pay-per-purchase vs pay-for-access). It's also connected to the relationships consumers form with the product (Morris, 2010, 2011).
Popular music professor Burkart (2008) tells of his conversation with Tarleton Gillespie in which they talk about two kinds of ownership in music: the first is financial ownership of the music product. I bought this, hence I own it. The second, and far more interesting, is what Gillespie calls cultural ownership. This is my collection, it is part of who I am (Belk, 1995). Cultural ownership is the more affective relationship between the consumer and the product (Fournier, 1998) and becomes more important then to
understand in the present day where consumers give up ownership for access to services (Burkart, 2008).
The first work on measuring music involvement came from Dixon (1980) who developed a music involvement scale that measured: the number of hours spent
listening to music, the frequency of concert visits, the number of purchases of recorded music, and the percentage of listening time they spent really listening. This is a
behavioral manifestation of involvement similar to the scale developed for traditional products (i.e. Engel and Blackwell, 1982). The aim of Dixon's study was to test how involvement related to music genre tastes. At the time (more than 30 years ago), new genres were coming up rapidly as the author noted himself. This effect on genre taste
however does not concern this study and in fact it is fairly safe to say that we don't see the rapid creation of new ground-breaking genres today anymore.
Pucely et al. (1988) researched music purchases and enduring music involvement. Involvement was found to be positively -but not strongly- associated with the intention to purchase, and strongly associated with the time an individual spends really listening and negatively to the propensity to be distracted by music while doing other tasks (i.e. active vs passive listening). Their work shows that enduring involvement related more strongly to active music consumption rather than to purchasing behavior.
Flynn, Eastman, and Newell (1993) studied rock-music consumption and found that consumers who were highly involved with rock music scored higher on reading rock magazines, money spent on rock-music, and time spent listening to rock. North and Oishi (2006) researched CD purchasing among British and Japanese consumers and found that involvement and the need to be in control (Lacher, 1989) had a strong
correlation with purchase behavior. Unfortunately very little attention is given to the theory of involvement.
Styven (2010) found that enduring music involvement is positively related to preference for the tangible physical format. However her involvement construct consisted of only two items, which is rather minimal if we keep ourselves to the multi-item recommendations of Bearden, Netemeyer and Mobley (1993).
There exists very little research that looks at consumer acquisition and possession processes in regards to digital cultural goods. Sterne (2006, 2012) argues that consumers treat MP3s as ´cultural artifacts´, which do still possess value. Research into collecting in the digital realm (McCourt, 2005) remains sparse however (Handke, 2010). Nevertheless, the success of digital stores like Amazon and iTunes is an indication that despite the downloading and file-sharing consumers still purchase and value owning the digital format.
H.1 High product involvement leads to preference for ownership of instead of access to music.
2.2.2 Purchase-decision involvement
Purchasing decision involvement, the second form of involvement, as used here is a combination of two seperate involvement forms (O'cass, 2000): purchasing involvement and purchase-decision involvement. The former was initially developed by Slama and Tashchian (1985) to find the relation with demographic characteristics. Purchasing is an activity that some consumers are in general much more involved in than others. Since purchasing was traditionally done at retail stores, this construct was primarily used to define shopping segments (Lokshin et al., 1997).
Purchase-decision involvement, the later developed and more widely used scale by Mittal (1989), refers to the extent to which a consumers cares about what he/she buys and correspondingly the extent to which he or she is interested in the choice making. It also entails (for traditional goods) a scrutiny in choosing the right brand. Mittal (1989) argues that purchase decision involvement is the type of involvement with which marketing practitioners should be most concerned with because purchasing involvement is related with effort in shopping (Slama and Tashchian, 1983) and obtaining the best value (Ohanian and Tashchian, 1992). Marketing managers are interested in positioning their brand in the mind of the consumer and to make sure the purchase is one of interest, concern and care.
In regards to looking for best value there is a contrast with product involvement as product involvement entails low price consciousness. Lichtenstein, Bloch and Black (1988) for instance found a strong link between WTP and a consumer’s product
involvement. Consequently, consumers more concerned with the product than with price are characterized as product involved and these consumers generally are more willing to pay higher prices. Persons high in purchasing involvement are expected to respond more to deals and discounts, whereas consumers with low purchasing
involvement would likely be more interested in anything that would reduce the effort involved in shopping.
An example of a combination of high purchase-decision and low product involvement is when a person may be highly involved in the purchase of a new car but may have no enduring involvement in cars beyond that time and situation when it is necessary to make a purchase. We would expect then that consumers with low
purchase-decision involvement would be more inclined towards access due to the reduced load (or in fact almost zero load) on the need to purchase. Consumers high in purchasing decision involvement enjoy and care about purchasing the product and making decisions which new album to buy, so likely these consumers will prefer ownership.
Product and purchase-decision involvement often occur separately. However when a product is involving for a consumer the purchase-decision will not be made casually. Thus “product involvement is an antecedent of purchase-decision
involvement” (Mittal and Lee, 1989 p. 365), meaning the former is one of the causes of the latter, although not a necessary cause (as antecedent does not imply this). This would lead us to assume that some consumers who are are highly involved in music are also highly involved in purchase-decision making of music.
H.2 High purchase-decision involvement leads to preference for ownership of instead of access to music.
H.2a High purchase-decision involved consumers value owning music on mp3. H.3 Product involvement is an antecedent of purchase-decision involvement in the relationship with preference for ownership.
2.2.3 Consumption involvement
The last form of involvement identified by O'Cass (2000) was actually not
operationalized as a stand-alone scale in prior research, but was – after rigorous testing and validating – developed as a separate dimension. Formerly the aspect of the
consumption of the product was often lumped together with product or enduring involvement. Theory on this construct is thus unfortunately lacking but the strong emphasis on the act of consuming (wearing fashion clothing in the research of O'Cass) captures the experiential aspect of consuming. The focus on the actual consumption experience locates it outside the purchase and general interest in the product. It's about (in this case) the listening experience and the feeling of fulfillment and satisfaction it gives.
It would seem reasonable to assume that consumers who are highly product involved with music are also highly consumption involved. At the same time
considering the fact that music can now be consumed for free without purchase through access, it's possible that high consumption involvement consumers prefer access rather than ownership. So I expect to find a positive relationship but a weak one however.
H.4 High consumption involvement effects preference for ownership of instead of access to music.
H.4a High consumption involved consumers value owning music on mp3. H.5 Product involvement is an antecedent of consumption involvement in the relationship with preference for ownership
3. METHOD
3.1 SAMPLE
My survey sample consists of friends, relatives, fellow business students and internet forum visitors. The questionnaire was sent out via:
- a selection of my facebook contact list (65 people) - personal email to friends and relatives (15 people) - the UvA Business Studies facebook page and
- posting on the FOK forums, one of the largest forums in the Netherlands covering all fashionable topics one can think of.
Given my personal experience with receiving many (often unappealingly formulated) emails with questionnaires, I opted for a more captivating approach by using a bit of humor. I also decided to give an accurate and honest description of the time it takes to finish the survey (5 – 8 minutes). Furthermore I asked if people would be so helpful as to forward it to their friends. On the FOK forums I created a more elaborate introduction writing to explain the specific theme (access vs ownership) of my research and invited people to start a discussion on the topic. The survey design invited both critique and praise. The feedback was incorporated. A reminder was sent/posted after one week.
3.2 MEASUREMENT
Questionnaire design• The independent variables were tested in questions 6 – 9; • the dependent variables in question 10
• the control variables in questions 1 – 5 and 11 and further.
Instrument
• The dependent and independent variables were tested by indicating the extent to which the respondent agrees with a statement. The scale was measured with a 7-point
Likert scale (ranging from 'totally disagree' to 'totally agree') with an 8th scale for ´not applicable to me´.
• Control variables were measured primarily on categorical level (dummy variables were created where needed).
3.3 VARIABLES
Independent variablesMy independent variables consist of three involvement constructs taken from O'Cass (2000). While the research for traditional consumer goods stresses the importance of advertisement involvement (as can be witnessed from the vast amount of literature coming from this corner of research), I decided to exclude this factor because marketing of music does not rely on advertisements so much as a communication channel (word-of-mouth plays a bigger role).
As mentioned before involvement refers to the extent to which the consumer views the object as a central part in their life; meaningful, engaging and important. This definition applies to all the forms of involvement, they just refer to a different context (the product, the purchase decision or the consumption of the product). Furthermore although O'Cass tested his model of involvement on fashion clothing, his
conceptualization of involvement does not require different definitions or changes in the scales for different products. They are applicable to a consumer's involvement with any given object.
High levels of involvement imply a relatively complete engagement of core aspects of the self in the object, in this case music (product, purchase and consumption). Low or even no involvement implies apathy or automatic behavior towards the object (O'Cass, 2000). Two of the three constructs - product involvement and purchase decision involvement – stem from a long line of research in consumer behavior studies.
Consumption involvement however was newly introduced by O'Cass (2000) to distinguish between involvement with the product in general and involvement in/during the actual consumption experience.
Two key terms in all of the items are 'importance' and 'relevance'. As described in earlier work (Mittal, 1989, 1992; Mittal and Lee, 1989; Zaichkowsky, 1986), these terms
are part of involvement but do not alone constitute involvement perse. So consumers might view the purchase-decision as important without being actually involved in the purchase-decision. So many consumers might view music as important or relevant, however they may not have any attachment to and/or significant interest in the product. They might be involved only when actually listening to music without involvement in purchasing or general involvement with the music product.
1. Product involvement
This scale contains items that measure the extent to which the consumer is involved in the product. The original scale contains 15 items. Items contain questions such as '[This
product] means a lot to me', 'I am very much involved with/in [this product]' and 'I would say [this product] is central to my identity as a person'. I deleted the item 'Some individuals are completely involved with [this product], attached to it, absorbed by it. For others [this product] is simply not that involving. How involved are you with [this product]?'. This item was found to
be too long. North and Oishi (2006) also deleted this item in their research on music consumption.
The main thrust of this construct is that it represents the consumer's enduring involvement with the product, rather than the situational involvement such as when a product is on sale, when a product is needed for at a certain time (new refrigerator to give the worst example) or when a new product is introduced that becomes popular and is desired at that moment for trendy reasons. As such it is an an inner state of the
individual that reflects a long term product interest or attachment (Bloch, 1981), meant to represent a stable characteristic that may only change over a longer period of time and/or with changes in the consumer´s value set (O´Cass, 2000). For example students' involvement with partying that is present during college years will undoubtedly
become lower as the person begins working full-time after graduation. Or teenagers' involvement with rock music will undoubtedly decline as they reach adulthood (Richins and Bloch, 1986).
→ Reliability analysis of this scale yielded a Cronbach's alpha of .978.
This scale measures the degree to which the consumer is involved in the purchasing and purchase-decision of the product. The original scale contains 10 items with questions such as 'Making purchase decisions for [this product] is signifcant to me', I attach great
importance to purchasing [this product]' and 'I like being involved in making purchases of [this product]'. I deleted 2 items resulting in an 8 item scale. The first was the exact same item
as in product involvement that I deleted (only now with purchase-decision in it). The second item 'I think a lot about my choices when it comes to [this product]' did not properly explain whether it refer to the purchase aspect, which made it's meaning ambiguous in the context of music (people might interpret it as the choice which music to listen to).
→ Reliability analysis of this scale yielded a Cronbach's alpha of .971.
3. Consumption involvement
This scale contains items that measure the extent to which the consumer is involved in/during the consumption of the product. The scale contains 7 items with questions such as 'I feel a sense of personal satisfaction when I wear [this product]', Using/consuming
[this product] is one of the most satisfying and enjoyable things I do', and 'Using/consuming [this product] is important to me. High levels of consumption involvement imply active
listening (Shankar, 2000). Many people listen passively to music (Bradshaw and Holbrook, 2008), they are not engaged or very little. Music is after all-pervasive in our daily lifes and may be taken for granted: it can be listened to easily and for free via tv, radio (Mizerski et al., 1988) and online. So the person might enjoy music and even listen to it a regular basis it might not be important or significant to his or her life.
The strong emphasis on the act of consuming (wearing fashion clothing in the research of O'Cass) captures the experiential aspect of consuming rather than
involvement on a general level with a product. This experiential focus lies at the heart of hedonic consumption literature (e.g. Hirschman and Holbrook, 1982), which moves away from the traditional information processing stream of research that focuses only on purchasing.
→ Reliability analysis of this scale yielded a Cronbach's alpha of .941.
My main dependent variable captures consumers' preference for ownership of music (be that in the form of CD, LP or mp3) rather than access to music (with a Spotify account, via Youtube or radio). Subsequently I investigate whether this preference for
ownerships also means that consumers value ownership of the digital format.
1. Preference for ownership instead of access
´For me it is more valuable to own my music ( on CD, LP, mp3) rather than merely having access to music and not owning it (i.e. with Youtube, Spotify or radio).'
1a. Value ownership of digital format
'I value owning my music on mp3 (or some other digital format).'
Control variables
In the context of my case of music consumption I will control for:
1. Music-specific factors
In the liteture a few key factors have been identified that have significant effect on the consumption of music: genre (Leyshon et al., 2005), pop concert attendance (Bakker, 2005; Williamson and Cloonan, 2012) and music listening device (Molteni and Ordanini, 2003; Styven, 2010). Furthermore I control for the amount of music consumers
download.
- genre: three categories are used: indie, dance/club/house/techno and jazz/classical (indie fans may perceive indie to be more unique and hence more worth to own for instance, dance genres usually imply experiencing the music in the club and
jazz/classical are genres popular among collectors and connoisseurs (Holbrook, 1987)) - frequency of pop concert attendance (whether concert attendance is a substitute for purchasing)
- digital music device as preferred audio system: (whether the consumer uses one of the following three as his/her preferred device for listening to music: pc/laptop, mp3
player or mobile. Listening primarily on one's pc may for instance indicate the consumer prefers the ease and convenience of access)
- frequency of downloading (whether download intensity relates to preference for mp3)
2. Demographics
In consumer research socioeconomic and demographic variables – age, gender, income and education – were traditionally considered very important bases to segment (Slama and Tashchian, 1985), but it is recognized that these factors may be less relevant for marketing today (Cova and Cova, 2002; Nutall et al., 2011) because our societies are becoming increasingly more technologically advanced (Flynn et al., 1993) and for music specifically this holds true as music is deeply intertwined with technological
innovations.
- age - gender - income level - education level
4. RESULTS
4.1 DESCRIPTIVE STATISTICS
Tables and graphs are listed below. Noteworthy statistics are:
a sample size of N = 154.
average age of 30 with a mode of 28.
a male-female ratio of 3:2 (94:60), which is fortunate as samples for research into music consumption predominantly consist of males (Styven, 2010).
a highly educated sample with more than 50% of the sample possessing a university degree.
a relatively small percentage of students (22%) and a reasonably low percentage of unemployed people (6,5%), which is fortunate for this research since – as mentioned in the introduction – most research looking into music purchasing focus on students and this was often noted as being a limitation (e.g. Chiang and Assane, 2009; Filimon et al., 2010; Sinha and Mandel, 2008)
income levels are concentrated on the first three scales.
These statistics indicate that this sample mostly represents the younger working middle-class with a reasonable to good education.
Statistics
Age Gender Education Employement N Valid 138 154 154 153 Missing 16 0 0 1 Mean 30,30 1,39 5,50 1,85 Mode 28 1 7 1 Std. Deviation 11,059 ,489 1,401 ,985 Minimum 14 1 1 1 Maximum 71 2 7 4
Gender
Frequency Percent Valid Percent Valid MALE 94 61,0 61,0
FEMALE 60 39,0 39,0
Total 154 100,0 100,0
Education
Frequency Percent Valid Percent Valid High School (MAVO/VMBO) 2 1,3 1,3
High School (HAVO) 3 1,9 1,9 High School (VWO) 10 6,5 6,5 MBO 16 10,4 10,4 HBO 40 26,0 26,0 University (undergraduate) 36 23,4 23,4 University (postgraduate) 47 30,5 30,5 Total 154 100,0 100,0 Employement
Frequency Percent Valid Percent Valid Paid employment 77 50,0 50,3
Self employed 32 20,8 20,9 Student 34 22,1 22,2 Unemployed 10 6,5 6,5 Total 153 99,4 100,0 Missing System 1 ,6 Total 154 100,0 Income level
Frequency Percent Valid Percent Valid 0 to € 800 38 24,7 25,2 € 800 to € 1600 35 22,7 23,2 € 1600 to € 2400 42 27,3 27,8 € 2400 to € 3000 21 13,6 13,9 € 3000 to € 3800 9 5,8 6,0 more than € 3,800 6 3,9 4,0 Total 151 98,1 100,0
Missing System 3 1,9
4.2 COLLINEARITY ANALYSIS
Multicollinearity poses a problem for the multiple regression model when two or more independent variables are strongly correlated (e.g. 0.9), leading to a distortion of the beta coefficients (Field, 2009). In an exploratory correlation analysis between the independent variables I noticed a very high collinearity (0.907) between product involvement (PI) and consumption involvement (CI).
In the subsequent model I checked 'collinearity diagnostics' to examine – what is the most common method to diagnose multicollinearity – the variance inflation factor (VIF). This can be seen in the coefficients table (table 1). Here we see VIF values of 5.717 and 5.970 for CI and PI respectively. Now, although Field (2009) cites Myers (1990) who suggests a VIF value greater than 10 is cause for concern, recommended maximum VIF values of 5 (e.g. Rogerson, 2001) and even 4 (e.g. Pan & Jackson, 2008) can be found in the literature.
A final step in analyzing multicollinearity is to look at the variance proportions (table 3) to see if there are independent variables that have high proportions on the same eigenvalue (Field, 2009). Furthermore we should only look at small eigenvalues, i.e. the bottom rows of the table. This would indicate they are dependent (Field, 2009). Here we find values of .97 and .92 for CI and PI on dimension 5 (the lowest eigenvalue).
This is the final indication of their interdependency. As is known unfortunately however there's not much that can be done about it. Field (2009) notes there is no way of knowing which variable to omit and that furthermore theoretical conclusions about the subsequent new model (where one of the high VIF variables is omitted) are meaningless. The only option here is to keep product involvement in the model since that is the most theoretically developed construct10.
10 The analysis for the second dependent variable (value digital format) is not included, but yielded identical figures: VIF values of 6.0 and 5.7 and variance proportions values of .92 and .97 (for CI and PI resp.).
TABLE 1 Coefficien tsa Model Unstandardized Coefficients Standard ized Coefficie nts t Sig. Colline arity Statisti cs B Std.
Error Beta Tolerance VIF 1 (Constant) ,752 ,733 1,026 ,306 CONSUMPTION INVOLVEMENT ,171 ,288 ,093 ,595 ,553 ,175 5,717 PRODUCT INVOLVEMENT ,102 ,266 ,062 ,385 ,701 ,167 5,970 PURCHASE DECISION INVOLVEMENT ,581 ,075 ,537 7,741 ,000 ,894 1,119 a. Dependent Variable: PREF1 I prefer owning music (as CD, LP, mp3) rather than being able to only 'access' music (like with Youtube, Spotify or radio). TABLE 2 Collinearity Diagnosticsa Mode l Dimension Eigenvalu e Condition
(Constant) CONSUMPTIO N INVOLVEMENT PRODUCT INVOLVEMENT 1 1 3,799 1,000 ,00 ,00 ,00 2 ,168 4,760 ,02 ,00 ,00 3 ,029 11,378 ,82 ,02 ,08 4 ,004 31,140 ,16 ,97 ,92
a. Dependent Variable: PREF1 I prefer owning music (as CD, LP, mp3) rather than being able to only 'access' music (like with Youtube, Spotify or radio).
4.3 HYPOTHESES TESTING
To test my theoretical model I employ a multiple regression using blockwise entry in SPSS. The three remaining main hypotheses propose that the two consumer's levels of involvement are positively and strongly correlated with their preference for ownership of music and that furthermore product involvement is an antecedent of purchase-decision involvement. The sub hypotheses subsequently test whether consumers' levels of involvement determines their valuation of owning music on a physical and digital format.
The hypotheses will be tested in three separate regression analyses. Regression 1 will test the interrelationship of the three main variables, more specifically whether there is mediating effect (H.3). Regression 2 will test the effect of the independent variables on the main dependent variable (H.1 and H.2). Finally regression 3 will investigate whether involved consumers also value digital ownership (H.1a and H.2a).
4.3.1 Regression 1
The first regression aims to investigate whether there is a case of mediation among the variables. This is done in three subsequent steps11. In the first step I regress PI on
purchase-decision involvement (PDI). In the summary (table 1) we see that the
correlation of PI with PDI is .307. In the coefficients table (table 2) we see that PI has a significant effect on PDI (a = .471 and SEa = .118 with p = .000)
In the second step I use block-wise entry to first regress PI on preference for ownership (PREF1) and then PI and PDI on PREF1 (table 3). In model 1 we see that the correlation between PI and PREF1 is .311 and that PI accounts for .097 of the variance in PREF1. In model 2 we see that the multiple correlation of PI and PDI with PREF is .594 and that these together account for .353 of the variance in PREF1.
In the coefficients table (table 4) we see that the B coefficient of PI is initially .516 with p = .000 but becomes more than twice as small after accounting for PDI in the
11
I followed the steps outlined by Baron and Kenny (1986)
second model (B coeffcient = .245 with p = .033). So the effect of PI on PREF1 becomes more than twice as small and less significant. Furthermore for PDI we see that b = .576 and SEb = .074 with p = .000. This would us to assume there is a case of partial
mediation of PDI on PI here.
The final step entails inputting these values into the Sobel test12 which gives a
significant result of p = .000.
→ H.3 is supported
Step 2 and 3 are repeated (table 4a) to assess the mediation with the second DV (PREF2), The b coefficient becomes less than twice as small but the effect is now insignificant with PDI being signficant (b = .190 and SEb = .076 with p<.05). The Sobel test gives a
significant result. So here we find full mediation.
TABLE 1 Model Summary Mode l R R Square Adjusted R Square Std. Error of the Estimate 1 ,307a ,094 ,088 1,912
a. Predictors: (Constant), PRODUCT INVOLVEMENT
TABLE 2
Coefficientsa
Model Unstandardized Coefficients
Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) ,994 ,680 1,462 ,146 PRODUCT INVOLVEMENT ,471 ,118 ,307 3,978 ,000 a. Dependent Variable: PURCHASE DECISION INVOLVEMENT
TABLE 3
12 http://www.danielsoper.com/statcalc3/calc.aspx?id=31
Model Summary Mode l R R Square Adjusted R Square Std. Error of the Estimate 1 ,311a ,097 ,091 2,065 2 ,594b ,353 ,345 1,753
a. Predictors: (Constant), PRODUCT INVOLVEMENT
b. Predictors: (Constant), PRODUCT INVOLVEMENT, PURCHASE DECISION INVOLVEMENT
TABLE 4
Coefficientsa
Model Unstandardized Coefficients
Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 1,548 ,734 2,109 ,037 PI ,516 ,128 ,311 4,037 ,000 2 (Constant) ,976 ,628 1,556 ,122 PI ,245 ,114 ,148 2,149 ,033 PD ,576 ,074 ,532 7,740 ,000
a. Dependent Variable: PREF1 I prefer owning music (as CD, LP, mp3) rather than being able to only 'access' music (like with Youtube, Spotify or radio).
TABLE 4a
Coefficientsa
Model Unstandardized Coefficients
Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 2,521 ,643 3,924 ,000 PI ,251 ,112 ,180 2,247 ,026 2 (Constant) 2,363 ,635 3,724 ,000 PI ,157 ,116 ,113 1,354 ,178 PDI ,190 ,076 ,210 2,516 ,013
4.3.2 Regression 2
The second regression represents the core of the research aim: the effect of consumer involvement on preference for ownership. Variables are entered in three subsequent stages in order of importance as deducted from the literature. The two independent variables (IV's) are entered in the first model and the control variables in model 2 and 3 (situational factors and demographics respectively):
The model summary (table 5) show that model 3 (with the demographics) does not contribute significantly to the regression (p > .5).
In model 1 we see that 35,3% of the variance is explained by the IV's.
Model 2 accounts for 41,6% of the variance in the dependent variable (PREF1) with p<.05.
In the coefficients table (table 6) we see that aside from PI and PDI 'Mainly listening on pc/laptop' and 'Genre – Dance, Club/House and Techno' are significant, but negatively correlated with PREF1.
Product involvement
The stand-alone contribution of PI is significant and positive but 'loses' a bit of it's predictive power because the relationship with PREF1 is partially mediated by PDI, as was found in the first regression. In model 2 the inclusion of the situational factors strengthens it's power, as the b coefficient increases with 20% and p becomes two times smaller. More specifically, PI becomes stronger when we control for consumers who mostly listen to dance genres and consumers who mostly listen on their pc/laptop.
→ hypothesis 1 is supported
Purchase-decision involvement
This is evidently the main predictor of PREF1 and is responsible for the largest part of the variance in PREF1 (see regression 1). We do see that the b coefficient becomes slightly weaker in model 2 with the b coefficient decreasing with 10% (whereas PI increaased).
Genre and music device' effect
As can be seen two of the seven situational factors show a highly negative and significant effect. The effect of consumers who mostly listen to dance genres is
straightforward: these genres are mostly experienced in the club/disco without a direct relationship with purchasing records (Leyshon et al. 2005). For consumers who mostly listen on their pc/laptop, we could think of a number of reasons of why they exhibit a lower preference for ownership: music for them serves more as distraction while surfing; they care less for sound quality; they favor the ease and convenience (Burkart, 2008) of quickly having access without the hassle of purchasing; or on a more abstract level, they might have an entirely different customer value set when it comes to owning music or cultural goods in general.
While it's difficult to adequately say anything about these consumers'
involvement levels or the exact relationship with involvement, a possible explanation for the increase in PI and decrease in PDI is that these consumers may be product involved while having low purchase-decision involvement. So by controlling for them, we remove the suppressing effect on PI (hence the increase in power of PI).
TABLE 5 Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate Change Statistics R Square
Change F Change df1 df2 Sig. F Change 1 ,594a ,353 ,345 1,753 ,353 41,257 2 151 ,000
2 ,645b ,416 ,379 1,706 ,063 2,203 7 144 ,037
3 ,651c ,424 ,371 1,718 ,008 ,508 4 140 ,730
a. Predictors: (Constant), PURCHASE DECISION INVOLVEMENT, PRODUCT INVOLVEMENT
b. Predictors: (Constant), PURCHASE DECISION INVOLVEMENT, PRODUCT INVOLVEMENT, device - Mobile phone, Genre - Jazz and Classical, device - MP3 Player, Genre - Indie pop / rock, device - PC / Laptop, Genre - Dance, Club and Techno, Pop concert attendance
c. Predictors: (Constant), PURCHASE DECISION INVOLVEMENT, PRODUCT INVOLVEMENT, device - Mobile phone, Genre - Jazz and Classical, device - MP3 Player, Genre - Indie pop / rock, device - PC / Laptop, Genre - Dance, Club and Techno, Pop concert attendance, Income, Education, Gender_male, Age
