"Ease trumps effort" : a sociological research of how higher educated consumers choose music in a digital era

“Ease trumps effort”

A sociological research of how higher educated consumers choose

music in a digital era

Master thesis Master Sociology

Universiteit van Amsterdam Jannes Broekman, 10796592 jannes.broekman@student.uva.nl 1st reader: dhr. prof. dr. O.J.M.

(Olav) Velthuis

2nd reader: dhr. dr. A.T. (Alex) van Venrooij

Word count: 23,995 Amsterdam: July 2, 2019

Abstract

In my thesis, I ask the question: How do higher educated consumers choose music in a digital era? This question has gained new relevance in a digital era. Digitalization has made possible an overabundant and endless supply of music, and coincidingly a proliferation of sources to choose music. This question is concerned with how consumers choose music, instead of what music consumers choose. How consumers choose can be divided into a material and a normative dimension. The material dimension comprises the sources consumers use, which can be i.a. critics, the personal network, rankings and ratings but also music streaming services and algorithms. According to literature, these are judgment devices that serve to reduce the uncertainty which consumers face in a market of music. To research how consumers choose music, I have conducted 14 semi-structured or in-depth interviews combined with solicited diaries. 7 respondents have grown up in a digital era, and 7 have grown into a digital era. All respondents were higher educated. My research revealed that digitalization has led to an ease of choice, because one can easily delegate choices to Spotify and its personal recommendations. However, the personal network remains important and complements rather than substitutes Spotify’s personal recommendations. Moreover, my respondents use critics and rankings and ratings only as a preliminary assessment, because they are not concerned with the actual judgments they provide. Digitalization reduces the necessity to choose the ‘right’ product, causing consumers to use judgment devices not for reducing their uncertainty but rather only to navigate through the digital music market. Consumers’ normative dimension of how they choose consists of often delegating choice to sources, while retaining their emphasis on autonomy in judgment. Thus, central to this thesis is the preference of ease over effort in a digital era.

Table of contents

1 |

Introduction ... 4

2 |

Theoretical framework ... 7

2.1 | Culture of choice ... 7

2.2 | Sources for choice ... 8

2.2.1 | Judgment devices ___________________________________________________________________ 8 2.2.2 | Personal network _________________________________________________________________ 10 2.2.3 | Critics and reviews ________________________________________________________________ 10 2.3 | Rankings & ratings ...12

2.4 | The digital era ...15

2.4.1 | Music streaming services & recommendation systems __________________________________________ 16 2.5 | Ethics of choice ...19

3 |

Methodology ... 22

3.1 | Strategy & design ... 22

3.2 | Methods ... 23 3.3 | Research population ... 25 3.4 | Data analysis ... 26 3.5 | Operationalization ... 27

4 |

Everything Now! ... 28

4.2.1 | Shared development ________________________________________________________________ 37 4.2.2 | Relevance of personal recommendations ___________________________________________________ 40 4.3 | Critics & commensuration ... 43

4.3.2 | Contradiction in judgment ____________________________________________________________ 47 4.3.3 | Expert rankings and ratings __________________________________________________________ 48 4.3.4 | Buyers’ rankings and ratings __________________________________________________________ 50

4.4 | Autonomy ...51

4.4.1 | Autonomy in choice ________________________________________________________________ 52 4.4.2 | Autonomy in judgment ______________________________________________________________ 53 4.5 | Conclusion ... 55

5 |

Conclusion & discussion ... 57

6 |

Bibliography ... 61

7 |

Appendices ... 66

7.1 | Respondent characteristics... 66

7.2 | Interview questions (in Dutch) ... 67

1 | Introduction

“Every song that I’ve ever heard / Is playing at the same time, it’s absurd”, sings the popular Canadian indie rock band Arcade Fire in the song Everything Now (Metrolyrics.com, n.d.). This line is a stab at the modern on-demand streaming services that make access to almost every song immediately available. Arcade Fire’s latest album Everything Now is a pessimistic, yet bombastically festive, satire on modern society. Their main, explicit message repeated throughout the album is that we, in a digital era, have everything now, which barely leaves space for nothing now. Arcade Fire even went beyond lyrical satire in their social critique. They published a fake news website on which they posted a “premature premature evaluation” of their yet-to-be released album (Young, 2017). By this pre-emptive critical review of their album they made fun of professional music critics, speculating on critics’ usual speculation of albums that have yet to come out. This was a critique on the modern immediacy and absurdity of the valuation process of music, where people are so eager to judge music even before it has been released. Ironically, after release the album itself was deemed only mediocre by many music critics, receiving a mean grade of 6.6 (out of 10) (Metacritic.com, n.d.). However, the album did gain a large popularity, debuting at No. 1 on the Billboard chart (Caulfield, 2017).

Although it is debatable whether this is just a publicity stunt (Empire, 2017), Arcade Fire nonetheless introduces us to several interesting concepts in sociology that have explicit societal relevance in an age of digitalization. Firstly, they describe a part of the process by which music is given value and receives judgment. Critics are recognized as authorities of music valuation. But they are not the only ones that are important in this regard, because even though critics did not grade the album very well, it still became popular. There are other sources that guide consumers in their decisions in choosing music. For example, as Caulfield (2017) shows on Billboard.com, a highly ranked place in charts influences consumer choices as well.

Importantly, Arcade Fire pessimistically show that music judgments and consequent consumer behavior should be placed in the context of digitalization. Everything Now urges us to see that music is everywhere. However, this leaves consumers with a paradoxical situation (Luck, 2016). There is such an abundance of music on streaming services that they need to find ways to navigate through this. They need to use sources that inform them on the content and quality of music in order to consume it. Arcade Fire’s stunt has already enlightened some of these sources, namely critic reviews, and rankings and ratings, but not all. Digitalization has made possible a proliferation of these sources to choose music, for example music streaming services and algorithms (Schwarz, 2018). It has changed the way in which music is composed, distributed, and consumed (Théberge,

2015). In a digital era, choosing music has gained new relevance for sociological study. Thus, Arcade Fire’s latest album and their satirical stunt have raised the following research question for me: How do higher educated consumers choose music in a digital era?

To formulate an answer to this question, I will conduct semi-structured interviews and solicited diaries with consumers of two generations – one younger part with respondents aged 19 to 26 and one older part aged 53 to 61. This division is relevant, because it separates consumers that have grown up in a digital era with consumers that have grown into a digital era. In this way, what digitalization adds to and how it changes the way consumers choose music can be thoroughly investigated. Furthermore, the respondents need to be consumers with higher education, because, firstly, it is convenient for me to approach them and, secondly, this choice demarcates the respondent population.

To come to a full understanding of the sources consumers currently use in a digital era, this research will be set up inductively and exploratively by leaving my research questions relatively open. This gives respondents the opportunity to tell freely what sources they use and how they regard them, which guides my subsequent analysis. I will show – from the consumer’s point of view – how sources operate in a digitalized world, which is permeated by algorithms and rankings and ratings.

In order to accomplish this, Schwarz (2018) advises cultural sociologists to investigate the

how question of culture rather than the what question. This how question of culture has gained

momentum in the social scientific debate. He argues that culture shapes the specific processes and practices of consumers for choosing. The how question includes both a material dimension of culture – the sources that are available to consumers to help them make choices – as well as a normative dimension – why consumers use these sources. In this thesis I will relate both dimensions to each other in order to get a complete understanding of how people choose music in a digital era. This connects the culture of choosing with the cultural market of music. Furthermore, according to Schwarz (2018), digitalization shapes choice processes in many ways, which provides the how question with another dimension. For example, the debate on rankings and ratings has renewed relevance in an era of digitalization. I will add to this debate by setting rankings and ratings as a central theme in this thesis. Espeland and Stevens (2008) argue that numbers and quantification have been largely overlooked in sociology. It should, however, not be underestimated how rankings and ratings can exert a great authority over the behavior of people (Espeland & Stevens, 2008). By answering this research question, I will add to these sociological debates.

This research also has societal relevance. As stated by Arcade Fire’s promotion of their album

in contemporary society. Arcade Fire’s concern is that the immediate availability of music leads consumers to prefer quick, superficial satisfaction over the appreciation of the profundity of music. At the core of this thesis lies the preference of ease over effort in music consumption, in which there is a tension between ease and autonomy. This could be extended to other markets of cultural consumption as well, such as movies. In fact, the growing importance of quick and easy consumption, is visible throughout society in general, especially in fields where algorithms gain importance. My research provides a valuable insight into how this development changes the nature of consuming and choosing in society. So, this research will be able to provide – from a consumer’s perspective – insight into larger societal developments, like digitalization and rankings and ratings, and their consequences.

2 | Theoretical framework

2.1 | Culture of choice

Today, making choices is ubiquitous in our everyday life. There are concrete moments in which we are faced with the necessity of making direct decisions, but there are also vague moments in which we have to deliberately choose our paths of life (Schwarz, 2018). Choices are directly noticeable in consumerism. We make choices when choosing which clothes to buy, which movie to see, or which music to listen to, etc. “In consumer culture, choices are understood simultaneously as authentic manifestations or realizations of already-existing selves, and as the entrepreneurial project of self-making” (Schwarz, 2018, p. 857). Thus, in the context of choosing music, central to this thesis is the concept of choice, and its relation to culture. It is necessary to delineate a theoretical foundation upon which I can interpret my interview results.

In sociology of choice, there have been two main streams (Schwarz, 2018). Firstly, the modernity path argues that culture preceded choice in modernization. Culture is being equated with tradition, and the erosion of tradition means the opportunity for individualized and free choice (Giddens, 1991). Secondly, the dispositional path considers free choice an illusion, because choice has already been determined by culture in the form of internalized dispositions (or, habitus) (Bourdieu, 1984). However, Schwarz (2018) argues that the problem of these theories is that the first renders culture disappeared, whereas the second renders choice disappeared.

Therefore, Schwarz (2018) wants to emphasize how people choose, in order to keep choice and culture connected. Culture offers people “repertoires of ways of choosing, culturally specific

techniques of choice” (Schwarz, 2018, p. 851). Firstly, these techniques of choice are shaped by their

materiality, which constructs the alternative options or techniques to choose from. There are multiple sources that help consumers to choose from divergent music products. These sources are material settings or tools for consumers (Schwarz, 2018). They occupy the position between producers and consumers of music, which is essential to a capitalist society. For example, judgment devices provide such options, to which I will return in the following section. And especially digitalization plays an important role in shaping the sources that are available to people, for example through algorithms (Schwarz, 2018).

Secondly, culture comes with normative prescriptions – or, ethics of choice – that delineate which techniques (or sources) are appropriate to use for choosing music. These can be seen as underlying narratives that determine how people choose. Moreover, these normative prescriptions indicate modes or styles of consumption through which music is consumed and distinction is

played out (Schwarz, 2018). After a discussion of the sources consumers use, I will discuss the normative prescriptions of choice.

In this thesis I focus primarily on the material dimension, while also investigating the normative dimension of consumer choices in music. I will use Schwarz’s (2018) theory as a theoretical foundation in this thesis, because it redirects sociology of choice from questions about

what music is consumed to questions about how music is consumed. This view connects culturally

specific choice processes with the necessity of consumers to choose from a music market. I will return to this later.

2.2 | Sources for choice

In my thesis I want to examine in what way consumers choose music in a digital era. As Schwarz (2018) argues, consumers may use different techniques for choosing music, which comprise a specific materiality. From sociological literature it follows that this material dimension can be understood as sources that help consumers in guiding their choices.

2.2.1 | Judgment devices

As Schwarz (2018) argues, there is a material dimension to how people choose music. Karpik (2010) provides a theory of how this material dimension can best be understood. According to him, judgment devices guide consumers in making choices in markets of singularities, in this case music (Karpik, 2010). He argues that judgment devices provide ways to deal with the uncertainty that consumers experience when choosing music. Thus, judgment devices are important from the perspective of consumers, in the sense that consumers rely on the authority of judgment devices to choose music.

Judgment devices are especially relevant in a market of singularities (Karpik, 2010). Singularities refer to “at the very least artworks, products of the cultural industry, and professional services” (Karpik, 2010, p. 10). Firstly, singularities are multidimensional, which means that they cannot be reduced to a single quality or characteristic. All the dimensions of the singularity are related to each other. Secondly, singularities are characterized by uncertainty. They are produced through a certain interpretation of the dimensions but also consumed through a certain interpretation. This gives a singularity a strategic uncertainty in the sense that the uncertainty defines the product and its consumption. But singularities also have a quality uncertainty, which means that it is unclear what the quality of the product is. People have to form a particular judgment

rely on authorities to guide their judgment. Thirdly, a singularity is incommensurable. As said, it comprises multiple dimensions, whose qualities need to be judged through different interpretations. It is, thus, singular, unique, and not objectively comparable to other singularities (Karpik, 2010).

Music is a perfect example of a singularity. Music products entail multiple dimensions and are therefore uncertain in quality and not measurable. Because of music’s multidimensionality, quality uncertainty, and incommensurability, people need help to decide what music they want to consume. This help comes from personal and impersonal judgment devices (Karpik, 2010, p. 14). “Judgment devices reduce the cognitive deficit that characterizes consumers in the market of singularities” (Karpik, 2010, p. 44). They provide consumers with oriented knowledge and expertise, so that consumers are able to grapple with the incommensurability of singularities. In this way, judgment devices construct the exchange relationship between producers and consumers. Consumers put their trust into these devices, and thereby voluntarily delegate this social relation. On the basis of this delegation, judgment devices receive their authority. Each judgment device operates from its own point of view, and this can lead to different judgments of singularities (Karpik, 2010).

Karpik (2010) distinguishes between five categories of judgment devices. Firstly, appellations “are names associated with the attributes and meanings that define singular products or families of singular products” (Karpik, 2010, p. 45). In the case of music, these can for example be quality labels or genre labels, etc. Secondly, confluences are the techniques and marketing strategies used to attract buyers, such as public advertisements or the spatial organization of shops (Karpik, 2010).

Thirdly, Karpik (2010, p. 46) describes rankings as “hierarchical arrangements of singularities according to one or several criteria”. Karpik’s influential book was originally written in the 1990s, which means that the impact of the Internet was yet to be assessed. Nowadays, these rankings are even more pervasive in music consumption than they were back then (Beer, 2016). As a consequence of Internet, rankings and ratings have increased, as well as the forms they take on, organizing an increasingly abundant cultural field of music (Wright, 2011). Rankings and ratings, therefore, deserve a larger and more thorough discussion as a “sociological phenomenon in its own right” (Espeland & Stevens, 2008, p. 402). According to Espeland and Stevens (2008), it is especially important to investigate their social implications.

For this reason, I will turn to a discussion on the possible impact of rankings and ratings on the consumer choices in music in the next section. After rankings and ratings, I will discuss another societal development that has social implications in the way consumers choose music, namely music streaming services and recommendation systems. But first, it is important to discuss Karpik’s

(2010) two remaining judgment devices in order to come to an understanding of the sources that may guide consumers’ choices, namely networks and critics. These deserve separate subsections, because they are often and thoroughly discussed in sociological literature concerned with music consumption.

2.2.2 | Personal network

Networks are invisible and spontaneous but nonetheless powerful in influencing consumer choices.

There are personal networks – which consist of peers, family and contacts, and “operates by circulation of the spoken word” (Karpik, 2010, p. 45); trade networks – which are made up of sellers and buyers; and practitioner networks – which circulate knowledge of organizations. The personal network is based on the “distribution of credible product knowledge” (Karpik, 2010, p. 102). In this sense, consumers trust people from their personal network for multiple reasons. They believe that the others have a shared taste, personalized knowledge of others’ taste, interpersonal trust, and that they are looking for excellence in music (Karpik, 2010). According to Karpik (2010), accepting recommendations is based on the conviction that they possess these criteria. The trade network includes a professional authority that “produces the trust necessary to exchange” (Karpik, 2010, p. 102). This means that they prove the control for consumers based on a logic of excellence and trust. For example, these can be sellers from a local music store. The practitioner network has a logic based on moral cooperation between firms and practitioners. Karpik (2010) is less clear about how the practitioner network can function as a judgment device.

More can be said about the operation of the personal network as a source to guide consumers’ music choices. Tepper and Hargittai (2009) argue that opinion leaders are central to the discovery of new music for consumers. Opinion leaders listen to more hours of music as well as more variety in music. Furthermore, they have denser social networks with which they often share music. Because of their involvement with and expertise of music, they are confident to recommend to others. In this way, opinion leaders function as those members of a social network that often recommend music to others on the basis of their involvement with and expertise in music (Tepper & Hargittai, 2009). From a consumer’s point of view, it is necessary to see how consumers use recommendations by opinion leaders to guide them in their music choice.

2.2.3 | Critics and reviews

coined by Bourdieu. According to Bourdieu (1984, p. 325), these are typically “the producers of cultural programmes on TV and radio or the critics of ‘quality’ newspapers and magazines and all the writer-journalists and journalist-writers”. To organize the vast collection of theory on cultural intermediaries, Smith Maguire and Matthews (2014) have made a working definition. According to them, cultural intermediaries are, firstly, market actors who assign value and meaning to cultural goods (such as music), thereby mediating how these goods are perceived and engaged with by consumers, and, secondly, cultural intermediaries have to have a claim to professional expertise in taste and value (Smith Maguire & Matthews, 2012; 2014). On the basis of this professional expertise, they derive their professional credibility. Bourdieu (1984) would say that this professional credibility indicates cultural intermediaries’ symbolic capital, because they have the power to consecrate and legitimize certain knowledge as valued, thereby exerting a recognized authority. This recognized authority on the subject of music determines the trust consumers have in cultural intermediaries. This conviction of critics’ professional credibility leads consumers to take their valuation of music seriously (Smith Maguire & Matthews, 2014). Through this cultural and symbolic capital critics are able to relate to and identify with consumers that they speak to, and consequently influence their consumption behavior.

Critics’ professional expertise is conveyed to consumers through three central elements in reviews: the provision of “descriptive, informative elements that are familiar or easily accessible to the public”; the provision of “analytical elements that supply an interpretative context in which to understand the work”; and the provision of “evaluative elements in the form of positive or negative judgments” (Debenedetti, 2006, p. 31; Shrum, 1991). The impact of a review, then, depends on the professional expertise of valuation, and the critic’s social and cultural proximity to consumers. In this sense, when critics are in the same social space as consumers, and therefore have a similar taste, consumers tend to trust the reviews as a source. Moreover, consumers who are more involved with music, or more expert, are more inclined to read reviews (Debenedetti, 2006). This signifies why certain consumers use reviews as a source for choosing music, while others do not.

By valuating music critics take on a mediating role between consumers and producers in the field of cultural production that Bourdieu describes. This field consists of the sub-field of small-scale production – which is oriented toward artistic products – and the sub-field of large-small-scale production – which is oriented towards commercial products (Hesmondhalgh, 2006). Bourdieu’s “classical model” contrasts the aesthetic criteria of evaluation with the commercial (Beljean, Chong & Lamont, 2015). The critic legitimizes the autonomous logic of the small-scale field by evaluating music on the basis of a “certain aesthetic sophistication and its creator being inspired by a sort of

inner truth rather than by practical considerations – ‘art for art’s sake’” (Debenedetti, 2006, p. 38). In this way, critics determine and reassess what is good music through an aesthetic logic.

So, critics are part of the relationship between culture and economy. They assign value to cultural products, framing them as legitimate and valuable – what Bourdieu (1984) calls ‘symbolic imposition’. Critics, thus, accomplish the ‘qualification of products’ (Smith Maguire & Matthews, 2010). According to Karpik (2010), this qualification of products is influential in consumers’ music choice. Karpik (2010, p. 97) argues that judgment devices are located between producer and consumer, thereby acting as the main qualifiers of music. This qualifying of music is of help to consumers for choosing music. According to Debenedetti (2006, p. 35), “for the consumer, reviews can reduce the fundamental uncertainty associated with an experiential product, the quality of which cannot be evaluated fully before purchase”. Critics give consumers the specific knowledge of the quality of music by their valuation of music, in a market that is defined by its uncertainty (Karpik, 2010).

To conclude, as we have seen, judgment devices – including networks and critics – can be important for influencing consumer behavior. By means of assigning music a judgment, critics and networks reduce the uncertainty of the quality of music that has yet to be chosen. Their valuation indicates what is ‘good’ or ‘bad’ music, or what fits consumers’ tastes. However, as I have said, it is possible that the above named sources are not the only ones influential in consumer choices. In order to fully answer my research question, therefore, it is first important to explore what kind of sources consumers turn to when they choose music. Moreover, it is necessary to research how consumers use these sources and how they regard them. Therefore, my first research subquestion is as follows: What kind of sources do consumers use to choose music?

2.3 | Rankings & ratings

As I have said, an important judgment device Karpik (2010) denotes are rankings. Here, I will provide this judgment device with a more thorough theoretical discussion. Espeland and Stevens (1998) describe how the process of commensuration is central to this judgment device. This applies not only to rankings, but also to ratings. Commensuration means “the transformation of different qualities into a common metric” (Espeland & Stevens, 1998, p. 314). Commensuration has changed the way we value. Today, it seems self-evident that the value of cultural products is standardized, quantified, and objectified into a number. Capitalism is pierced by commensuration and completely depends on it. Capitalism makes it necessary for cultural products to be assigned a price, so that they can be objectively measured, understood and consumed. Different qualities are transformed

them into a single metric. Through rankings and ratings this metric can be compared between different cultural products, from which an objective value can be deduced (Espeland & Stevens, 1998).

Let’s take music. As I have explained before, music is a singularity on the cultural market. It entails different qualities that are hard to define, and is characterized by uncertainty. This makes it for consumers unclear as to what the value of a particular music product is. It can be seen as an incommensurable, as Espeland and Stevens (1998) describe. Incommensurables are socially unique, not comparable to others. However, as Karpik (2010) shows, even incommensurables (or singularities) are being commensurated as much as possible. Commensuration, then, can come in the form of a ranking or a rating, in which a music product is defined into one common metric, such as a place in a Top 40 list or a grade of 4 stars out of 5. This commensuration is helpful for informing consumers about the supposed value of music. In other words, it helps to reduce their cognitive deficit (Karpik, 2010). Thus, rankings and ratings are important in assigning value to music, albeit a commensurated, standardized value. They stand between producer and consumer, guiding consumers which choices to make.

This process of valuation faces some consumers (either wittingly or unwittingly) with a dilemma. Fourcade’s (2016) theory of classificatory judgments explains this dilemma. How we judge things – like music – goes according to three basic principles: nominal, cardinal, and ordinal judgments. “Nominal judgments are oriented to essence”, which means looking at the inherent properties of music (Fourcade, 2016, p. 176). It is, in this sense, a qualitative judgment. Cardinal judgments allow for comparisons on the basis of the numerical value of the properties of music. It is about the quantitative dimension. But most importantly, ordinal judgments are evaluative, and are “oriented to relative positions according to a stable rank-ordering criterion” (Fourcade, 2016, p. 178). As Fourcade (2016, p. 178) argues “modern ordinal judgments often tend toward numerical commensuration”. This commensuration means that the value-making shifts from a nominal judgment, based on the pricelessness of inherent qualities, to an ordinal judgment, based on the commodification of a common metric (Fourcade, 2016).

Karpik (2010) provides a useful framework for interpreting these rankings and ratings, and judgment devices in general. According to him, all markets of singularities are regulated by a particular ‘regime of economic coordination’ – a model that has its own “system of relations between particular qualified products, particular judgment devices, and particular forms of consumer commitment” (Karpik, 2010, p. 96). Judgment devices can be divided into impersonal (appellations, cicerones, confluences, and rankings) and personal judgment devices (networks). I have already explained these networks in more detail in the previous section.

Impersonal judgment devices can be divided into substantial and formal devices. Substantial devices are concerned with the specific content of a singularity. A music product is evaluated according to its particular characteristics, on the basis of aesthetic criteria (Karpik, 2010, p. 100). This is generally done by critics. We can distinguish between two regimes of economic coordination here. The authenticity regime is primarily located on small markets, in which originality is important and music products receive moderate profit. The mega regime is located on large markets, where devices (such as critics) have to balance between aesthetic criteria and high profit (Karpik, 2010, p. 101). On the contrary, formal devices are concerned with “the relative positions of the singularities”, and these qualify singularities on the basis of rankings and ratings (Karpik, 2010, p. 100). The expert-opinion regime typifies the rankings by experts on a small market. They select the best music products relative to other music products, and consumers trust in them do to so. Expert rankings and ratings are official awards and prizes handed out by professionals in the music industry. The common-opinion regime, however, operates on a large market and is primarily oriented toward high profit. Therefore, these rankings are primarily buyers’ rankings, which are determined by the selling figures of music products, such as best-seller lists (Karpik, 2010).

Furthermore, there are forms in which multiple judgment devices are combined. Professional critic reviews, for example, often combine a detailed discussion of the inherent qualities of a music product with a rating. They combine a subjective valuation with an objective, commensurated measure in order to be able to compare them with others. And on the Internet consumer reviews are often combined with a ranking (Mellet et al., 2014). In these cases, substantial and formal devices are combined.

In this thesis, therefore, my broad understanding of rankings and ratings will be that they designate a common metric to a product of music – either in the form of its relative place to other products of music (ranking), or in the form of an objective measure of its supposed quality (ratings), or both. I want to let the interview respondents freely indicate which types of rankings and ratings they deem important (or not) for their consumer choices.

Bialecki, O’Leary and Smith (2016) researched the importance of these kinds of ratings and rankings for film viewer choices. They concluded that film viewers often use rankings and ratings – of the movie database IMDB – in their consumer choices and are highly influenced by them in their opinion of cultural products. Consumers primarily use these rankings and ratings as a screening

tool in order to assess the quality uncertainty of movies before watching. Moreover, consumers are

more inclined to use rankings and ratings as a screening tool from particular reviewers of which they know possess a similar taste in movies (Bialecki, O’Leary & Smith, 2016). Because of the similarity of their research to this aspect of my research, I base my second research subquestion on

theirs. The second subquestion of this thesis is as following: How are rankings and ratings implicated in

consumers’ choice of music?

I will now turn to a societal development that has reorganized the way in which consumers choose music, namely digitalization. In particular, I am interested in how consumers use music streaming services and the algorithms they offer.

2.4 | The digital era

We are now living in a digital era, in which music is largely composed, performed, recorded, distributed, and consumed digitally (Théberge, 2015). Digitalization has not been a sudden change, but rather a “relatively long, transformative process of economic, technological, social and cultural change that has taken place over a half-century or more” (Théberge, 2015, p. 329). Due to the rise of technology in the music industry, it was possible for artists to find newer and creative ways to make music. Not only production developed, also the digital distribution and consumption rose to a great extent. Where people previously had to buy CDs or LPs, it is now possible to listen straight from an Internet-connected mobile phone (Hesmondhalgh & Meier, 2018). The easy accessibility of MP3 files on the Internet and the availability of music streaming services have presented consumers with consumption values based on “speed of access, minimized storage requirements and convenience” (Théberge, 2015, p. 334). Immediate and unlimited access to music has redefined the choices consumers have to make regarding the cultural product of music (Théberge, 2015).

Digitalization has given people access to an increasing volume and diversity of music (Wright, 2011). Coincidingly, this means a proliferation of types of sources to inform music choice in a digitalized world. For instance, rankings and ratings have become more pervasive in more different forms (Beer, 2016). Another example is that electronic word-of-mouth communication has become an important alternative to traditional word-of-mouth communication for consumers. In eWOM communication, consumers share their opinions and experiences of cultural products, such as music, on Internet forums (Hennig-Thurau et al., 2004). Furthermore, Hanrahan (2013) shows how professional music criticism has largely made way for new online forms of evaluation. For example, sites as Pitchfork, independent bloggers, and social media personalities have become more and more important in guiding consumers what kind of music they should choose.

Especially young people, who are born in a digital era, are always connected to digital forms of media – listening to music on their iPods or mobile phones, or finding, sharing, and downloading music via the Internet (Swerdlow, 2008; Palfrey & Gasser, 2008; Hargittai & Hinnant, 2008). However, Tepper and Hargittai (2009) show that, for young people, alongside these digital sources,

as well. According to them, young people “will continue to rely on traditional mass media (radio, television, newspapers, films) as important sources for the discovery of new music” (Tepper & Hargittai, 2009, p. 232). Reversely, it needs to be researched whether older people, who have grown into a digital era, are starting to rely on digital sources in their music choice.

So, digitalization makes possible new techniques of music consumption. According to Schwarz (2018, pp. 853-854), in order to choose what music they want to listen to, consumers delegate their “cognitive, evaluative, calculative screening and choice processes to non-humans – online forms and algorithms that organize information about choice options; offer evaluative criteria to formalize users’ preferences; and rank and rate options”. I will now turn to a discussion of these new digitalized techniques, namely music streaming services and algorithms.

2.4.1 | Music streaming services & recommendation systems

In recent years, lakes of ink have been spilled over the concept of algorithms. To keep this discussion compact, I will only discuss algorithms in relation to streaming music services and music consumption, of which the personalized recommendation system is the most notable algorithm. Moreover, a discussion of algorithms easily slips into a technical debate, so I will only refer to algorithms in a general sense, highlighting the way in which they work to influence consumer behavior.

According to the Global Music Report 2018 (IFPI, 2018), 54 percent of the global music revenue is accounted for by digital music streaming in 2018, which makes it the most important source for listening to music. In the Netherlands, Spotify is used by 1 out of 5 people, which amounts to a total of 3.3 million consumers. Apple Music is the second best used streaming service in the world, amounting to 2.2 billion consumers (BTG, 2017). More than half of the music is being consumed through these streaming services. Even more so, in 2019, iTunes will be discontinued by Apple, which signifies the transition from digital ownership of music (on iTunes) to music streaming services as the main source for consumption (Roettgers, 2019).

Sociologists are ambivalent about this development. Some say it will lead to more “control and supervision, coupled with over-abundance, inflation of artistic quality, meaninglessness, repetition and disappointment” (Andersson, 2010, p. 69). Others say it will liberate digital culture, where everyone can access whatever one wants (Andersson, 2010). Music streaming services give consumers new ways to explore, collect, organize, and choose music. Discovering new music is an important consideration for joining paid subscriptions to streaming services. These streaming services make music ubiquitous for consumers, and as a consequence lead them into the possibility

music creates a ‘paradox of choice’, according to Luck (2016). There is so much choice that consumers do not know what to choose anymore. They have to find ways to navigate through this supply, otherwise most people start listening to music they are already familiar with (Luck, 2016).

Marshall (2015) discusses three types of music streaming services, which all work on the basis of complex algorithms. Firstly, streaming radio functions often as a personalized radio, which automatically picks and plays songs based on one’s previous listening choices, like Pandora. This is a non-interactive mode, which means that the consumer does not directly choose the songs. Secondly, locker services are a small part of the streaming industry, and function as a locker for one’s own music files, which then offers mobile streaming access to that locker. Examples are iTunes match and iCloud music library. Thirdly and most importantly, on-demand streaming services affords the consumer the option to choose which music is played. This is an interactive form in which the consumer can choose music from a large database of songs available. The most obvious example is Spotify. Spotify, however, also incorporates both the non-interactive streaming radio form and the locker form (Marshall, 2015).

Music streaming services, especially Spotify, make use of a complex algorithm whereby they can automatically recommend music. An algorithm can be understood as a “complex assemblage of information exchange consisting of multiple elements and processes” (Karakayali, Kostem & Galip, 2017, p. 5). On the basis of multiple aural factors, like danceability and pitch, The Echo Nest (the algorithm technology corporation behind Spotify) measures songs one has listened to before. The Echo Nest also scans the world context for information about artists, genres, etc. On the basis of this information, Spotify is able to create a taste profile of a user, and consequently recommend undiscovered songs one might like. In this way, Spotify maps the consumer as a musical identity (Prey, 2017). An important aspect of Spotify is its ‘Discover Weekly’. This is “a personally tailored playlist of 30 new tracks that is delivered to each subscriber every Monday morning” (Prey, 2017, p. 1090). Another example is Spotify’s ‘Daily Mixes’, which are playlists assembled by an algorithm in order to recommend music while distinguishing between certain genres, moods, or styles. This is meant by Spotify as a playlist fitted to one’s daily needs, and therefore does not discriminate between songs one has or has not heard before, as ‘Discover Weekly’ does. Generally, on the basis of these algorithms Spotify is able to suggest personal recommendations of songs, albums, artists, and playlists (created both by Spotify and users). It could be that these recommendations are subject to commensuration as well, in the sense that songs are objectively measured and placed in a ranking of the degree to which songs fit one’s taste. However, I have not been able to find any literature that provides more transparency of these intricate workings of algorithms of Spotify.

Karakayali, Kostem and Galip (2017) argue that such a recommender system mediates between users and their cultural environment. By doing so, streaming music services reify both the music and the individual consumer. This can make consumers uncomfortable about personal recommendations (Prey, 2017). Prey (2017) regards this as a process of algorithmic individuation, in which streaming music services help creating the individual through personalized recommendations. In this way, it can function as a technology of the self whereby the consumer can work on a specific kind of taste through algorithms.

Thus, this recommender system can function as an intimate expert, which exerts a certain authority to help consumers choose music (Karakayali, Kostem & Galip, 2017). This can be tied back to the discussion on cultural intermediaries. Morris (2015) explains how these recommendation systems reconfigure the process of cultural intermediation. He calls companies, like Spotify, that centralize these algorithms infomediaries: “organizations that monitor, mine and mediate the use of digital cultural products” (Morris, 2015, p. 447). According to him, recommendation systems are involved in the process of value-making and representation of music. Through an interaction of consumer and code the value of music is framed in music streaming services. These infomediaries influence consumer behavior, because “they expand the capabilities of digital commodities by helping users discover new cultural goods or providing novel experiences of familiar cultural content, and they do so largely through persistent and pervasive data collection” (Morris, 2015, p. 453). Traditional cultural intermediaries have authority based on their expertise and knowledge of what is good music (editorial logic), whereas infomediaries have authority based on the effectiveness of their taste profiles of consumers (algorithmic logic) (Morris, 2015; Gillespie, 2014). According to Gillespie (2014), this editorial logic might come to be replaced by an algorithmic logic.

I have argued that digitalization and especially its products of music streaming services combined with recommendation systems provide a new way of helping consumers to choose music. Digitalization can be seen as setting a new context for techniques of choice, such as rankings and ratings, but also algorithms. The recommendation system is a source that guides consumers’ music choice, because they exert a kind of authority in the valuation and representation of music. In this thesis I am interested in how consumers use these digital techniques of choice and how they regard them. Therefore, my third subquestion is as follows: How are music streaming services and

2.5 | Ethics of choice

In the previous, I have tried to explain how sources – such as critics and networks, but especially rankings and ratings and music streaming services – help consumers make choices in music consumption. Next to this materiality of techniques of choosing, the normativity is an important aspect to consider, which Schwarz (2018) calls ethics of choice. “Culture equips us […] with normative prescriptions, regarding which choosing techniques are appropriate where and why” (Schwarz, 2018, p. 851). In this regard, the normative dimension reveals why consumers make use of certain sources to guide their choices of music; why they might prefer these sources over others; and why they might use them in these ways. This normative dimension explains how consumers use sources to choose music, while adding a cultural explanation to this behavior.

Karpik (2010) provides a theory of why consumers use certain sources to guide their music choice. He argues, “in the market of singularities, when one is confronted with the multiplicity of possible choices, the search for the ‘right’ product can be conducted” through certain consumer commitments (Karpik, 2010, p. 104). Through an active commitment, consumers want to autonomously make their own choice for the right product, whereas through a passive consent, consumers defer the choice to make the right product to others. And through autonomy, consumers want to develop their own personal judgments, whereas through heteronomy, consumers are inclined to adopt other judgments as their own. This leads to four types of consumer commitments: activity and autonomy; activity and heteronomy; passivity and autonomy; passivity and heteronomy (Karpik, 2010). These consumer commitments relate in several ways to judgment devices and the economic regimes that I have discussed earlier.

Firstly, consumers that have an active commitment to make personal choices in music and seek to assert their autonomy in doing so, adhere to the authenticity regime. They see choice as an adventure, and only use sources because they are necessary in an overcrowded music supply. They find authenticity and originality in small markets, and consequently listen to critics with a full aesthetic valuation of music. Secondly, consumers that have an active commitment to make personal choices, but are heteronomous in their judgment, adhere to the mega regime. They also value active personal choice, but rely more on sources in their judgment because the (financial) risk as a consumer is too great to be autonomous. Therefore, they refer to sources that balance aesthetic criteria with financial success of music (often critics), so they know their choice is worth the expense. Thirdly, consumers that passively consent to others to choose the right product, but want to be autonomous in their own judgment, belong to the expert-opinion regime. They primarily use expert rankings and ratings, because these tell them what music is ‘better’, while they can still assert their autonomous taste in small markets. They are able to take autonomous, critical positions

toward others’ opinions. Lastly, consumers that have a passive consent and are heteronomous belong to the common-opinion regime. The latter, however, favor the convenience of others’ opinions and are not inclined to autonomously maintain their own judgment. This leads to the use of buyers’ rankings and ratings, and subsequently choices that veer towards conformity (Karpik, 2010).

This begs the question what explains these different types of consumer commitments of choosing music. According to sociological literature, these ethics of choice are structured by a new turn of ‘emerging’ cultural capital. This focus on how consumers choose music leads away from a focus on what kind of music consumers choose. According to Jarness (2015, p. 67), “liking the same things does not necessarily mean having the same tastes”. A focus on what consumers choose may lead to the idea that taste directly predicts a certain choice of music (Jarness, 2015). Instead, a focus on how consumers choose music shows that a similar choice of music can be appropriated and appreciated in different ways (Daenekindt & Roose, 2017). Correspondingly, this “analytical decoupling of the whats and hows of cultural consumption” has led to a different view on cultural distinction (Jarness, 2015, p. 66). Cultural distinction happens not on the basis of specific choices in music (as a result of taste), but on the basis of how choices in music are made (Prieur & Savage, 2011). ‘Emerging’ cultural capital is a measure of the certain ways of engagement with – or appropriation of – music, instead of taste differences (Friedman et al., 2015; Prieur & Savage, 2013). Emerging cultural capital is located on the axis of the modality of consumption – the embodied performance of distinction through how consumers choose music (Friedman et al., 2015, p. 6). The amount of ‘emerging’ cultural capital can be a result of, for example, educational level, as a factor of social class (Jarness, 2015).

A particular contemporary ‘reflexive appropriation’ of culture lies at the basis of distinction through how music is consumed (Prieur & Savage, 2013; Bennet et al., 2009). Through being self-reflexive consumers know how to appropriate certain cultural products. In other words, culture provides the different ways of choosing music, and consumers know how to navigate between these ways. This requires a certain ‘distanced and ironic attitude’ through which consumers are able to choose between the vast supply of music (Prieur & Savage, 2013). This attitude even allows for a disinterest in the traditional distinction mechanisms of ‘highbrow’ and ‘lowbrow’ culture (Prieur & Savage, 2013). Young people tend to care less for this traditional aesthetic distinction. For example, in an ironic way, they happily consume music that is generally considered bad (Bennet et al., 2009).

These theories suggest the ethics of choice that underlie how consumers choose music. Consumer commitments can be seen as normative prescriptions – to be either autonomous,

heteronomous, active, or passive in one’s choices (Karpik, 2010; Schwarz, 2018). And they indicate the perspective through which certain sources are regarded. Also, these ethics of choice, and the corresponding use of certain sources, might form the basis of distinction. Jarness (2015) shows that how consumers choose music differs between educational level, as a factor of social class. My research, however, is limited to a discussion of consumers with higher education. Even though I cannot compare between educational levels, it is still necessary to investigate what the underlying ethics of choice for higher educated consumers are, and how these differ for them. It is the question whether these ethics of choice correspond to Karpik’s (2010) consumer commitments. In any case, it is necessary to not only ask how consumers choose music, but also why they do it in these ways. Therefore, my fourth subquestion is as follows: Why do consumers use these sources in these ways for

3 | Methodology

3.1 | Strategy & design

In this thesis, I aim to show how higher educated consumers choose music in a digital era. I have structured the research in an inductive and explorative manner, in order to get broad and thorough responses. I wanted to let consumers explain their wide variety of methods to choose music in a digital era, and how they regard these methods. Therefore, it is important to put the focus on the interpretation of the consumers themselves. This thesis is not concerned with the quantitative data that indicates what amount of consumers use certain sources, or the significant effect of one variable on another. Rather, it is concerned with how consumers try to navigate a digital era that is complex and varied, and what this means to their music consumption. This thesis digs in to the experiences and opinions of consumers of the process of choosing music; it is about subjectivity. This thesis, therefore, requires an open-ended and qualitative research strategy (Bryman, 2012).

Correspondingly, this thesis is based on an interpretivist epistemology and a constructionist ontology. I am interested in the “subjective meaning of social action” (Bryman, 2012, p. 30). This means that I aimed to discover the interpretation, view, meaning, and value, my respondents have of their processes of choosing music. In this thesis, I center social actors and their actions. Continually, by social processes such as choosing music, people recreate meanings they attach to the social world. In this thesis, the concept of culture must be understood as a process that is not pre-given or deterministic at all, but constantly construed by social actors (Bryman, 2012).

To support this epistemological and ontological perspective, I primarily used an inductive approach. Inductive means adhering to an iterative process: going back and forth between theory and data. From the data I collected I aimed to see what fits within the theoretical framework, but also how I can adjust that theoretical framework to support my data (Bryman, 2012). Due to this thesis’ explorative character, I believe it is essential to let the data guide the rest of the research. The data guided the linking between theoretical framework and results. For this purpose, I have set up my first subquestion (What kind of sources do consumers use to choose music?). The answers to this question given by my respondents have led me to revisit the theoretical framework. By collecting the data, I found the most important ways in which people choose music, which subsequently led me to place an emphasis on these ways in the theoretical framework. So, I have constantly revisited and realigned the theoretical framework with the empirical data and vice versa, in order to come to new theoretical insights. In this sense, my approach borders on an abductive approach, because I did not set aside all preconceived theoretical ideas during the research project, but I developed the theoretical repertoire throughout the research process (Timmermans & Tavory, 2012).

The research design I wielded is a cross-sectional design. It is often used when a researcher interviews a number of people according to in-depth, unstructured, or semi-structured interviews (Bryman, 2012). This design implied that I collected empirical data from several respondents, compared the interview data across the multiple respondents I have interviewed, and analyzed what the similarities and differences are. It is a useful method for qualitative research to interpret the subjective meaning respondents assign, and to compare them with each other (Bryman, 2012).

3.2 | Methods

My main method was qualitative interviews, for a number of reasons. Firstly, this is the method that fits best with an interpretivist epistemology and a constructivist ontology. Secondly, qualitative interviews allowed me to centralize the respondent’s point of view. I wanted the respondents to tell me what kind of sources they use, how they use them, and why they use them. This gave me rich and detailed reports of their processes of choosing music. Thirdly, this particular method allowed me to go about the research in an explorative manner. I did not want to restrict them by providing a list of sources or reasons for choosing music, thereby possibly risking missing relevant ones. Who can tell me better how they choose music than the consumers themselves? Quantitative surveys, instead, would have only offered me a fixed set of questions, without room for possible digressions or other relevant information. Fourthly, the topic of choosing music could be, for most people, implicit in their everyday life. In most cases they do it automatically, quick, without reflection, and they may use many sources. This means that consumers need some time to think about it. Conducting an interview with these consumers means giving them the time and helping them to think their processes through. Surveys, on the contrary, would rush them into answering. They would probably not take (enough) time to really think their processes through, which could lead to imperfect and incomplete answers.

Before collecting my regular data, I have conducted a test interview. In this test interview I went in-depth to see what kind of patterns, questions and themes are important and keep returning. From these, I was able to construct semi-structured questions. I followed up this test interview with the actual interviews. I allowed for “rambling off”, which often made the interviews more in-depth than semi-structured. We were able to go into a detailed ‘discussion’ or ‘story’ of how and why they choose music. However, sometimes my respondents were inclined to answer immediately without really thinking their answers through. At other times, I realized I made the mistake of following my questions up too quickly without giving respondents time to think, or of asking closed questions. But after a few interviews I felt more comfortable and realized that I simply had to give

questions. Another useful method was to provide a small answer or story of my own, so they could understand what I meant and consequently were able to think through and voice their own thoughts. Because of these reasons, the interview questions did not determine the actual order and content of the interviews. Most of the time, it was necessary to deviate from the interview question scheme, which was only useful as a topic list for guidance (appendix 7.2).

All interviews took place where the respondents felt most comfortable. Nearly all respondents wanted to interview face-to-face inside their homes. One respondent wanted to sit in a café, and another insisted on doing it over the phone, since he had little time. The face-to-face character of the interviews appeared to be very helpful for a comfortable setting. There was hardly any difficulty in creating a rapport with the respondents, because they were all very open and enthusiastic to talk about the subject of music.

There is, however, a shortcoming to the interview method. Interviews are very useful, but only to a certain extent. They were able to provide me an in-depth view of the practices of consumers, but I had to take into account the possibility that the respondents at the time of the interview were not able to be fully explicit about these practices, because they did not realize them at the moment. It is, therefore, important as a researcher to cope with the attitudinal fallacy: “what people say is often a poor predictor of what people do” (Jerolmack & Khan, 2014, p. 178). As Swidler (2001) has argued in Talk of Love, people often possess contradictory values, which do not directly correspond to their actions. People tend to pick a culturally determined way of acting for a specific situation, which might not fit with the values they adhere to. This shows the necessity of a triangulation method (Bryman, 2012). Jerolmack and Khan (2014) argue that in order to reduce the attitudinal fallacy it is necessary to complement interviews with observations. However, in my case, it was not possible to conduct small-scale observations in combination with the interviews, because these would not indicate properly consumers’ day to day practices in choosing music. The respondents would simply give an example of their practices of choosing music (of which they already explicitly know they use).

Therefore, I needed to complement interviews with a method that allowed for keeping track of consumers’ day-to-day practices. Solicited diaries have been proven to be useful in this regard, and they are able to provide insight into “social life ‘as it happens’” (Dowling, Lloyd & Suchet-Pearson, 2016, p. 683). The ‘diary-interview’ method is useful when “the investigator is unable to make firsthand observations or wishes to supplement those already collected” (Zimmerman & Wieder, 1977, p. 481). Therefore, firstly, I conducted an interview with respondents and, secondly, I asked them to keep a short diary. In the interview I could discuss with the respondents what they think their normal practices of choosing music are. In these interviews I needed to make the

respondents aware of their actions, making their implicit behavior explicit. Then, when they were asked to keep their diaries they were more aware of their actual practices and more clearly able to keep track of them. For example, the respondents might not have been aware before the interview that they use algorithms to choose new music. If they would have recorded their diaries beforehand, they would not have noticed this. In this sense, it was my purpose to let the respondents revisit the interviews in their diaries themselves. It is a ‘self-reflective’ method that helped me to define more specifically what consumers’ practices for choosing music are (Dowling, Lloyd & Suchet-Pearson, 2016).

I realized that these diaries were a time-consuming practice for many respondents, so the respondents were free to reject. Of all respondents only one respondent did not submit the diary to me. I asked the respondents to record a brief entry once every day for three days. This entry was primarily focused around the question of how they choose music in their day-to-day life. To make it easier for the respondents I provided them with a set-up diary schema (appendix 7.3). My aim was that the respondents would think more explicitly about their use of certain sources to choose music. In practice, this method seemed a useful triangulation method. It provided me with specific examples of how they use certain sources, which supplemented my data analysis. The respondents often indicated to use the same methods as they told of in the interviews. This leads to me to say that the attitudinal fallacy barely occurred during the interviews (Jerolmack & Khan, 2014).

3.3 | Research population

I have conducted 14 interviews (appendix 7.1). I kept the variance of possible respondents as large as possible, without setting too much characteristics that they needed to meet, apart from age. However, in the end this provided a shortcoming to the research. If I had set more strict criteria, I would have been better able to provide a specific answer to my last subquestion. It could have given me insight in differences in choosing music between different genders, educational levels, and classes, which would, for example, have showed the works of distinction in how people choose music. I did not set gender and occupation as selection criteria, but I did ask my respondents for these characteristics.

Nonetheless, there were several considerations I made in order to find suitable respondents for my research. Firstly, the respondents needed to consider themselves consumers of music. They, at the least, needed to listen to, stream, or buy, either digital or material music 2 or 3 times a week. A second important consideration was age. To investigate the effect of digitalization on how consumers choose music, I divided two age generations. The first generation needed to have grown

This resulted in a first group of respondents aged between 19 and 26 years old (7 respondents), and a second group of respondents aged between 53 and 61 years old (7 respondents). By dividing between these two age groups I was able to find interesting results that show how people have adapted to digitalization and how their practices differ from people that have grown up during digitalization. Thirdly, during the respondent selection process I became aware that I was only able to find higher educated consumers (appendix 7.1). These respondents were most convenient for me to approach for interviews due to my own level of education. Therefore, I decided to demarcate the respondent group to include only respondents with higher education. The consequence is that the following findings only apply to consumers with higher education. However, even though this selection inhibits the possibility to compare between educational levels, it proved useful to provide a more in-depth account of music choice of higher educated consumers. In this way, I could thoroughly investigate the differences and similarities within this respondent group.

With these considerations in mind I used purposive sampling (Bryman, 2012). Because of time and place constraints this sampling took place in my own locality. I started with my own personal network to find suitable respondents. This sampling method was convenience sampling, in which I simply tried to find respondents on the basis of accessibility (Bryman, 2012, p. 201). Furthermore, the initial respondents were able to propose other respondents to me that were suitable. In this way, I also gained access to others’ networks. So, I supplemented it with a snowball sampling method (Bryman, 2012). However, these methods entail the disadvantage that the selection of respondents is not random. This was of little consequence, because there was much variation in their answers, regardless of their personal connection.

3.4 | Data analysis

I aimed to transcribe each interview directly after I conducted it. However, in practice, this was difficult due to time constraints. During the period of data collection and analysis, I assembled notes about the process, the respondents, and the interviews. In these notes I already compared multiple interview findings. In a later stadium, these notes proved helpful for interpreting the data. Subsequently, after conducting the first 12 interviews, I started with the coding process with the software package Atlas.ti. Because of the explorative and inductive character of the research, I coded the interviews inductively, while keeping the research questions and interview topics in mind. The same goes for the solicited diaries. This resulted in 133 codes. Then, I organized these codes into 5 groups that resembled the research questions. The codes and code groups made it possible to compare the interviews while keeping an eye on my research questions, even though the

3.5 | Operationalization

My research question is: How do higher educated consumers choose music in a digital era? As I have explained, I aimed to keep this research as explorative and inductive as possible. My theoretical framework functioned more as a guideline for the interviews and as a framework for interpretation of the data. Therefore, I have not operationalized the concepts into clear dimensions and indicators.

However, there are some concepts I have operationalized more clearly than others. I have divided the concept ‘choice of music’ into ‘techniques of choice’ and ‘ethics of choice’. Techniques of choice indicate the sources people use in their choice, whereas ethics of choice indicate the underlying normative prescriptions (Schwarz, 2018). The sources consumers use in their music choice were operationalized according to Karpik’s (2010) judgment devices: networks, appellations, cicerones, confluences, and rankings. Karpik (2010) organizes these sources into regimes of economic coordination. I have highlighted some of these sources in the theoretical framework, even though the interviews made clear that many more are used.

The personal network is based on trust in personalized knowledge, interpersonal trust, and a shared taste (Karpik, 2010). Critics (cicerones) are used on the basis of trust in their expertise and their cultural proximity to consumers (Debenedetti, 2006). Furthermore, rankings and ratings are understood as the commensuration of music, and can be divided into expert – assessed by critics – and buyers’ rankings and ratings – e.g. best seller lists (Karpik, 2010). Moreover, music streaming services make immediate access to music possible (Marshall, 2015). They provide algorithms, which take on the role of recommendation systems, that can automatically recommend new and unheard music based on consumers’ earlier preferences (Morris, 2015; Prey, 2017). Finally, ethics of choice are operationalized by consumer commitments: activity and autonomy; activity and heteronomy; passivity and autonomy; passivity and heteronomy (Karpik, 2010).