Document [original]

The Behavioural Economics of Music

A framework for investigating music decision making

vorgelegt von

M. Sc.

Manuel Anglada-Tort

ORCID: 0000-0003-3421-9361

an der Fakultät I – Geistes- und Bildungswissenschaften

der Technischen Universität Berlin

zur Erlangung des akademischen Grades

Doktor der Philosophie

- Dr. phil. -

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender: Prof. Dr. Stefan Weinzierl

Gutachter: Prof. Dr. Jochen Steffens

Gutachter: Prof. Dr. Daniel Müllensiefen

Tag der wissenschaftlichen Aussprache: 16. Oktober 2020

Berlin 2021

Acknowledgements

I am grateful for the inspiration, opportunities and support that have enabled me to realise

this thesis. I would like to express this gratitude in particular to Prof. Dr. Stefan Weinzierl

for the opportunity to complete my PhD at the Technische Universität Berlin. I would also

like to deeply thank Prof. Dr. Daniel Müllensiefen, who was so significant in my decision

to pursue a PhD. From playing football together at Goldsmiths to our collaborations on many

different research projects, I have learned so much from Daniel. The continuous support,

motivation, and expertise which I have received from Prof. Dr. Jochen Steffens, have also

been so important to guarantee the successful completion of this work. In addition, I wish

to thank Dr. Nikhil Masters, whose critical approach and expertise in economics has made

him truly indispensable in the consolidation of the framework presented in this thesis.

I gratefully acknowledge the funding for my PhD from the Studienstiftung des deutschen

Volkes. The stability and prestige provided by this scholarship has been crucial to my

research. I would like to further extend my sincere thanks to the collaborators that have

contributed to the scientific publications conducted within this thesis, namely: Prof. Dr.

Adrian North, Dr. Amanda Krause, Dr. Diana Omigie, Dr. Sanfilippo, Steve Keller, and

Tabitha Trahan. I have also been fortunate to work with some incredibly talented Masters

students: Björn Thorleifsson, Thomas Baker, Till Noé, Kerry Schofield, Emily-Beth Hill,

Heather Thuering, and Pattera Sutanthavibul; and I thank them for their enthusiasm and

dedication.

It is really important to me to also acknowledge the support and love of my family - my

parents, Lluís and Marta, and my sisters, Berta and Mei. None of this would have been

possible without them giving me such a wonderful childhood and great education. My

mother and grandfather first awakened my enthusiasm for music and my father taught me

the most important thing: to observe the world around us and ask critical questions. I also

wish to thank my friends in Barcelona, London, and Berlin for providing the happy

distractions in and out of the research world enabling me to enjoy my life to the full. Finally,

special thanks to my partner, Haia, for her unconditional love, support and valuable insights.

Abstract

Music-related decision making encompasses a wide range of behaviours including those

associated with music composition and performance, listening choices, music consumption,

and decisions involving music education and therapy. Although research programmes in

psychology and economics have contributed to an improved understanding of music-related

behaviour, historically these disciplines have been unconnected. In this thesis, I present The

Behavioural Economics of Music (BEM), a novel research framework that promotes the

study of musical behaviour using the tools of behavioural economics. Behavioural

economics aims to increase the explanatory power of neoclassical economics by relaxing

the rationality assumptions of homo economicus, incorporating insights from an array of

disciplines, including psychology, sociology, anthropology, biology, and neuroscience.

Thus, the BEM offers an empirically supported set of new concepts and methods that can

prove particularly suited to study the multi-faceted nature of music. Ten scientific

publications were conducted to support this thesis at three distinct levels. Firstly, two

literature reviews helped in understanding the need and value of the BEM. Secondly, four

empirical studies examined whether core concepts from behavioural economics, such as

bounded rationality and cognitive heuristics, can increase our understanding of music

decision making. Findings from these studies indicated that listeners are not utility

maximisers who use all information and time available to make optimal musical choices.

Instead, they are boundedly rational and, therefore, limited by their cognitive ability, time,

and information available. Finally, a further four studies focused on the application of the

BEM to improve music-related decision making in the real world. Two studies investigated

decision making in the context of choosing music for branding and advertising and two other

studies explored alternative methods to examine responses to music in the real world.

Overall, these studies emphasize the synergistic benefits of the BEM to both psychologists

and economists. Specifically, the BEM draws upon insights from a range of disciplines

(including important work from music psychology and music cognition), whilst

incorporating models from behavioural economic theory, thereby providing a framework

that is wide in scope as well as internally consistent.

Table of contents

Introduction 1

1.1

Decision making in music psychology . . . . . . . . . . . . . . . . . . . . 1

1.2

Music decision making in economics . . . . . . . . . . . . . . . . . . . . . 3

1.3

The Behavioural Economics of Music (BEM) . . . . . . . . . . . . . . . . 5

1.3.1

Bounded rationality in music . . . . . . . . . . . . . . . . . . . . . 6

1.4 Research questions and aims . . . . . . . . . . . . . . . . . . . . . . . . . 8

Scientific Publications 9

2.1

Why the BEM? .................................................................................................... 11

2.1.1

Visualizing music psychology (S1) ........................................................ 11

2.1.2

The Behavioural Economics of Music (S2) ............................................ 12

2.2

Bounded rationality in music decision making ................................................... 15

2.2.1

The repeated recording illusion (S3) ...................................................... 15

2.2.2

False memories in music listening (S4) .................................................. 17

2.2.3

Names and titles matter: Linguistic fluency and the affect heuristic (S5) 17

2.2.4

The effect of name recognition on listener choices (S6) ........................ 18

2.3

Real world applications ....................................................................................... 20

2.3.1

Source effects on the evaluation of music for advertising (S7) .............. 20

2.3.2

The effect of music recognition on consumer choice (S8) ..................... 21

2.3.3

The busking experiment: A field study (S9) ........................................... 23

2.3.4

Popular music lyrics and musicians’ gender over time (S10) ................. 23

Discussion 25

3.1

Theoretical contributions ..................................................................................... 25

3.2

Practical contributions ......................................................................................... 28

3.3

Future directions .................................................................................................. 31

3.4

Conclusion ........................................................................................................... 37

Table of contents v

References 39

Appendix A Visualizing music psychology (S1) 44

Appendix B The Behavioural Economics of Music: Systematic review (S2) 72

Appendix C The repeated recording illusion (S3) 102

Appendix D False memories in music listening (S4) 130

Appendix E Names and titles matter: Linguistic fluency and the affect heuristic (S5) 161

Appendix F The effect of name recognition on listener choices (S6) 193

Appendix G Source effects on the evaluation of music for advertising (S7) 211

Appendix H The effect of music recognition on consumer choice (S8) 238

Appendix I The busking experiment: A field study (S9) 266

Appendix J Popular music lyrics and musicians’ gender over time (S10) 283

Chapter 1

Introduction

This thesis is organised as follows. The following parts of Chapter 1 introduce two

disciplines that have been long concerned with the study of music decision making, i.e.,

music psychology and economics. The chapter follows by introducing the field of

behavioural economics, highlighting its potential benefits for music research, such as the role

of bounded rationality and related concepts. The chapter ends by outlining the research

questions and aims that guided this work. Chapter 2 summarises the ten scientific

publications conducted to support this thesis (see Table 2.1 for a list of publications; see

Appendix A-J for the full texts), emphasising their value to the BEM. Chapter 3 provides a

discussion of this work, focusing primarily on contributions and future directions.

1.1 Decision making in music psychology

Music psychology is a branch of psychology and musicology that aims to understand the

psychological processes by which music is perceived, processed, responded to, created, and

integrated into everyday life (see Hallam, Cross, & Thaut, 2016; Tan, Pfordresher, & Harré,

2017; Deutsch, 2013, for reviews). Decision making is inherent to many of these processes.

In this thesis, this is referred to as music decision making - i.e., the cognitive process of

evaluating and choosing between alternative options in any human behaviour related to

music. This section introduces some of the most prominent areas in music psychology that

rely on our understanding of music decision making.

Decision making is inherent in any creative process by which music is composed and impro-

vised. The process of composing and improvising music can be understood as a complex

sequence of judgments and decisions that involve both formal theoretical basis of music

1.2 Music decision making in economics

theory and creative basis mainly relying on aesthetic evaluations (e.g., Maeshiro, Nakayama,

& Maeshiro, 2011). For instance, deciding how to frame a harmonic accompaniment in the

context of a newly composed melody, how to extend an improvisation by using a motive

heard earlier, or how to produce a hit song that results in millions of sales. Psychologists

have begun to examine how composers make choices as part of the creative process (see

Impett, 2016, for a review) or how musicians decide which note to play next while impro-

vising (see Ashley, 2016, for a review). However, the decision making process underlying

music composition and improvisation is still poorly understood and, therefore, could benefit

significantly by considering insights from other disciplines, such as theoretical modelling

and utility theory from economics.

Decision making also plays a fundamental role in music performance evaluation (see

Waddell, 2018, for a review). Yet research on this topic within the field of music psychology

raises serious questions about the reliability and consistency of music performance

evaluation (Waddell, 2018). For example, jurors’ decisions in a high profile musical

competition were significantly influenced by the order in which the candidates performed:

those who performed first had a lower chance to win the competition, whereas those who

performed later had a higher chance (Flôres & Ginsburgh, 1996). Other studies have

identified several non-musical factors that can significantly bias the evaluation of music

performances, including musicians’ body movements (Wöllner & Behne, 2011), race and

gender (Elliott, 1995), and physical attractiveness (Griffiths, 2008). These findings illustrate

the need to better understand and improve decision making processes in the context of

evaluating music performances, such as increasing awareness between experts or using

evaluative methods that are less prone to human errors.

Music decision making is also central to the study of music preferences and listening

behaviour (see Lamont & Greasley, 2016; Lamont, Greasley, & Sloboda, 2016, for reviews).

Advances in technology have played a major role in the way people listen to music in daily

life, enabling them to listen to music in a wide variety of situations, such as whilst working,

exercising, travelling, or relaxing (Lamont et al., 2016). In these contexts, researchers have

identified several psychological needs that underlie listening behaviour, including

distraction, motivation, attention, emotional regulation, and stress reduction (e.g., Greb,

Schlotz, & Steffens, 2018; Linnemann, Wenzel, Grammes, Kubiak, & Nater, 2018;

Saarikallio & Erkkilä, 2007). Nevertheless, despite the wide range of psychological

approaches that have been used to investigate music preferences and listening behaviour,

there is currently no unified theory that has successfully addressed the complexities of this

topic and there is no model that can predict accurately a person’s preference or choice for

music at any given point in time (Lamont & Greasley, 2016).

Finally, the process of selecting music in applied contexts is heavily reliant on our under-

standing of music decision making. A clear example is the use of music in advertising and

branding. In this context, choosing an effective music branding strategy can have a positive

impact on consumers’ buying behaviour and attitudes towards brands (see Allan, 2007;

1.2 Music decision making in economics

North & Hargreaves, 2008; Oakes, 2007, for reviews). Nevertheless, a failure to adequately

use music can result in detrimental effects on communication effectiveness and consumer

behaviour (Allan, 2007; Lantos & Craton, 2012). Moreover, industry professionals often

rely on their gut instinct and personal experience to make musical choices (Schramm &

Spangardt, 2016). Thus, when choosing music for advertising is important to improve

current practices by designing efficient, reliable, and unbiased methods.

Overall, music psychology has examined music decision making in a wide variety of sit-

uations. However, this body of research is still relatively young and would benefit from

using a more sophisticated and unified understanding of the processes underlying human

judgments and decision making. This could be achieved by incorporating knowledge from

other disciplines that have been long concerned with the study of human decision making,

such as economics.

1.2 Music decision making in economics

Independent from music psychology, music-related decision making has also been studied

through the lens of economics (see Byun, 2016; Cameron, 2015, 2016; Krueger, 2005;

Tschmuck, 2017, for reviews). This research is mostly focused on economic decision

making related to music, such as the behaviour of firms in the music industry (Burke, 1996;

Sweeting, 2013; Rayna & Striukova, 2009), the economy of live music events (Decrop &

Derbaix, 2014; Hiller, 2016; Holt, 2010; Larsen & Hussels, 2011; Mortimer, Nosko, &

Sorensen, 2012), predicting music popularity in the charts (Bradlow & Fader, 2001; Elliott

& Simmons, 2011; Hendricks & Sorensen, 2009; Stevans & Sessions, 2005; Strobl &

Tucker, 2000), music consumption (Byun, 2016; Cameron, 2015), and music copyright and

piracy (see Oberholzer-Gee & Strumpf, 2009; Varian, 2005, for reviews).

This body of research provides valuable insights to understand key phenomena that

influence music consumption and shape the music market. For example, studies have found

associ- ations between economic conditions and changes in the characteristics of popular

music over time (Maymin, 2012; Pettijohn & Sacco, 2009; Pettijohn, Eastman, & Richard,

2012; Zullow, 1991): songs with faster tempo were most popular in good economic and

social times (Pettijohn et al., 2012), whereas pessimistic rumination in popular music lyrics

signifi- cantly predicted changes in consumption expenditures and General National Product

(GNP) growth (Zullow, 1991). The economic effects of music piracy (illegally downloading

and sharing music) have also generated great interest amongst economists (see Oberholzer-

Gee & Strumpf, 2009; Varian, 2005). This is, in part, because the controversial results in the

literature: while some studies show a relationship between the rise of music piracy and the

decline in music sales (e.g., Liebowitz, 2004, 2006), others found little evidence for this

causal relationship (e.g., Oberholzer-Gee & Strumpf, 2009). Another area of interest in the

economics literature is the growing importance of the live music business. Here, studies

suggest that the large decline in artists’ income from record sales has led to the increase in

1.2 Music decision making in economics

price for concert tickets (Larsen & Hussels, 2011; Mortimer et al., 2012; Tschmuck, 2017),

which grew by 82% from 1996 to 2003 (Krueger, 2005).

The economic literature on music-related decision making relies on a simple but powerful

model of behaviour put forward by standard economics: when making judgments and

decisions, people do so by maximizing a utility function, using all information available,

and processing this information with the appropriate time and mental resources (see

Coleman & Fararo, 1992, for one of many reviews on rational choice theory). Thus, this

body of research tends to assume that stakeholders such as composers, producers, artists,

labels, and listeners are rational actors. This rational model of behaviour is useful in

providing a coherent and internally consistent body of theory that offers rigorous and

falsifiable models of human behaviour. For instance, using standard economics one can

model a consumer’s rational choice between choosing to consume music over other goods,

or the production and supply curves in a rational music market that leads to equilibrium

price (Byun, 2016).

Overall, research on music decision making in economics provides highly valuable insights

to understand how people consume music and key factors that shape the music market.

Nevertheless, this body of research focuses mostly on economic decision making and does

not consider the psychological underpinnings known to be involved when people create,

perform, listen to, and respond to music. Importantly, the assumption of rationality put

forward by standard economics has been strongly challenged in the last years. For example,

laboratory and field experiments in both psychology and economics show that when making

judgments and decisions, people are limited by their mental capacity, use heuristics to solve

complex problems, their preferences are inconsistent over time, and their choices are

influenced by social information, their current emotional state, and the context (see

Dellavigna, 2009, for a review).

1.3 The Behavioural Economics of Music (BEM)

In recent years, researchers have begun to utilise theories and tools from behavioural eco-

nomics to study music-related decision making (e.g., Anglada-Tort & Müllensiefen, 2017;

Lonsdale & North, 2012). Behavioural economics is a scientific discipline that integrates

economics, psychology, and other social sciences to understand and improve judgments and

decision making of individuals and groups in a variety of domains (see Angner, 2012;

Cartwright, 2018, for introductory books; see Dhami, 2016, for a rigorous and detailed

analysis of the methodologies behind the field; see Kahneman, 2011; Tahler, 2015; Ariely,

2008, for informative books aimed to the general public). Behavioural economics arose from

a general discontent with the rational model of behaviour, stimulating diverse streams of

work that have expanded at a remarkable rate (Dhami, 2016). Thus, by increasing the

psychological underpinnings of standard economics, behavioural economics offers a more

realistic and comprehensive account of human judgments and decision making, generating

more accurate predictions and suitable methods for the study of human behaviour. This

thesis proposes The Behavioural Economics of Music (BEM), a novel interdisciplinary

approach to study the multi-faceted nature of musical behaviour by using knowledge and

tools from behavioural economics. The BEM emerges from the intersection between

behavioural economics and music research, aiming to contribute bidirectionally to both

fields (see Figure1.1).

Fig. 1.1 The Behavioural Economics of Music.

Heuristics and biases

Social preferences

Music perception

Music composition

Peer effects

Behavioural

Music performance

Time preferences

Dual process theory

Economics

Music

Research

Music preferences

Music consumption

1.3 The Behavioural Economics of Music (BEM)

When investigating music-related decision making, researchers from both economics and

psychology stand to gain from utilising the behavioural economics toolkit. Economists

studying music can benefit by moving away from the rigid neoclassical assumptions that

music composers, performers and listeners are rational actors, and instead apply more

empirically supported evidence when examining musical behaviour. Equally, since

behavioural economics still maintains its economic identity in terms of having falsifiable

models that are mathematically rigorous, psychologists can benefit from incorporating such

theoretical approach to address key issues in music research that have eluded researchers so

far.

Furthermore, behavioural economics is an established paradigm that has been hugely

influential in a variety of domains, leading to new interdisciplinary fields. Examples include

the behavioural economics of health (Blumenthal-Barby & Krieger, 2015; Rice, 2013),

education (Jabbar, 2011), climate change and energy use (Arne Brekke & Johansson-

Stenman, 2008; Frederiks, Stenner, & Hobman, 2015). At the same time, behavioural

economics is being applied to a wide variety of institutions and companies all over the world,

including governments, NGOs, and business entities such as Airbnb, Uber, and Google (e.g.,

Samson, 2017). Given that research in music decision making is still relatively young and

has no research agenda dedicated to this area exclusively, there is great potential in

incorporating insights from behavioural economics to study music decision making. This

work can be an important step towards creating this.

Finally, today it remains unclear to which extent music may prove exceptional or congruent

with general theories of human judgment and decision making. However, the rich,

multisensory, aesthetic, social, and highly emotional nature of music may challenge some

of these theories, enhancing their generalizability and scope. Music not only offers a unique

spectrum of situations to study human decision making, but it is also an important cultural

product, an artefact of society that is shaped with political and socioeconomic changes. Thus,

investigating music decision making provides a novel testing ground for general theories

that could reveal unique insights into behavioural economics.

1.3.1 Bounded rationality in music

Bounded rationality (Simon, 1955, 1982) is a core concept in behavioural economics that

challenges the rational model of behaviour put forward by standard economics. According

to bounded rationality, humans are limited by their mental capacity, available information,

and time when making judgments and decisions. As a consequence, individuals seek

decisions that satisfice (e.g., are good enough) rather than optimize (choosing the best

possible decision). A central part of this thesis is to better understand bounded rationality in

the context of music, as well as related concepts from behavioural economics that may help

increase our understanding of music decision making. To illustrate how some of these

concepts are particularly relevant in a music context, three examples are provided below:

1.3 The Behavioural Economics of Music (BEM)

First, extensive empirical evidence shows that to make efficient judgments and decisions

under bounded rationality, people rely on cognitive biases and heuristics. Heuristics are

mental shortcuts used by individuals to simplify complex decisions into easier to calculate

operations. Although they allow individuals to make decisions quickly and efficiently,

heuristics can systematically fail and lead to cognitive biases (see Dhami, 2016; Kahneman,

2011; Dawes & Hastie, 2010, for reviews). Heuristics and biases are highly relevant in a

music context, both in terms of the perception of music and how people respond to it. For

example, songs with more repetitive lyrics, which are easier to process in terms of

information, may be perceived as being liked more (processing fluency); individuals may

falsely remember sounds that come easily to the mind (availability heuristic); and listeners

may evaluate a music experience based on the most intense moment and the end (peak-end

rule).

Second, dual-process theory proposes a cognitive architecture based on two systems of

processing that support bounded rationality (Kahneman, 2003). Whereas the emotional

System-1 is unconscious, implicit, automatic, effortless and rapid, the cognitive System-2 is

conscious, explicit, controlled, effortful and slow (see Evans, 2008, for a review). Our

capacity for mental effort is limited and, consequently, mental processes will be assigned to

one of the two systems based on how much mental effort they require. Thus, dual-process

theories exploring the interaction between emotional and cognitive processes in the brain

can be useful for understanding music performance and listening behaviour. For example,

to determine the extent to which musicians’ decisions while performing are conscious or

unconscious, and how these decisions may be influenced by music expertise.

Finally, in 2007, the critically acclaimed band Radiohead surprised the music industry by

offering their new album “In Rainbows” as a digital download using a pay-what-you-want

(PWYW) agreement. Essentially, this meant that fans could pay as much as they liked for

the album, including a zero option. Although at odds with neoclassical economic theory, in

which consumers would download the album for free, many fans made voluntary payments

for the album. One possible explanation for such generous payments under PWYW is that

individuals exhibit social preferences, i.e., they care about the preferences of others (see

Fehr & Rangel, 2011, for a review). Thus, social preferences are an example of how

individuals do not always act as self-interested decision-makers and their decisions are

influenced by the context and information available.

1.4 Research questions and aims 8

1.4 Research questions and aims

This thesis had two main goals: (i) to gain a solid understanding of the role that behavioural

economics can play to increase our understanding of music decision making, and (ii) to

provide fruitful directions for future research. To achieve these goals, five research

questions guided the present work:

•

RQ1 - Where does music decision making sit within the music psychology literature?

•

RQ2 - To date, which studies have utilised behavioural economics for research on

music-related decision-making?

•

RQ3 - How can insights from behavioural economics, such as bounded rationality,

increase our understanding of musical behaviour?

•

RQ4 - How can behavioural economics improve music-related decision making in

real world applications?

•

RQ5 - Which are the most fruitful areas for future research on the BEM?

Ten scientific publications were conducted to address these research questions (see section

2 for a summary of these studies highlighting their main contributions to the present thesis;

see Appendix A-J for the full texts). First, a bibliometric study visualizing all published

literature on music psychology enabled the identification of an important gap in the music

psychology literature, namely, a lack of a research agenda dedicated exclusively on music

decision making (RQ1; see Appendix A for the full text). Second, a systematic literature

review was conducted to provide an up-to-date account of all studies that have utilised

behavioural economics for music research thus far (RQ2; Appendix B). The systematic

review also identified fruitful avenues for future research on the BEM (RQ5). Furthermore,

four empirical studies examined the role of bounded rationality and related concepts to

increase our understanding of music decision making (RQ3; Appendix C-F). Finally, four

studies focused on applying insights from behavioural economics to improve music-related

decision making in real world situations (RQ4; Appendix G-J).

Chapter 2

Scientific Publications

This thesis was written cumulatively and comprises a total of ten scientific publications

(seven are published in scientific journals and three were in preparation at the time this

thesis was submitted). The full texts of the ten publications are provided in the appendices

of this thesis as pre-prints. This section briefly summarises these publications (see Table

2.1 for a list of publications), highlighting their main contribution to the BEM. Although the

methods and scientific goals of these studies are diverse, they can be categorised into three

groups depending on how they support this thesis: Why the BEM? (see section 2.1), bounded

rationality in music decision making (see section 2.2), and real world applications (see

section 2.3).

Table 2.1 List of scientific publications.

ID Reference RQ

S10

Anglada-Tort, M., & Sanfilippo, K. R. M. (2019). Visualizaing Music

Psychology: A Bibliometric Analysis of Psychology of Music, Music

Perception, and Musicae Scientiae from 1973 to 2017. Music & Science, 2,

2059204318811786. https://doi.org/10.1177/2059204318811786

Anglada-Tort, M., Masters, N., Steffens, J., North, A., & Müllensiefen,

D. (in prep.). The Behavioural Economics of Music: Systematic Literature

Review and Future Directions. Manuscript in preparation.

Anglada-Tort, M., & Müllensiefen, D. (2017). The repeated recording

illusion: the effects of extrinsic and individual difference factors on

musical judgments. Music Perception: An Interdisciplinary Journal,

35(1), 94-117.

https://doi.org/10.1525/mp.2017.35.1.94

Anglada-Tort, M., Baker, T., & Müllensiefen, D. (2019). False mem-

ories in music listening: exploring the misinformation effect and

individual difference factors in auditory memory. Memory, 27(5), 612-627.

https://doi.org/10.1080/09658211.2018.1545858

Anglada-Tort, M., Steffens, J., & Müllensiefen, D. (2019). Names

and titles matter: The impact of linguistic fluency and the

af fect heuristic on aesthetic and value judgments of music.

Psychology of Aesthetics, Creativity, and the Arts, 13(3), 277-292.

http://dx.doi.org/10.1037/aca0000172

Steffens, J., Till, N., & Anglada-Tort, M. (in prep.). I know that song: The

effect of name recognition on listener choices when searching for music in

playlists. Manuscript in preparation.

Anglada-Tort, M., Keller, S., Steffens, J., & Müllensiefen, D. (2020).

The impact of source effects on the evaluation of music for advertising:

Are there differences in how advertising professionals and consumers

judge music? Advanced online publication, Journal of Advertising

Research. http://dx.doi.org/10.2501/JAR-2020-016

Anglada-Tort, M., Schofield, K., Tabitha, T., & Müllensiefen, D. (in

prep.). I’ve heard that brand before: The effect of music as a recognition

cue to influence consumer choice. Manuscript in preparation.

Anglada-Tort, M., Thueringer, H., & Omigie, D. (2019). The busking

experiment: A field study measuring behavioural responses to street

music performances. Psychomusicology: Music, Mind, and Brain, 29(1),

46. https://doi.org/10.1037/pmu0000236

Anglada-Tort, M., Krause, K., & North, A. C. (2019). Popular music

lyrics and musicians’ gender over time: A computational approach.

Psychology of Music. 0305735619871602.

https://doi.org/10.1177/0305735619871602

2, 5

3, 5

4, 5

’RQ’ denotes research question.

2.1 Why the BEM?

To understand the need and value of the BEM, two literature reviews were conducted

addressing three research questions (see Table 2.1; RQ1, 2 and 5). The first literature review

identified a notable gap in the music psychology literature, showing that to date, there is no

research agenda dedicated exclusively to decision making in the context of music (RQ1).

The second review constitutes one of the core parts of this thesis, providing an up-to-date

account of studies utilising behavioural economics for research on music decision making

(RQ2), as well as identifying fruitful avenues for future research on the BEM (RQ5).

2.1.1 Visualizing music psychology (S1)

This study aimed to analyse all literature published in the three most prominent scientific

journals in the field of music psychology (see Appendix A for the full text). Namely,

Psychology of Music, Music Perception, and Musicae Scientiae. Using all available

literature in Scopus, a total of 2,089 peer-reviewed articles, 2,632 authors, and 49 countries

were analysed, covering a period of 44 years (1973–2017). Visualization and bibliometric

techniques were used to investigate the growth of publications, author and country

productivity, collaborations, and research trends. Thus, the results of this study present

objective and measurable patterns seen across the development of music psychology

research included within these three journals. For example, from 1973 to 2017, there was a

clear increase in music psychology research, with a total growth rate of 11%. A core feature

of this paper is the visualization network map of music psychology (see Figure2.1). The

network map shows the most influential keywords in the literature (i.e., keywords used by

the authors to define their publications) as well as how they co-occur with others, creating

different research trends and themes. For instance, the keywords “music” and “emotion” are

the most influential keywords in the literature, which is in line with the general interest and

significant increase in research on music and emotion (e.g., Eerola & Vuoskoski, 2013). The

map also provides an overlay visualization that adds a time dimension to each keyword (i.e.,

colour-coding each keyword based on the average publication year), which suggests

different trends in the popularity of each keyword over time. Thus, the network map is useful

in summarizing the complex field of music psychology in a single picture.

Overall, this study contributed to the literature by providing the first large-scale bibliometric

analysis that investigates general research trends and gaps in the field of music psychology.

Using bibliometric techniques to visualize the past and present of research in music

psychology enables critical observations and conclusions, opening many interesting avenues

for future research in the field. In the context of the BEM, this was important to identify a

gap in the literature. Namely, there is currently no research agenda dedicated exclusively to

study decision making in music. Despite this, decision making is inherent to many of the

influential concepts and research trends identified in the study, including, music

performance, creativity, improvisation, perception, music preference, music listening, and

2.1 Why the BEM?

music therapy.

Fig. 2.1 Network visualization map of keyword co-occurrences in music psychology.

The map shows the 75 most influential keywords used by researchers to describe their articles and how often

they co-occur with others, indicating main research trends and themes in music psychology. The width of the

line shows the strength of the co-occurrence between keywords, while the size of the circle indicates the total

number of occurrences. The colour of the circle indicates the average year of publications.

2.1.2 The Behavioural Economics of Music (S2)

This paper is the core theoretical contribution of this thesis. The paper conducted a

systematic literature review to provide an up-to-date account of studies utilising behavioural

economics to investigate music decision making (see Appendix B for the full text). Using a

robust search strategy that is highly representative of the behavioural economics literature,

the systematic review identified a total of 33 papers within four BEM areas that readily

apply to music decision making. Thus, the paper is organised around these areas, enabling

the reader to fully understand the scope of existing research as well as giving direction for

future research within each area. The main findings, organised by BEM area, are

summarised below:

Heuristics and biases: the systematic review identified 16 studies applying six cog- nitive

biases and heuristics to various aspects of musical behaviour (see Table 2.2 for definitions

and music examples). Several studies confirmed that individuals rely on judgmental heuristics

when listening to and evaluating music, allowing them to simplify complex decisions into

easier-to-calculate operations, but also leading to systematic errors. Studies on processing

fluency showed that fluency manipulations in music, such as repetition and consonance, can

2.1 Why the BEM?

influence music perception and, in turn, affect preferential judgments. Finally, studies on

framing effects found that presenting the same music stimuli with different contextual

information can systematically change a person’s preferences for the music.

Social decision making: nine studies showed that music decision-making is largely influenced

by social preferences and information. Firstly, social preferences, such as reciprocity and

guilt, are important to understand consumers’ motivation to engage in different revenue

models for music consumption, including voluntary payments for music. Social preferences

can also help better understand pricing strategies in the concert industry. Secondly, peer

effects can play a determinant role in music preferences and choices, which in turn can

influence the music market and determine outcomes such as the next successful artist or hit

song.

Behavioural time preferences: four studies found that behavioural time preferences can give

a deeper insight into how music is valued and consumed over time. Two studies found that

when consuming music online, consumers disproportionally prioritise im- mediate benefits

over future gains, providing solid evidence for hyperbolic discounting in music consumption.

The two other studies focused on time preferences for music consumption in terms of its

hedonic value. They showed that listeners’ ability to predict pleasure in their future music

consumption is rather low and when choosing music repeatedly over time, listeners do not

always choose music that maximizes their pleasure but instead seek variety.

Dual-process theory: four studies showed that exploring the interaction between System-1

and System-2 processes in the brain can help increase our understanding of how musicians

make decisions while performing, as well as the mediating role of music expertise.

2.1 Why the BEM?

Overall, the examination of these studies enabled to gain a solid understanding of the role

that behavioural economics has played in music research thus far. Furthermore, it was clear

from the review that the BEM is a relatively new approach and behavioural economics has

just begun to be applied in the domain of music. For example, the vast majority of the

retrieved studies were published in the last 10 years and over half of them in the last five

years. The findings also suggested that the BEM is fairly multi-disciplinary. While half of

the studies came from the music psychology literature, the other half came from behavioural

economics. Finally, the paper concluded by providing fruitful ideas and directions for future

research, both within the identified BEM areas and beyond.

Table 2.2 Heuristics and biases identified in the systematic review (S2).

Definition Music Example

Processing

fluency

Availability

heuristic

Represent-

ativeness

heuristic

Affect

heuristic

Framing

effect

Peak-end

rule

Human tendency to evaluate easy-to-process

information more positively than similar but

more difficult-to-process information (Reber

et al., 2004).

When judging the frequency and probability of

events, people rely on the ease with which

examples come to their minds (Tversky &

Kahneman, 1974).

People estimate the likelihood of an event by

comparing it to an existing event of similar

characteristics that already exists in their

minds (Tversky & Kahneman, 1974).

Human tendency to rely heavily upon our

emotional state when making judgments and

decisions (Slovic et al., 2002).

People make decisions based on how the

options are presented or “framed” (e.g., as a

loss or as a gain) (Kahneman & Tversky,

1979).

People judge an experience largely based on

how they felt at its peak (i.e., the most intense

point) and at its end (Kahneman &

Fredrickson, 1993).

Songs with more repetitive lyrics, which

are easier to process in terms of

information, are perceived as being liked

more (Nunes et al., 2015).

Listeners falsely remember sounds that

come easily to their minds (Vuvan et al.,

2014).

Stereotypes between music genres and fans

can be misjudged (Lonsdale & North,

2011).

Individuals evaluate music based on an

associated emotional feeling (Anglada-

Tort et al., 2018)

Contextual information presented with mu-

sic can systematically affect a person’s

judgment of the music (North &

Hargreaves, 2004)

Listeners evaluate a music experience

based on the most intense moment and at

the end (Rozin, et al., 2004).

2.2 Bounded rationality in music decision making

To further develop the BEM within this thesis, four empirical studies explored whether

bounded rationality and related insights from behavioural economics may prove valuable

for music research (RQ3). The four studies showed that when making musical judgments

and decisions, listeners are limited by their mental capacity (e.g., memory constraints), time,

and information available (e.g., song titles, post-event information, or descriptions about the

performer). Consequently, listeners rely on cognitive biases and heuristics that do not

depend on the music stimuli themselves.

2.2.1 The repeated recording illusion (S3)

This study investigated the extent to which listeners are limited by memory constraints and

the context when evaluating music performance (see Appendix C for the full text). To do

so, a novel experimental paradigm was developed, namely, the repeated recording illusion.

In this paradigm, participants (N= 72) were told to listen to three “different” musical

performances of an original piece. However, unbeknownst to them, they were exposed to

the same repeated recording three times in succession. Each time, the recording was

accompanied by a text suggesting a low, medium, or high prestige of the performer.

Participants evaluated the music using several rating scales (i.e., liking, timing, tone quality,

pitch accuracy, emotional quality, and overall quality). The procedure was repeated using a

piece of highly familiar rock music and a piece of unfamiliar classical music. Potentially

related extrinsic factors (i.e., explicit information and repeated exposure) and individual

differences were investigated.

Results showed that most participants (75%) believed that they had heard different musical

performances while, in fact, they were identical. In the two music conditions, participants

evaluated the same recording significantly more positively when it was presented with a

high-prestige text compared to low and medium texts. However, the position of the

recording only had a significant impact on the familiar music condition. To capture higher-

order interactions between extrinsic and individual difference factors, a regression tree

model based on permutation tests was computed. The dependent variable was a one-factor

solution indicated by a Principal Component Analysis, where a negative score indicated an

overall negative evaluation and a positive score an overall positive evaluation. The predictor

variables were prestige effect, the position of the recording, music conditions (classical-

unfamiliar vs. rock-familiar), and seven individual difference variables, including age,

personality, and musicality. Figure2.2 shows the structure of the regression tree the model,

which shows that only 3 of the predictor variables had a significant impact on performance

evaluation, i.e., explicit information, repeated exposure, and the music condition. Note that

none of the individual differences were significant predictors.

Overall, these findings highlight the fallibility of music evaluation and support the notion of

bounded rationality in musical behaviour, showing that musical judgments are limited by

2.2 Bounded rationality in music decision making

memory constraints, cognitive biases, and the context. The influence of explicit information

and the partial effect of repeated exposure are discussed in terms of the anchoring heuristic

(Tversky & Kahneman, 1974) and processing fluency (Reber, Schwarz, & Winkielman,

2004).

Fig. 2.2 The influence of non-musical factors on music performance evaluation.

The regression tree model is useful in identifying the most predictive variables influencing music performance

evaluation, as well as specific conditions that lead to particularly high (left node, 3) and low (right node, 9)

ratings. The tree model can be interpreted by starting at the top and following each branch down, to arrive at a

terminal node. A path to a terminal node describes the interaction of experimental conditions that lead to a

particular subset of ratings.

2.2 Bounded rationality in music decision making

2.2.2 False memories in music listening (S4)

When people listen to music or experience music in a live performance, they are normally

exposed to related information at some point after the event. This study examined for the

first time whether post-event misinformation can induce false memories in music (see

Appendix D for the full text). Though misinformation effects have been demonstrated

extensively within visual tasks, they have not yet been explored in the realm of non-visual

auditory stimuli. Besides, the study explored individual difference factors potentially

associated with false memory susceptibility in music, including age, suggestibility,

personality, and musical training. In two music recognition tasks, participants (N = 151)

listened to an initial music track, which unbeknownst to them was missing an instrument.

They were then presented with post-event information which either did or did not suggest

the presence of the missing instrument. The presence of misinformation resulted in

significantly poorer performance on the music recognition tasks (d = .43), suggesting the

existence of false musical memories. A random forest analysis indicated that music

expertise was not significantly associated with misinformation susceptibility. These

findings support previous research on the fallibility of human memory and demonstrate, to

some extent, the generality of the misinformation effect to a non-visual auditory domain. In

the context of the BEM, this is important to further support the notion of bounded rationality

in music decision making, in particular demonstrating the fallibility of memory-based

judgments of music.

2.2.3 Names and titles matter: Linguistic fluency and the affect heuris-

tic (S5)

This study manipulated the song titles and artist names presented with music to examine the

influence of two well- known heuristic principles on aesthetic and value judgments of music:

processing fluency (Experiment 1) (Reber et al., 2004) and the affect heuristic (Experiment

2) (Slovic, Finucane, Peters, & MacGregor, 2002) (see Appendix E for the full text). In

Experiment 1, the same music excerpts were presented with easy-to-pronounce (fluent) and

difficult-to-pronounce (disfluent) names. The names consisted of a list of Turkish names

that were shown in a previous study to be fluent or disfluent to English speakers (Shah &

Oppenheimer, 2007). Native English-speaking participants (N= 48) listened to the music

stimuli and provided evaluations on different scales measuring aesthetic properties (e.g.,

like, emotional expressivity, quality) and subjective value of the music (e.g., likelihood to

attend a concert of the artists or to recommend the song to a friend). Results indicated a main

significant effect of fluency. In particular, participants evaluated the same music excerpts

significantly more positively when presented with fluent names than when presented with

disfluent names.

In Experiment 2 (N= 100), the same procedure was used, but instead manipulating the

emotional content of the titles. Thus, the music excerpts were presented with positive (e.g.,

2.2 Bounded rationality in music decision making

Kiss), negative (e.g., Suicide), and neutral (e.g., Window) titles. This time, at the end of the

experiment, participants also performed an unexpected free recall task (i.e., write down the

songs they remembered). In both aesthetic and subjective value evaluations, presenting the

music with negative titles resulted in the lowest judgments. When looking at the effects of

emotionality on memory, results showed that music excerpts presented with neutral and

negative titles were remembered significantly more often than positive titles.

Overall, these findings suggest that like any other human judgments, evaluations of music

also rely on heuristic principles that do not necessarily depend on the aesthetic stimuli

themselves. These heuristics operate even when the information processed is minimal, such

as changing the linguistic properties of titles presented with music.

2.2.4 The effect of name recognition on listener choices (S6)

When searching for and choosing music in playlists, individuals may rely on judgment

heuristics to make fast (in terms of computing time) and frugal (in the use of information)

decisions. This study addressed this issue by investigating for the first time the role of the

recognition heuristic on musical choices when listeners search for music in playlists (see

Appendix F for the full text). The recognition heuristic states that when people are faced

with recognised and unrecognised options, they infer that the recognized one has the higher

value with respect to the criterion being judged and, therefore, they tend to choose it

(Goldstein & Gigerenzer, 2002). In particular, the study extended the paradigm used in

Oeusoonthornwattana and Shanks, 2010 to a listening task with 10 alternative choices,

simulating a common listening playlist. Before the main experimental task, participants

(German and English speakers) had to learn a list of Spanish names. This manipulation made

it possible to create playlists using novel music paired with Spanish names that had been

previously learned (i.e., recognisable) or were completely novel. In the main choosing task,

participants were presented with ten songs in a playlist format and had to choose their

favourite five. To study the role of recognition-based heuristics in the presence and absence

of music information, participants searched for and selected music in two playlist conditions:

a titles-only condition (where they could only choose music based on visual cues – i.e., the

Spanish names) and a titles-and-music condition (where they could choose music based on

both visual and auditory cues – i.e., they could also listen to the music).

2.2 Bounded rationality in music decision making

Figure 2.3 depicts the mean choice proportion of music clips paired with learned and novel

names in the two choosing conditions. Results confirmed that there was a significant effect

of name recognition in the two choosing conditions, but this effect was larger when

participants chose music based on visual information only (titles condition). Moreover,

participants’ preferences for the selected music were also influenced by recognition - i.e.,

the same music clips were significantly more liked when paired with learned names than

when paired with novel ones. These results show that listeners rely on the recognition

heuristic when both deciding which songs to choose in a playlist and developing music

preferences.

Fig. 2.3 Mean choice proportion of music clips when paired with learned and novel names

in both playlist conditions (Error bars represent 95% CI).

1.00

0.75

0.50

0.25

0.00

Titles Only

Titles and Music

Name recognition Novel Learned

Listeners rely on recognition cues when searching for and choosing music in the two playlist conditions.

However, the effect of name recognition was larger when listeners chose music only based on visual cues (title

only) than they chose music based on both visual and music cues.

Mean choice proportion of music clips

2.3 Real world applications

The last part of this thesis focused on the application of the BEM to improve music-related

decision making in the real world (RQ4). Two studies applied insights from behavioural

economics to better understand the decision making process to select music for branding

and advertising, whereas the other two studies explored alternative methods to examine

music preferences in the real world, including field research and naturalistic data

approaches.

2.3.1 Source effects on the evaluation of music for advertising (S7)

This study focuses on improving music-related decision making in branding and advertising

(see Appendix G for the full text). Music choices can have profound effects on brand

communications, but the process of evaluating and selecting music for advertising and

branding is poorly understood. When choosing music for advertisements, professionals are

influenced by a large number of factors that could impair their judgment. Based on insights

from behavioural economics, this study examined source effects in the evaluation of

advertising music by professionals and nonprofessionals (a group of general consumers). In

experiment 1, a group of advertising and marketing professionals listened to and evaluated

music from three assigned sources: generic music libraries, music commissioned by

production companies, or "real" artists (performing artists in the market). In experiment 2,

the same procedure we repeated with a sample of general consumers. Results showed that

advertising professionals gave significantly more favourable evaluations— higher in quality,

authenticity, and expected cost— when they thought the music was sourced from performing

artists compared with less credible and attractive sources. In contrast, consumers were not

affected by source cues at all. Importantly, ad professionals were more aware of the

influence of source cues than the group of consumers (see Figure 2.4), highlighting the

difficulty for domain experts in advertising to build up effective cognitive defences against

source effects. The differential effects of music source in the two groups could prove costly

for brands. Professionals may recommend that their clients pay a premium for music coming

from performing artists, but brands may see little or no added benefits if the source of the

music does not matter to the listening public. Potential solutions to mitigate source effects

include increasing awareness among professionals and measuring the impact of advertising

music and source on target consumers.

2.3 Real world applications

Fig. 2.4 Awareness of sources effects in consumers and ad professionals.

Consumers Professionals

Advertising professionals were significantly more aware of the influence of source effects when

choosing music for ads than the group of general consumers. This suggests that for domain experts

in advertising it is difficult to build up effective cognitive defences against this bias. In the Figure,

violin plots are used in addition to box plots to show the probability density of the data at different

values (smoothed using a kernel density estimator).

*** Denotes that the difference between groups is highly significant, as indicated by an independent t-test.

2.3.2 The effect of music recognition on consumer choice (S8)

This study is another example of how insights from behavioural economics can be

successfully applied to music decision making in the context of branding and advertising. In

particular, two experiments aimed to quantify the effectiveness of using music as a

recognition cue to influence consumer choice by means of the recognition heuristic (see

Appendix H for the full text). A pilot study was conducted (N= 2,854) to select 24 unfamiliar

excerpts of advertising music and 24 unfamiliar brands. Prior to the main experimental task,

participants memorised part of these unknown music clips. In a choice task, participants

were then presented with pairs of brands, one presented with previously learned music and

the other with novel music. Their task was to choose which brand they would purchase when

buying different products (e.g., headphones, cameras). Results revealed that pairing brands

with music that can be recognized by target consumers increased the likelihood that they

will choose the brand by 6%, which corresponds to a small but significant effect size

****

Source effect awareness

2.3 Real world applications

(d = .21). Furthermore, music preferences were a key moderating factor in the success of

recognition-based heuristics. Exploratory results indicated that participants only relied on

music recognition when they liked the music, whereas recognition-based heuristics did not

play an influential role when the music was disliked (see Figure 2.5). Therefore, when using

music to influence consumer behaviour, it is important to consider how recognition cues are

processed in combination with other information, such as music preferences. This is valuable

to inform brands in terms of measuring the value of their investment when working with

music.

Fig. 2.5 Mean choice proportion of brands paired with learned and novel clips when the

music was liked and disliked (Error bars represent 95% CI.).

1.00

0.75

0.50

0.25

0.00

Dislike

Music recognition Novel Learned

Music recognition only had a significant effect on participants’ choices when the music was liked, whereas

recognition-based heuristics did not play a significant role when participants disliked the music. The relative

difference between choosing a brand paired with a learned and a novel music clip when the music was liked

was 18%, whereas when the music was disliked the difference was 5.1%.

Mean choice proportion of brands

2.3 Real world applications

2.3.3 The busking experiment: A field study (S9)

This study applied methods from behavioural economics to a different music problem (see

Appendix I for the full text). That is, what makes a successful street musician? And which

aspects of the performative act might influence people’s economic responses? To address

this question, a field experiment was conducted with a professional busker in the London

Underground over the course of 24 days. The study primary aim to investigate the extent to

which performative aspects influence behavioural responses to music street performances.

Two aspects of the performance were manipulated: familiarity of the music (familiar vs.

unfamiliar) and body movements (expressive vs. restricted). The amount of money donated

and the number of donors were recorded. A total of 278 people donated over the experiment.

The music stimuli, which was selected in a previous study to differ only in familiarity, had

been previously recorded by the busker. During the experimental sessions, the busker lip-

synced to the pre-recorded recordings. Thus, the audio input in the experiment remained

identical across sessions and the only variables that changed across conditions were the

familiarity of the music and the expressivity of performed body movements. The results

indicated that neither music familiarity nor the performer’s body movements had a

significant impact on the amount of money donated or the number of donors. Importantly,

the results do not support previous literature investigating the influence of familiarity and

performers’ body movements, typically conducted in laboratory and artificial environments.

The findings are further discussed with regard to potential extraneous variables that may be

crucial to control for in similar field experiments (e.g., location of the performance, physical

appearance, and the bandwagon effect) and the advantages of field versus laboratory

experiments.

2.3.4 Popular music lyrics and musicians’ gender over time (S10)

This study applied a naturalistic data approach to investigate preferences for popular music

in the UK over time (see Appendix J for the full text). Data on the singles sales charts from

1960 to 2015 was analysed as a proxy of music preferences. Note that singles sales charts is

determined by weekly sales, downloads, and streaming of music. With this data, the study

focused on how the gender distribution of the United Kingdom’s most popular artists has

changed over time and the extent to which these changes might relate to popular music lyrics.

Using data mining and machine learning techniques, all songs that reached the UK weekly

top 5 sales charts from 1960 to 2015 were analysed (4,222 songs). A computational analysis

of the lyrics was conducted to measure a total of 36 lyrical variables per song. Results

showed a significant inequality in gender representation on the charts. However, the

presence of female musicians increased significantly over the period covered in the study.

The most critical inflection points leading to changes in the prevalence of female musicians

were in 1968, 1976, and 1984. Linear mixed-effects models showed that the total number of

words and the use of self-reference in popular music lyrics changed significantly as a

function of musicians’ gender distribution over time, and particularly around the three

2.3 Real world applications

critical inflection points identified. One of the most interesting trends found in the study is

shown in Figure 2.6: Regardless of gender, there was a drastic increase in the total number

of words over time, whereas the diversity of vocabulary (i.e., the number of different words

in a song divided by the total words) decreased significantly over time, suggesting that UK

popular music lyrics have become more repetitive over time. The use of data mining and

machine learning techniques (e.g., classification tree models and random forest) offered

several advantages in comparison to the statistical tools used in earlier studies.

Fig. 2.6 Mean words per song (left) and diversity of vocabulary (right) in UK popular music

lyrics from 1960 to 2015.

600

Total number of words

0.5

Diversity of vocabulary

400

0.4

0.3

Artist Gender

Both

Female

Male

200

1960

1980

Year

2000

0.2

1960

1980

Year

2000

There was a drastic increase in the total number of words used in popular songs from 1960 to 2015 (left plot).

In contrast, when looking at a measure of the diversity of vocabulary (i.e., the number of different words

divided by the total words in a song), the results showed that popular songs became less varied and more

repetitive over time. This finding occurred regardless of the gender of the artists or band. Note that Both

indicate bands or artists with both female and male members.

Mean words per song

Diversity of vocabulary

Chapter 3

Discussion

This section starts by discussing the main theoretical and practical contributions of this

thesis, with a focus on what have we learned from applying behavioural economics in the

context of music. The section ends with discussing directions for new insights and valuable

future research and, finally, concludes.

3.1 Theoretical contributions

The main theoretical contribution of this thesis is the conception of the BEM, an

interdisciplinary but unified framework with which we can increase our understanding of

musical behaviour (see Figure 1.1). Two literature reviews and eight empirical

investigations (see Table 2.1 for a list of publications; see Appendix A-J for the full texts)

demonstrate the value and potential of this novel approach.

The BEM contributes most significantly to the existing bodies of music research in both

standard economics and psychology. Economists interested in music will benefit by moving

away from the more rigid assumptions of standard economics and consider the

psychological underpinnings known to be involved in musical behaviour. For example, in

several studies conducted within this thesis, we learned that listeners are not utility

maximisers who use all information and time available to make optimal musical choices.

Instead, there are several psychological constraints that limit their ability to evaluate and

choose music, such as memory and the contextual information often presented with music.

Incorporating these insights will help building a more realistic and comprehensive account

of music-related decision making.

3.3 Future directions

In comparison to other stimuli, music is experiential, multisensory, aesthetic, social, and

highly emotional. Such intrinsic properties may prove particularly useful to test economic

theories and enhance their generalizability and scope. For example, music is highly effective

in evoking strong emotions in their listeners, such as chill experiences – i.e., the

phenomenon of chills or goosebumps caused by intense emotion that come from listening

to a specific piece of music (Goldstein, 1980). Thus, music can be an efficient and

inexpensive stimulus to study the role of emotion in decision making. For instance, S5 (see

section 2.2.3 - Appendix E) found an interaction between the affect heuristic and the

emotional content of the music, suggesting that the impact of this heuristic on decision

making may differ when using music stimuli in comparison with other stimuli. Similarly,

music is social and largely influenced by culture. By investigating properties of popular

music (e.g., lyrics) and characteristics of the artists (e.g., gender), S10 (see section 2.3.4 -

Appendix J) showed how popular music can be used as a cultural product to study how the

preferences and values of a society are shaped by political and socioeconomic changes.

On the other hand, phycologists will gain from considering behavioural economics as a

toolkit by which to address key music problems that. In particular, the BEM approach allows

psychologists to rethink the study of musical behaviour using a new (and empirically

supported) set of concepts and theories, such as bounded rationality, dual-process theory,

and behavioural game theory. Whilst these insights have been highly influential in the study

of human behaviour and decision making, they have rarely been applied to examine musical

behaviour. For example, to date, the notion of heuristic processing has been mostly

overlooked in the music psychology literature. Nevertheless, four studies conducted in this

thesis (see section 2.2) indicated that heuristics play a central role in music listening and

choice behaviour.

The studies conducted in this thesis are important in demonstrating the value of applying

insights from behavioural economics to study music decision making. However, they mostly

focused on one BEM area (i.e., cognitive biases and heuristics) and one aspect of music

decision making (i.e., music preferences and listening behaviour). To address this issue, S2

(see section 2.1.2 - Appendix B) provided an up-to-date account of all studies that utilised

behavioural economics for research on music-related decision making. This study

contributes significantly to the literature by showing which areas within behavioural

economics can generate new and valuable insights into the study of music decision making,

both in terms of research methods and theory. The systematic review identified 33 studies

organised in four distinctive BEM that readily apply to music decision making: cognitive

biases and heuristics, social decision making, behavioural time preferences, and dual-

process theory. Each of these BEM areas adds value to the existing bodies of music research

in both psychology and economics. For instance, social decision making is an area within

behavioural economics that examines how decisions are influenced by social information

and preferences. Although at odds with neoclassical economic theory, social preferences

(i.e., altruism, reciprocity, and fairness concern) can explain why consumers choose to pay

3.3 Future directions

voluntarily for music, a phenomenon that has puzzled researchers for a long time.

Similarly, behavioural time preferences can enable a deeper understanding of how music is

valued and consumed over time. Notably, individuals exhibit present-biased time

preferences, i.e., they have a strong preference for immediate gratification (O’ Donoghue &

Rabin, 1999). Since music is a hedonic good (i.e., multisensory based on experiential

consumption), individuals may place an even higher weight on outcomes that occur in the

present rather than the future. This has implications for how consumers select music,

particularly with the emergence of music streaming platforms providing music

instantaneously. A further area of behavioural economics, dual-process theory, explores the

interaction between emotional and cognitive processes in the brain. Dual-process theory can

be used to study decision making in the context of music composition and performance. For

example, investigating the interaction between these two systems can help better understand

conscious states while musicians perform and how these may impact on the quality of their

performances.

Another theoretical implication of this thesis is the focus on understanding the role of context

in music evaluation and decision making. Research within music psychology has identified

three main interconnected factors that influence people when listening to and evaluating

music: the music, the listener, and the listening context (see Hargreaves, North, & Tarrant,

2006; LeBlanc, 1982, for theoretical models considering the three factors; see Greasley &

Lamont, 2016; North & Hargreaves, 2008, for research reviews). Traditionally, the vast

majority of studies have focused on the music and the listener. Comparatively, less attention

has been paid to the listening context. In this thesis, six empirical investigations manipulated

contextual factors presented with the music stimuli to investigate its effects on musical

behaviour, including artists names, song titles, information about the artists, post-event

information about the music piece, and the source of the music. These studies consistently

show that music decision making does not happen in a vacuum, but is significantly

influenced by the context. More specifically, contextual information can lead listeners to

perceive different musical performances when in fact they are identical (S3; see section 2.2.1

- Appendix C); generate false musical memories of a past music event (S4; see section 2.2.2

- Appendix D); influence music judgments and decision even when the contextual

manipulation is minimal, such as only changing linguistic aspects of titles presented with

music (S5 and S6; see section 2.2.3 and 2.2.4 - Appendix E and F); and cause potentially

negative biases amongst ad professionals when choosing music for advertising (S7; see

section 2.3.1 - Appendix G).

Furthermore, the studies conducted in this thesis contribute towards a better understanding

of the role of music expertise in music evaluation and decision making. Previous research

consistently shows that highly trained musicians outperform non-musicians in several

musical tasks, such as short-term and working memory tasks with music stimuli (see

Talamini, Altoe, Carretti, & Grassi, 2017, for a review). Thus, it seems plausible that since

musicians’ cognitive abilities to perceive and process music are higher than non-musicians,

3.3 Future directions

they should perhaps be less influenced by contextual factors and cognitive heuristics.

Several of the studies conducted in this thesis addressed this issue by collecting data on

participants’ musical background, including both musical training and active engagement to

music. These studies, which collected data on more than 500 participants, showed that music

expertise does not have a protective effect against contextual factors. Besides, highly trained

musicians are not any more or any less susceptible to cognitive biases and heuristics than

non-musicians. Thus, contextual factors and heuristics seem to influence listeners regardless

of their previous experience to music. Although these results might seem counterintuitive at

first, they are consistent with the behavioural economics literature on the "expert problem"

(e.g., Hall, Ariss, and Todorov, 2007; Reyna, Chick, Corbin, and Hsia, 2014; Taleb, 2007),

showing that in certain conditions and domains, more knowledge and expertise does not

necessarily lead to more accurate and less biased judgments and decisions.

3.2 Practical contributions

A main practical contribution of this thesis is the wide variety of methods and paradigms

used to investigate different aspects of music decision making. For example, S3 (see section

2.2.1 - Appendix C) proposed the repeated recording illusion, a novel paradigm that is useful

to investigate non-musical factors in music evaluation because it allows for the study of

their effects while the music remains the same. S4 (see section 2.2.2 - Appendix D) applied,

for the first time, the misinformation paradigm using music instead of visual materials,

showing that listeners generate false memories in a music context. Both S5 (see section

2.2.3 - Appendix E) and S6 (see section 2.2.4 - Appendix F) adapted successfully existing

paradigms in the behavioural economics literature to study the effects of cognitive heuristics

in music evaluation and decision making. S5 adapted a well-known experiment from

behavioural economics (Shah & Oppenheimer, 2007) to examine linguistic fluency,

whereas S6 adapted a common paradigm to investigate the recognition heuristic in

preferential choice tasks (Oeusoonthornwattana & Shanks, 2010) to study musical choices

when listeners search for music in playlists. Overall, these studies emphasize the potential

of applying methods and paradigms from behavioural economics to study similar

phenomena in music.

This thesis also explored other methods beyond those commonly used in controlled and

artificial studies. This is important because controlled studies conducted in laboratories and

other artificial environments are susceptible, among other things, to two major problems

(Carpenter, Harrison, & List, 2005; Reis & Judd, 2000): a lack of external validity—the

extent to which the results are generalizable beyond the research setting and participant

pool—and a lack of ecological validity—the degree to which the results apply to the real

world situation under study. Note that issues related to poor ecological validity and

generalizability are taken particularly seriously by economists and behavioural scientists

(Harrison & List, 2004; Levitt & List, 2007). As argued by Levitt and List (2007), “Perhaps

3.3 Future directions

the most fundamental question in experimental economics is whether findings from the lab

are likely to provide reliable inferences outside of the laboratory” (p. 179). Thus, it was

important to consider further ways to examine behavioural responses to music in natural

environments, once sufficient scientific grounding has been obtained based on laboratory-

generated data.

This was one of the motivations for S9 (see section 2.3.3 - Appendix I), which used a field

research approach to investigate responses to musical performances in a naturalistic busking

environment. The charitable behaviour of passersby’ (i.e., amount of money donated) was

recorded while a professional busker performed in the London Underground over the course

of 24 days. Two factors of the performance, commonly investigated in lab studies, were

manipulated: familiarity of the music and body movements. Contrary to common findings

in lab studies looking at the same factors, the results indicated that neither music familiarity

nor the performer’s body movements had a significant impact on the amount of money

donated. These discrepancies might be due to differences in the ecological validity between

laboratory and field studies. For instance, in the laboratory, participants are always aware of

their participation in a scientific study and their only goal is to listen carefully to the music

while evaluating it in a highly controlled and quiet environment. Therefore, the advantages

of field studies over lab studies include high ecological validity and avoiding problems

associated with self-reported assessments. On the other hand, S10 (see section 2.3.4 -

Appendix J) used a big data approach that offers high ecological validity but also allows to

analyse large datasets. In particular, the study analysed naturalistic data from the singles

sales charts to examine how preferences for popular music have changed over time. This

approach allowed for the study of responses to music in the real world that are neither

obtained nor affected by the actions of researchers. Moreover, it made it relatively easy to

collect and analyse a large dataset (4,222 songs and 2,287 artists) covering 55 years (1960-

2015). Overall, S9 and S10 are useful in exploring alternative methods to study musical

behaviour that do not suffer from poor external and ecological validity.

Another contribution of this thesis includes the wide variety of analysis techniques used

across the eight empirical studies to examine human responses to music. Firstly, linear

mixed-effect models proved to be very efficient to test main hypotheses when using

repeated- measured designs, as they allowed the modelling of important sources of random

noise, such as participants and songs’ variability. Secondly, several studies applied machine

learning and data mining techniques that proved to be well-suited to examine certain music

problems. For example, to analyse a large set of individual differences, S3 (see section 2.2.1

- Appendix C) and S4 (see section 2.2.2 - Appendix D) showed that random forests can be

a very useful technique, as they can handle a large number of variables even when they are

correlated between themselves. Similarly, classification tree models were successfully

applied in S3 (see section 2.2.1 - Appendix C) and S10 (see section 2.3.4 - Appendix J) to

model higher-order interactions between several predictor variables and the dependent

variable of the study. In S10, for instance, this approach allowed for the identification of

three main inflection points in which the prevalence of female artists in UK’s popular music

3.3 Future directions

changed considerably over time, i.e., 1968, 1976, 1984. Interestingly, these inflection points

coincide with some significant moments in UK’s culture, such as the surge in popularity of

the women’s rights movements (1968), the rise of punk (1976), and the peak in popularity

of Margaret Thatcher’s prime ministership (1984).

Finally, the last part of this thesis focused on the application of the BEM to improve music-

related decision making in the real world. Two studies focused on music decision making in

advertising and marketing, an area where music choices are particularly relevant. Musical

choices can have profound effects on brand communications, consumer behaviour, and can

be costly for brands. Despite this, the process of choosing and evaluating music for

advertising is poorly understood. Thus, S7 and S8 applied insights from behavioural

economics to better inform the decision-making process of selecting music for ads. S7 (see

section 2.3.1 - Appendix G) found that source bias has a significant impact on how ad

professionals evaluate music for advertising purposes, whereas source cues have no effect

on consumers. The differences between these two groups can be costly for brands, as

professionals may recommend that their clients pay a premium for music coming from

specific sources (e.g., performing artists), but brands may see little or no added benefit if

ultimately, the source of the music does not matter to the consumer. Potential solutions to

mitigate this bias are discussed, including increasing awareness among professionals and

measuring the impact of advertising music and source on target consumers. In addition, S8

(see section 2.3.2 - Appendix H) provided a first estimation of the effectiveness of using

music as a recognition cue to influence consumer choice by means of the recognition

heuristic. The results showed that music can only be successfully used as a recognition cue

when it is liked by the target consumers, whereas recognition-based heuristics are not

influential when the music is disliked. This finding is valuable to brands in terms of the

importance of measuring the value of their investment when working with music.

3.3 Future directions

The diversity of the ten studies presented in Chapter 2 begin to illustrate the breadth and

potential of the role that behavioural economics can play in music research. Moreover, a

network visualisation map of behavioural economics demonstrates the rich array of

concepts and theories yet to be applied within the domain of music research (see Figure

3.12). The map was generated using 68,509 publications from the field of behavioural

economics and shows the 200 most frequently used concepts in these publications (the key

terms provided by the author (s) to describe their work). Thus, the map highlights the

potential of behavioural economics for future research on music decision making. Based on

this map and further accumulated insights from the ten empirical investigations summarised

in Chapter 2, this section proposes ideas and directions for future research within the BEM

(RQ 5). The future directions are organised around different areas within music research

that can significantly benefit from considering the behavioural economics toolkit.

Fig. 3.1 Network visualization map of behavioural economics.

The map shows the 200 most influential keywords in the behavioural economics literature (based on 68,509 publications) and how often they co-occur with

others, identifying main research areas and concepts in the field. On the map, each concept is represented by a circle, its size determining how frequently the

concept was used in the retrieved literature. The lines between concepts indicate how connected the concepts are with each other. The stronger the connection,

the wider the line. Highly connected concepts are grouped into clusters, with different colours representing different clusters.

3.3 Future directions

Music reward value

Music reward value is at the core of musical behaviour. Yet, if musical sounds have not

intrinsic reward for humans, why are they perceived so pleasurably by the human brain?

This issue has puzzled scientists and philosophers over the years, but recent advances in

neuro- science have made great progress in identifying multiple mechanisms underlying

reward in music (see Zatorre & Zald, 2011, for a review). Here, there is great potential for

Neuroeconomics, a fast-growing field in behavioural economics that aims to explain how

human decision making happens inside the brain (see Dhami, 2016; Glimcher & Fehr, 2014,

for reviews). By using neuroscientific methods, neuroeconomics both tests and develops

theories of human decision making while using algorithmic models and experimental

paradigms that are particularly suited to study human behaviour (Dhami, 2016; Glimcher &

Fehr, 2014). Therefore, neuroeconomics may be useful in identifying the brain regions that

allow humans to derive pleasure from music. Using this approach, Salimpoor et al., 2013

examined the neural processes when music gains reward value the first time it is heard.

While undergoing fMRI scanning, participants listened to previously unheard music

excerpts and indicated how much they were willing to spend on them using an auction

paradigm. The results showed that music reward arises from the interaction between

mesolimbic reward circuitry – a pathway in the brain associated with the dopaminergic

system and reward – and sensory cortices involved in auditory processing, which impacts

behavioural decisions about the value of music. Importantly, the benefits of neuroeconomics

for music research are not just limited to the study of reward in music. Neuroeconomics can

be useful to test and develop theories of decision making in other music areas as well, such

as music composition and improvisation choices, music preferences and listening behaviour,

and the use of music in advertising.

Music preferences and listening behaviour

Why do people listen to (and pay) for the music they do? And how do these decisions

influence the music industry? To address these questions, this thesis encourages future

research to consider the behavioural economics literature on cognitive biases and heuristics

and dual-process theory.

Several studies reported in this thesis show that listeners are boundedly rational and,

consequently, rely on cognitive biases and heuristics when evaluating and choosing music

(see section 2.2). However, we are far from having a clear picture of the role that heuristic

processing plays in music listening and decision making. In addition to the heuristics and

biases addressed in this thesis, researchers have proposed many others that have yet not been

applied to the music domain. This list is extensive, for example, the Decision Lab (Montreal,

Canada) identifies more than 80 heuristics and biases in human decision making

(https://thedecisionlab.com/biases). Thus, the scientific potential for music research here is

immense. As an example, one can consider how some of these unexplored heuristics and

biases may help better understand music listening behaviour in the current digital era. The

3.3 Future directions

drastic increase in music streaming services observed in recent years, such as Youtube and

Spotify, provides listeners with millions of tunes almost instantly (IFPI, 2019). Yet, it is

unclear how listeners search for and choose music in this seemingly endless range of music

choices. Here, it is worth considering research on choice overload, a cognitive phenomenon

that occurs as a result of too many choices being available to consumers, resulting in

negative outcomes, such as decision fatigue, choosing the default option, or having an

unpleasant experience (see Chernev, Böckenholt, & Goodman, 2012, for a review). Future

research on choice overload could determine how the size of music choice sets or playlist

affect people when listening to music in streaming services, helping identify the optimal

number of items to maximize the listening experience. Alternatively, mental accounting

refers to the tendency to treat one’s money or goods differently based on subjective criteria,

such as its intended use or its source (Thaler, 1985). Future studies on mental accounting

could examine differences in people’s value for music when the music is in digital format

versus physical format (e.g., CD or vinyl). Similarly, mental accounting could explain why

people easily engage in music piracy, whereas they would not engage in similar behaviour

when the format of the music is physical, such as stealing music CDs from a store.

Furthermore, there is great potential for research on dual-process theory in the context of

music listening and preferences. Examining the interactions between emotional (System 1)

and cognitive (System 2) processes in the brain can shed light on how music affects

individuals in a particular moment in time, providing a window into why people like the

music they do. Previous research suggests that people choose music depending on the

cognitive involvement of the two systems of processing (Konecni, 1982; Konecni &

Sargent- Pollock, 1976; North & Hargreaves, 2000, 1999). For example, participants were

less likely to select complex music when performing a cognitively-demanding task than

when performing a less demanding task (Konecni & Sargent-Pollock, 1976). Future studies

could examine this issue further by using a wider selection of music stimuli and cognitively

demanding tasks, as well as studying how they interact with listeners’ individual differences.

These insights could be used to improve current models of music preference and attempt to

predict listeners’ music choices in a given point in time based on the involvement of the two

systems of processing.

3.4 Conclusion

Music consumption and the market

The behavioural economics toolkit offers highly valuable insights into understanding

consumer behaviour and the market. This thesis emphasizes the potential of behavioural

pricing for future research exploring music consumption and how listener choices shape the

music industry. Behavioural pricing is an area in behavioural economics that examines how

consumers react to prices incorporating a psychological perspective (see Koschate-Fischer

& Wüllner, 2017, for a review). By investigating the intra-personal processes of price

perception, evaluation, and memory, researchers can better understand consumer behaviour

and its consequences to the market. With much of music content now online and the

continual threat of piracy, artists, labels, and music distributors have had to become more

innovative to develop new strategies for generating revenue from digital music content.

Thus, behavioural pricing can be useful to understand consumers’ reactions to pricing

decisions made by artists and music firms. For instance, future studies could focus on

understanding why consumers choose to pay voluntarily for music and which are the

advantages of this payment model compared to traditional fixed-price strategies. More

research is needed here looking specifically at the motivations behind such behaviour.

Similarly, future studies could examine the feasibility of other voluntary payment models,

such as giving the music completely for free or using crowdsourcing methods to raise funds

from fans. Other pricing decisions in music that could largely benefit from considering

behavioural pricing include the relationship between artists’ pricing decisions and the live

music industry. For example, future studies could investigate whether pricing strategies

could promote attendance of young audiences to classical music concerts, an ongoing

concern among classical music organizations (e.g., Dimaggio & Mukhtar, 2004; Kolb,

2000). Alternatively, future research could examine whether pricing strategies can

ameliorate the value gap created between artists and music streaming services – i.e., the

growing mismatch between consumed music in these services and the revenue returned to

musicians and the music community, which is proportionally very low (IFPI, 2019).

Music composition and improvisation

Given the highly demanding nature of music improvisation, including fast timing of note

events (often in the range of 40 milliseconds or less) and cognitive as well as bodily

limitations (Impett, 2016), it seems plausible that musicians strongly rely on fast and frugal

heuristics to make decisions while improvising. This thesis emphasizes the potential of

computational musicology to analyse large corpora of music performances and quantify the

role that different heuristics may play in music composition and improvisation. Using such

approach, a recent study (Beaty, Frieler, Norgaard, & Merseal, 2020) analysed a corpus of

hundreds of improvised solos from eminent jazz musicians. By extracting all melodic

sequences in the corpus and calculating relevant metrics (e.g., pitch variety, pattern

frequency), the authors quantified the level of complexity in each music sequence in the

corpus. The results consistently showed that expert jazz musicians tend to start their

3.4 Conclusion

improvisations with music sequences that are significantly easier than subsequent sequences

in their solos, where “easy” was defined as statistically more frequent and less melodically

complex. This finding is in line with the availability heuristic (Tversky & Kahneman, 1973),

as the less complex a music sequence is, the more available in memory and easier to

remember. This thesis encourages future studies to use similar methods to identify other

core heuristics underlying music composition and improvisation. For instance, the anchoring

heuristic (Tversky & Kahneman, 1974) refers to the human tendency to rely heavily on the

first piece of information offered to “anchor” subsequent judgments and interpretations.

Musicians may rely on this heuristic to simplify decision making while improvising,

adjusting their choices to the anchor (the first piece of music played or heard). By analysing

large corpora of improvised solos, one could examine the role of the anchoring heuristic by

comparing the melodic similarity of the first music sequence in a solo to subsequent

sequences.

Music performance evaluation

Further research on the BEM could focus on improving the decision making process

inherent to music performance evaluation, both in terms of identifying core cognitive biases

and heuristics and tailoring evaluative strategies to mitigate their effects. As an example,

consider how this process could work to improve jurors’ decision making in a well-known

musical competition such as the Queen Elisabeth Musical Competition. Firstly, using data

from previous years, one could systematically investigate whether juror’s judgments rely on

cognitive biases and heuristics. For instance, Flôres and Ginsburgh (1996) analysed the

rankings of the 12 semi-finalists in this competition over 21 years and found a clear order

bias in jurors’ decisions: those semi-finalists who performed first had a lower chance to win

the competition, whereas those who performed later had a higher chance. Order bias can be

understood within the context of bounded rationality, where jurors are simply not able to

remember equally well all performances and, consequently, their judgments rely on the

availability effects (Tversky & Kahneman, 1973), giving higher scores to performances they

remember better (those that performed more recently). Secondly, once the impact of

cognitive biases and heuristics is measured, the second step should focus on tailoring

effective strategies to mitigate their effects on the evaluative process. For example, making

jurors aware of such biases is one step towards mitigating their effects, as there is evidence

that awareness of bias can bring about change (Pope, Price, & Wolfers, 2013). Other

interventions could consist of using assessment methods that are less susceptible to

cognitive biases, such as providing moment-to-moment evaluations or using blind and

randomised procedures.

Music education

Research on music education has used a wide variety of psychological approaches to

understand what motivates people to take and persists with music lessons and practice (see

3.4 Conclusion

Austin, Renwick, & McPherson, 2006; Renwick & Reeve, 2012, for reviews). However, it

remains unclear why some individuals manage to persist through the challenges of learning

and practicing, whereas others eventually quit. Research on behavioural time preferences

can offer valuable insights to this issue. For example, research on delay discounting provides

a unifying theoretical approach that that has been successfully applied to a number of similar

issues in psychology, including those related to health, self-control, impulsivity, and risk

taking (see Green & Myerson, 2004; Koffarnus, Jarmolowicz, Mueller, & Bickel, 2014;

Peters & Büchel, 2011; Reynlods, 2006, for reviews). In the context of music education,

delay discounting predicts that in impatient students, the short-term temptation (e.g., quit-

ting or not practising enough) foregoes the long-term goal (e.g., learning how to play an

instrument). More importantly, the relationship between time and subjective reward can be

modelled accurately using mathematical functions with few parameters, such as discount

rate (how fast an individual’s subjective value decreases over time), which is associated with

impulsivity and impatience (see Peters & Büchel, 2011, for a review). Thus, there is great

potential for future research looking at whether the rate of delay discounting can be a reliable

trait marker for music practice and learning. By incorporating such insights, one could help

improve current educational methods in music, decreasing dropout rates and increasing the

learning experience. For instance, there is evidence that framing effects and episodic future

thinking can reduce significantly delay discounting (Koffarnus et al., 2014), as the better

that individuals can imagine future outcomes, the more they value them.

3.4 Conclusion

Music psychology has examined various aspects of decision making related to musical

behaviour using a wide variety of methods and techniques (see Deutsch, 2013; Hallam et

al., 2016; Tan et al., 2017, for reviews). This body of research, however, would benefit from

using a more sophisticated and unified framework dedicated exclusively on the study of

music decision making, as well as incorporating insights from social sciences and

economics. In contrast, economists have investigated music-related decision making with a

focus on rational economic analysis instead of the psychological underpinnings known to be

involved in music perception, cognition, and behaviour (see Byun, 2016; Cameron, 2016;

Krueger, 2005; Tschmuck, 2017, for reviews). To bridge this gap, this thesis proposes the

BEM, a novel research framework that integrates knowledge from psychology, economics,

and other disciplines to increase our understanding of human behaviours related to music.

Ten scientific publications (see Table 2.1 for a list of publications; see Appendix A-J for the

full texts) were conducted to demonstrate the value of this novel approach, generating new

insights into the study of music decision making.

The BEM has both theoretical and practical implications. First, it offers a multidisciplinary

but unified framework to understand and improve music decision making in a variety of

areas. Second, it merges two distinct bodies of research that have been largely unconnected

3.4 Conclusion

in the literature thus far. In particular, the BEM moves away from the rigid neoclassical

assumption of rationality by incorporating insights from psychology, while still relying on

falsifiable models from standard economics that are mathematically rigorous and can prove

significant to address key problems in music research. Third, the BEM offers a solid

understanding of those areas in behavioural economics that readily apply to music decision

making, but also provides valuable directions for future research aiming to explore new

areas, such as neuroeconomics, behavioural pricing, and behavioural game theory. These

avenues for future research can improve our current knowledge in many areas within music

psychology, including music composition and improvisation, performance evaluation,

music preferences and consumption, and music education. Finally, since music is a potent

and highly emotional stimulus, investigating music decision making can provide a novel

testing ground for general theories on human behaviour and decision making. Thus, the

BEM can be used as a toolkit to generate new ideas and accelerate progress in any area

concerned with human behaviours related to music.

References

Allan, D. (2007). Sound Advertising: A Review of the Experimental Evidence on the Effects

of Music in Commercials on Attention, Memory, Attitudes, and Purchase Intention.

Journal of Media Psychology, 12(3), 1–37.

Anglada-Tort, M., & Müllensiefen, D. (2017). The repeated recording illusion: the effects of

extrinsic and individual difference factors on musical judgments. Music Perception:

An Interdisciplinary Journal, 35(1), 94–117. doi:10.1525/mp.2017.35.1.94

Angner, E. (2012). A course in behavioral economics. Macmillan International Higher

Education.

Ariely, D. (2008). Predictably irrational. Harper Audio.

Arne Brekke, K., & Johansson-Stenman, O. (2008). The behavioural economics of climate

change. Oxford Review of Economic Policy, 24(2), 280–297. doi:10.1093/oxrep/grn012

Ashley, R. (2016). Musical Improvisation. In S. Hallam, I. Cross, & M. Thaut (Eds.), The

oxford handbook of music psychology. Oxford University Press.

Austin, J. R., Renwick, J. M., & McPherson, G. E. (2006). Developing motivation. In

G. E. McPherson (Ed.), The child as musician: A handbook of musical development

(pp. 213–238). Oxford University Press.

Beaty, R., Frieler, K., Norgaard, M., & Merseal, H. (2020). Spontaneous melodic productions

expert musicians contain sequencing biases seen in language production. Retrieved

from https://osf.io/p8gmj/download

Blumenthal-Barby, J. S., & Krieger, H. (2015). Cognitive biases and heuristics in medical

decision making: A critical review using a systematic search strategy. Medical Decision

Making, 35(4), 539–557. doi:10.1177/0272989X14547740

Bradlow, E. T., & Fader, P. S. (2001). A bayesian lifetime model for the “hot 100” billboard

songs. Journal of the American Statistical Association, 96(454), 368–381. doi:10.1198/

016214501753168091

Burke, A. E. (1996). How effective are international copyright conventions in the music

industry? Journal of Cultural Economics, 20(1), 51–66. doi:10.1007/s10824-005-

1060-z

Byun, C. H. C. (2016). The economics of the popular music industry. Springer.

Cameron, S. (2015). Music in the marketplace: a social economics approach. doi:10.5860/

choice.192595

Cameron, S. (2016). Past, present and future: music economics at the crossroads. Journal of

Cultural Economics, 40(1), 1–12. doi:10.1007/s10824-015-9263-4

Cartwright, E. (2018). Behavioral Economics. Routledge.

Chernev, A., Böckenholt, U., & Goodman, J. (2012). Choice overload: A conceptual review

and meta-analysis. 25(2), 333–358. doi:10.1016/j.jcps.2014.08.002

Coleman, J., & Fararo, T. (1992). Rational choice theory. Sage Publications.

References

Dawes, R., & Hastie, R. (2010). Rational choice in an uncertain world: The psychology of

judgment and decision making. Sage Publications.

Decrop, A., & Derbaix, M. (2014). Artist-Related Determinants of Music Concert Prices.

Psychology and Marketing, 31(8), 660–669. doi:10.1002/mar.20726

Dellavigna, S. (2009). Psychology and economics: Evidence from the field. Journal of

Economic Literature, 47(2), 315–372. doi:10.1257/jel.47.2.315

Deutsch, D. (2013). Psychology of music. Elsevier.

Dhami, S. (2016). The foundations of behavioral economic analysis. Oxford University

Press.

Dimaggio, P., & Mukhtar, T. (2004). Arts participation as cultural capital in the United States,

1982-2002: Signs of decline? Poetics, 32, 169–194. doi:10.1016/j.poetic.2004.02.005

Eerola, T., & Vuoskoski, J. K. (2013). A review of music and emotion studies: Approaches,

emotion models, and stimuli. Music Perception, 30(3), 307–340. doi:10.1525/MP.2012.

30.3.307

Elliott, C. (1995). Race and Gender as Factors in Judgments of Musical Performance. Bulletin

of the Council for Research in Music Education, (127), 50–56.

Elliott, C., & Simmons, R. (2011). Factors determining UK album success. Applied Eco-

nomics, 43(30), 4699–4705. doi:10.1080/00036846.2010.498349

Evans, J. S. B. T. (2008). Dual-Processing Accounts of Reasoning, Judgment, and Social

Cognition. Annual Review of Psychology, 59(1), 255–278. doi:10.1146/annurev.psych.

59.103006.093629

Fehr, E., & Rangel, A. (2011). Neuroeconomic Foundations of Economic Choice—Recent

Advances. Journal of Economic Perspectives, 25(4), 3–30. doi:10.1257/jep.25.4.3

Flôres, R. G., & Ginsburgh, V. A. (1996). The Queen Elisabeth Musical Competition How

fair is the final ranking. Journal of the Royal Statistical Society, 45, 97–104.

Frederiks, E. R., Stenner, K., & Hobman, E. V. (2015). Household energy use: Applying

behavioural economics to understand consumer decision-making and behaviour. 41,

1385–1394. doi:10.1016/j.rser.2014.09.026

Glimcher, P. W., & Fehr, E. (2014). Neuroeconomics: Decision making and the brain.

doi:10.1016/B978-0-12-416008-8.00003-6

Goldstein, D. G., & Gigerenzer, G. (2002). Models of ecological rationality: The recognition

heuristic. Psychological Review, 109(1), 75–90. doi:10.1037/0033-295x.109.1.75

Greb, F., Schlotz, W., & Steffens, J. (2018). Personal and situational influences on the

functions of music listening. Psychology of Music, 46(6), 763–794. doi:10.1177/

0305735617724883

Griffiths, N. K. (2008). The effects of concert dress and physical appearance on perceptions

of female solo performers. Musicae Scientiae, 2, 273–290. doi:https://doi.org/10.1177/

102986490801200205

Hallam, S., Cross, I., & Thaut, M. (2016). Oxford handbook of music psychology (second

edition). Oxford University Press.

Hendricks, K., & Sorensen, A. (2009). Information and the Skewness of Music Sales. Journal

Political Economy, 117(2), 324–369. doi:10.1086/599283

Hiller, R. S. (2016). The importance of quality: How music festivals achieved commercial

success. Journal of Cultural Economics, 40(3), 309–334. doi:10.1007/s10824-015-

9249-2

Holt, F. (2010). The economy of live music in the digital age. European Journal of Cultural

Studies, 13(2), 243–261. doi:10.1177/1367549409352277

References

IFPI. (2019). International Federation of the Phonographic Industry (IFPI) Global Music

Report 2018. Retrieved from http://www.ifpi.org/downloads/GMR2019.pdf

Impett, J. (2016). Making a mark: the psychology of composition. In S. Hallam, I. Cross, &

M. Thaut (Eds.), The oxford handbook of music psychology. Oxford University Press.

Jabbar, H. (2011). he behavioral economics of education: New directions for research.

Educational Researcher, 40(9), 446–453.

Kahneman, D. (2003). Maps of bounded rationality: Psychology for behavioral economics.

American Economic Review, 93(5), 1449–1475.

Kahneman, D. (2011). Thinking , Fast and Slow. Macmillan.

Kolb, B. M. (2000). You call this fun? Reactions of young first-time attendees to a classical

concert. MEIEA Journal, 1(2), 13–29.

Konecˇni, V. (1982). Social interaction and musical preference. In D. Deutsch (Ed.), The

psychology of music (pp. 497–516). Academic Press.

Konecni, V. J., & Sargent-Pollock, D. (1976). Choice between melodies differing in complex-

ity under divided-attention conditions. Journal of Experimental Psychology: Human

Perception and Performance, 2(3), 347–356. doi:10.1037/0096-1523.2.3.347

Koschate-Fischer, N., & Wüllner, K. (2017). New developments in behavioral pricing re-

search. Journal of Business Economics, 87(6), 809–875. doi:10.1007/s11573-016-

0839-z

Krueger, A. B. (2005). The economics of real superstars: The market for rock concerts in the

material world. Journal of Labor Economics, 23(1), 1–30. doi:10.1086/425431

Lamont, A., & Greasley, A. (2016). Musical preferences. In S. Hallam, I. Cross, & M. Thaut

(Eds.), The oxford handbook of music psychology. Oxford University Press.

Lamont, A., Greasley, A., & Sloboda, J. (2016). Choosing to Hear Music. In S. Hallam,

I. Cross, & M. Thaut (Eds.), The oxford handbook of music psychology (pp. 1–20).

Oxford University Press.

Lantos, G. P., & Craton, L. G. (2012). A model of consumer response to advertising music.

Journal of Consumer Marketing, 29(1), 22–42. doi:10.1108/07363761211193028

Larsen, G., & Hussels, S. (2011). The significance of commercial music festivals. In S.

Cameron (Ed.), Handbook on the economics of leisure (pp. 250–270). doi:10.4337/

9780857930569.00022

Liebowitz, S. J. (2004). Will MP3 downloads annihilate the record industry? The evidence

so far. Advances in the Study of Entrepreneurship, Innovation, and Economic Growth,

15, 229–260. doi:10.1016/S1048-4736(04)01507-3

Liebowitz, S. J. (2006). File sharing: Creative destruction or just plain destruction? Journal

of Law and Economics, 49(1), 1–28. doi:10.1086/503518

Linnemann, A., Wenzel, M., Grammes, J., Kubiak, T., & Nater, U. M. (2018). Music listening

and stress in daily life—a matter of timing. International Journal of Behavioral

Medicine, 25(2), 223–230. doi:10.1007/s12529-017-9697-5

Lonsdale, A. J., & North, A. C. (2012). Musical taste and the representativeness heuristic.

Psychology of Music, 40(2), 131–142. doi:10.1177/0305735611425901

Maeshiro, T., Nakayama, S. I., & Maeshiro, M. (2011). Representation of decision making

process in music composition based on hypernetwork model. In Lecture notes in

computer science (including subseries lecture notes in artificial intelligence and lecture

notes in bioinformatics) (Vol. 6771 LNCS, pp. 109–117). doi:10.1007/978-3-642-

21793-7_13

Maymin, P. (2012). Music and the market: Song and stock volatility. North American Journal

of Economics and Finance, 23(1), 70–85. doi:10.1016/j.najef.2011.11.004

References

Mortimer, J. H., Nosko, C., & Sorensen, A. (2012). Supply responses to digital distribution:

Recorded music and live performances. Information Economics and Policy, 24(1), 3–

14.

North, A. C., & Hargreaves, D. J. (1999). Music and driving performance. Scandinavian

Journal of Psychology, 40, 285–292.

North, A. C., & Hargreaves, D. J. (2000). Musical preferences during and after relaxation

and exercise. The American Journal of Psychology, 113(1).

North, A. C., & Hargreaves, D. J. (2008). The social and applied psychology of music. OUP

Oxford.

Oakes, S. (2007). Evaluating Empirical Research into Music in Advertising: A Congruity

Perspective. Article in Journal of Advertising Research, 47(1), 38–50. doi:10.2501/

S0021849907070055

Oberholzer-Gee, F., & Strumpf, K. (2009). File sharing and copyright. Innovation Policy and

the Economy, 10, 19–55. doi:10.1086/605852

Oeusoonthornwattana, O., & Shanks, D. R. (2010). I like what I know: Is recognition a non-

compensatory determiner of consumer choice? Judgment and Decision Making, 5(4),

310–325.

Peters, J., & Büchel, C. (2011). The neural mechanisms of inter-temporal decision-making:

understanding variability. Trends in cognitive sciences, 15(5), 20–37.

Pettijohn, T. F., Eastman, J. T., & Richard, K. G. (2012). And the Beat Goes On: Popular

Billboard Song Beats Per Minute and Key Signatures Vary with Social and Economic

Conditions. Current Psychology, 31, 313–317. doi:10.1007/s12144-012-9149-y

Pettijohn, T. F., & Sacco, D. F. (2009). The Language of lyrics An analysis of popular

Billboard songs across conditions of social and economic threat. Journal of Language

and Social Psychology, 28, 297–311. doi:10.1177/0261927X09335259

Pope, D. G., Price, J., & Wolfers, J. (2013). Awareness Reduces Racial Bias. Management

Science, 64(11), 4988–4995.

Rayna, T., & Striukova, L. (2009). Monometapoly or the Economics of the Music Industry.

Prometheus, 27(3), 211–222. doi:10.1080/08109020903127778

Reber, R., Schwarz, N., & Winkielman, P. (2004). Processing fluency and aesthetic pleasure:

Is beauty in the perceiver’s processing experience? Personality and Social Psychology

Review, 8(4), 364–382. doi:10.1207/s15327957pspr0804_3

Renwick, J. M., & Reeve, J. (2012). Supporting motivation in music education. In G. E.

McPherson & G. F. Welch (Eds.), Oxford handbook of music education (pp. 143–162).

Oxford University Press.

Rice, T. (2013). The Behavioral Economics of Health and Health Care. Annual Review of

Public Health, 34(1), 431–447. doi:10.1146/annurev-publhealth-031912-114353

Saarikallio, S., & Erkkilä, J. (2007). The role of music in adolescents’ mood regulation.

Psychology of Music, 35(1), 88–109. doi:10.1177/0305735607068889

Salimpoor, V. N., van den Bosch, I., Kovacevic, N., McIntosh, A. R., Dagher, A., & Zatorre,

R. J. (2013). Interactions between the nucleus accumbens and auditory cortices predict

music reward value. Science, 340(6129), 216–219. doi:10.1126/science.1231059.

arXiv: arXiv:1011.1669v3

Samson, A. (2017). The Behavioral Economics Guide 2017. Behavioral Science Solutions

Ltd.

Schramm, H., & Spangardt, B. (2016). Wirkung von Musik in der Werbung [Effects of music

in advertising]. In Handbuch werbeforschung (Siegert, G, pp. 433–449). Wiesbaden,

Germany: Springer VS.

References

Shah, A. K., & Oppenheimer, D. M. (2007). Easy does it: The role of fluency in cue weighting.

Judgment and Decision Making, 2(6), 371–379.

Simon, H. A. (1955). A Behavioral Model of Rational Choice. The Quarterly Journal of

Economics, 69(1), 99–118. doi:10.2307/1884852

Simon, H. A. (1982). Models of Bounded Rationality. MIT press.

Slovic, P., Finucane, M., Peters, E., & MacGregor, D. G. (2002). Rational actors or rational

fools: Implications of the effects heuristic for behavioral economics. Journal of Socio-

Economics, 31(4), 329–342. doi:10.1016/S1053-5357(02)00174-9

Stevans, L. K., & Sessions, D. N. (2005). An empirical investigation into the effect of music

downloading on the consumer expenditure of recorded music: A time series approach.

Journal of Consumer Policy, 28(3), 311–324. doi:10.1007/s10603-005-8645-y

Strobl, E. A., & Tucker, C. (2000). The dynamics of chart success in the U.K. pre-recorded

popular music industry. Journal of Cultural Economics, 24(2), 113–134. doi:10.1023/

A:1007601402245

Sweeting, A. (2013). Dynamic Product Positioning in Differentiated Product Markets: The

Effect of Fees for Musical Performance Rights on the Commercial Radio Industry.

Econometrica, 81(5), 1763–1803. doi:10.3982/ECTA7473

Tahler, R. (2015). Misbehaving: The Making of Behavioral Economics. WW Norton.

Tan, S., Pfordresher, P., & Harré, R. (2017). Psychology of music: From sound to significance.

Routledge.

Thaler, R. (1985). Mental Accounting and Consumer Choice. Marketing Science, 4(3), 199–

214. doi:10.1287/mksc.4.3.199

Tschmuck, P. (2017). The economics of music. Agenda Publishing.

Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and

probability. Cognitive Psychology, 5(2), 207–232. doi:10.1016/0010-0285(73)90033-9

Tversky, A., & Kahneman, D. (1974). Judgment under Uncertainty: Heuristics and Biases.

Science, 185(4157), 1124–31. doi:10.1126/science.185.4157.1124

Varian, H. R. (2005). Copying and copyright. Journal of Economic Perspectives, 19(2),

121–138. doi:10.1257/0895330054048768

Waddell, G. (2018). Time to decide: A study of evaluative decision-making in music perfom-

rance (Doctoral dissertation, Royal College of Music).

Wöllner, C., & Behne, K.-E. (2011). Seeing or hearing the pianists? A synopsis of an early

audiovisual perception experiment and a replication. Musicae Scientiae, 15(3), 324–

342. doi:10.1177/1029864911410955

Zatorre, R. J., & Zald, D. H. (2011). Music. In J. A. Gottfried (Ed.), Neurobiology of sensation

and reward. CRC Press.

Zullow, H. M. (1991). Pessimistic rumination in popular songs and news magazines pre-

dict economic recession via decreased consumer optimism and spending. Journal of

Economic Psychology, 12(3), 501–526.

Appendix A

Visualizing music psychology (S1)

This is an Accepted Manuscript of an article published by SAGE in Music & Science on

25th of January 2019. ©The Author(s) 2019. Reprinted by permission of SAGE

Publications

, and available online: https://doi.org/10.1177/2059204318811786. The paper

is not the copy of the record and may not exactly replicate the authoritative document

published in the journal. For presentation in this thesis, the appendices of the paper have

been removed and the passages referring to each Appendix in the text modified to indicate

where to find the materials online. Moreover, there may be minor modifications in the text

to guarantee a consistent typographic style throughout the thesis, such as the position of

figures and tables. Please do not copy or cite without author’s permission.

Citation

Anglada-Tort, M., & Sanfilippo, K. R. M. (2019). Visualizing Music Psychology: A

Bibliometric Analysis of Psychology of Music, Music Perception, and Musicae Scientiae

from 1973 to 2017. Music & Science, 2, 2059204318811786. DOI:

https://doi.org/10.1177/2059204318811786.

Author contribution

The paper was written together with Dr. Sanfilippo (Goldsmiths, University of London). I

conceived of the idea and the analysis strategy for the study, whereas all other aspects were

done collaboratively.

The paper is deposited under the rems of the Creative Commons Non Commercial CC BY-NC: This article

is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 License

(http://www.creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial use, reproduction and

distribution of the work without further permission provided the original work is attributed as specified on the

SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).

Visualizing music psychology: A bibliometric analysis of

Psychology of Music, Music Perception, and Musicae

Scientiae from 1973 to 2017

Music psychology has grown drastically since being established in the middle of the 19th

century. However, up to this date, no large-scale computational bibliometric analysis of the

scientific literature in music psychology has been carried out. This study aims to analyze all

published literature from the journals Psychology of Music, Music Perception, and Musicae

Scientiae. The retrieved literature comprised a total of 2,089 peer-reviewed articles, 2,632

authors, and 49 countries.

Visualization and bibliometric techniques were used to investigate

the

growth of publications, citation analysis, author and country productivity, collaborations, and

research trends. From 1973 to 2017, with a total growth rate of 11%, there is a clear increase

in music psychology research (i.e. number of publications, authors, and collabo- rations),

consistent with the general growth observed in science. The retrieved documents received a

total of 33,771 citations (M = 16.17, SD = 26.93), with a median (Q1 – Q3) of 7 (2 – 20).

Different bibliometric indicators defined the most relevant authors, countries, and keywords

as well as how they relate and collaborate with each other. Differences between the three

journals are also studied. This type of analysis, not without its limitations, can help understand

music psychology and identify future directions within the field.

Keywords: music psychology; psychology of music; bibliometrics; scientometrics; visu-

alization technique.

A.1 Introduction

The beginnings of what we now regard as music psychology started in the middle of the 19th

century as a branch of both psychology and musicology (Thaut, 2016). But music psychology

has evolved and grown drastically since then. From a focus on psychoacoustics, perception,

and the cognitive sciences, to health applications and the use of music in everyday life, music

psychology has shifted and blossomed, establishing programs, labs and journals covering

different research interests, geographical areas, and research groups.

Music Psychology can be defined as the scientific study of the psychological processes

through which music is perceived, created, responded to, and incorporated to everyday life

(Tan, Pfordresher, & Harré, 2017; Thompson, 2014). The field of music psychology

therefore embraces an incredibly diverse and wide variety of topics, including the origins of

music, music perception and cognition, responses to music (e.g. bodily, emotional, and

aesthetical), the neuroscience of music, music development, music education, music

performance, composition and improvisation, the use of music in everyday life, and music

therapy and wellbeing (Hallam, Cross, & Thaut, 2016). But the psychology of music can

also contribute to many other fields, including musical theory, ethnomusicology, computer

science, aesthetics, health sciences, marketing and advertising. Researchers from all over the

globe investigate these topics empirically, with more than 80 music cognition and science

labs around the world (www.musicperception.org/smpc-resources.html). Various music

psychology specific conference series have also begun to develop such as the International

Conference on Music Perception and Cognition (ICMPC), founded in 1989, and the European

Society for the Cognitive Sciences of Music (ESCOM), founded in 1991. In 2008, the

International Conference of Students of Systematic Musicology (SysMus) was founded for

students of systematic musicology, a broader field which encompasses music psychology.

The first research journal specifically dedicated to music psychology is Psychology of Music

established in 1973. This multidisciplinary journal’s aim is to, "increase scientific understand-

ing of all psychological aspects of music and music education". Music Perception, estab-

lished in 1983, was developed with a primary focus on cognitive-psychological research with

broader and multidisciplinary draw, including work from “psychology, psychophysics, neuro-

science, music theory, acoustics, artificial intelligence, linguistics, philosophy, anthropology

and

cognitive science” (mp.ucpress.edu). In 1997, the European Society for the Cognitive

Sciences of Music (ESCOM) was developed along with its journal Musicae Scientiae, which

aims to include “empirical, theoretical and critical articles directed at increasing understand-

ing of how music is perceived, represented and generated” (journals.sagepub.com/home/msx).

As a truly multidisciplinary subject, music psychology research is published in many other

A.1 Introduction

journals, including other APA journals and journals from related disciplines, such as musicol-

ogy, music theory, music therapy, music education, aesthetics, marketing, and neuroscience.

This includes, for example, the Journal of Research in Music Education, International Jour-

nal of Music Education, Journal of Music Therapy, Empirical Musicology Review, and

Psychomusicology: Music, Mind, and Brain

The current research focuses on the three most prominent scientific journals in music psy-

chology. Namely, Psychology of Music, Music Perception, and Musicae Scientiae. We used

two

criteria to select these journals: content and impact. Regarding content, the focus was in

journals covering specifically the psychology of music. Impact was determined by the SJR

ranking provided in SCImago (https://www.scimagojr.com/). This measure indicates the

average number of weighted citations per document received within the selected journal

during the previous three years. In June 2018, searching in the category “music”, Psychology

of Music was ranked the fourth, Musicae Scientiae the sixth, and Music Perception the

seventh. The first (i.e. IEE Signal Processing Magazine), second (i.e., Journal of Research in

Music Education), third (i.e., Music Education Research), and fifth (i.e., International Journal

of Music Education) journals did not meet the first criterion of content, focusing on other

topics rather than music psychology (i.e. signal processing or music education).

With the surge of interest in music psychology research, it is important as a discipline

to reflect objectively and systematically on what research has been published and what

gaps can still be filled. Bibliometrics and scientometrics allow for the measurement and

analysis of published scientific literature, giving objective and measureable data to help us

understand the discipline’s trajectory thus far. By using computational, mathematical and

statistical techniques, bibliometrics analyses the quantity and quality of published scientific

literature, including citation analysis, authorship and country productivity and collaborations,

impact of publications, and research trends (e.g., Blázquez-Ruiz, Guerrero-Bote, & Moya-

Anegón, 2016; Blažun, Kokol, & Vošner, 2015; Chen, Arsenault, Gingras, & Larivière,

2015; De Bellis, 2009; Laengle et al., 2017; Mryglod, Holovatch, Kenna, & Berche, 2016;

Naukkarinen & Bragge, 2016; Sweileh, 2017; Sweileh, Al-Jabi, AbuTaha, Zyoud, Anayah, &

Sawalha, 2017; Sweileh, Al-Jabi, Sawalha, Zyoud, 2016; Sweileh et al., 2016). Bibliometric

analysis has rarely been applied to music psychology. We have found two articles that used

a bibliometric approach to study perception and cognition research (Tirovolas & Levitin,

2011) and music and affect research (Diaz & Silveria, 2014). In Triovolas & Levitin (2011)

study, the authors looked at publications within one journal (i.e., Music Perception), covering

a total 578 articles. The retrieved literature was coded to look at the most frequent topics,

populations, stimuli, materials, outcome measures, and music styles, indicating their trends

A.1 Introduction

between 1984 and 2009. They also provided a list of the top 20 most highly cited articles

published in Music Perception and the top 20 articles published outside Music Perception

that were most cited in the journal. Finally, the authors showed the most productive countries

publishing in Music Perception. In the paper by Diaz and Silveria (2014), the authors looked

specifically at music and affective phenomena. They focused on three journals: the Journal

of Research in Music Education, Psychology of Music, and Music Perception. The authors

used a strict inclusion criteria to select articles related to topics relevant to affective aspects

of music, resulting in a total of 286 articles. While the study by Triovolas & Levitin (2011)

only focused on one scientific journal, the study by Diaz and Sliveria (2014) had a very

narrow topical focus. Thus, these studies cannot give insight into trends throughout music

psychology as a whole. Moreover, the studies used very few bibliometric indicators. For

instance, they did not provide information about growth of publications, more elaborated

citation analysis, or author productivity and collaborations. Another important limitation in

these two studies is the use of human coders to analyze the content of the articles.

The present study aims to produce a large-scale computational bibliometric analysis of the

scientific literature published in music psychology from 1973 to 2017. Using this method of

analysis we aim to better understand research trends, citations, authorships, collaborations,

as well as global contributions. This is important in identifying future directions within

the field. To reduce potential sources of bias and analyze systematically a large amount of

documents, the present study used the R package Bibliometrix (Aria & Cuccurullo, 2017),

a tool for quantitative research in bibliometrics that provides various functions to perform

citation, coupling, and scientific collaboration analysis. To visualize the data, we used

VOSviewer (Van Eck & Waltman, 2010), a software tool that applies advanced clustering

and natural language processing techniques for generating and visualizing maps based on

network data. VOSviewer software has been used in a large body of published literature

(http://www.vosviewer.com/publications), generating over 500 publications since 2006. To

the best of our knowledge, the software has not yet been applied to music psychology

literature.

In the present study we analyze, through visualization and bibliometric techniques, all

published literature from Psychology of Music, Music Perception, and Musicae Scientiae,

focusing on five key aspects of the retrieved literature: (1) growth of publications (i.e., annual

growth rate, relative growth rate, and whether there are significant temporal changes in the

number of publications over time), (2) citation analysis (i.e., number of citations per journal

and year, top cited authors and papers, and whether there are significant temporal changes

in the number of citations over time), (3) authorship analysis (i.e., productivity,

A.2 Methods

dominance, collaboration index, visualizations of authorship collaboration, and Lotka’s

law), (4) country analysis (i.e., productivity, visualization of country collaboration, and

geographical distribution of the publications), and (5) the main conceptual language used in

the retrieved literature.

A.2 Methods

A.2.1 Data collection and search strategy

The data used in this study was retrieved from Scopus, a bibliographic database that covers

over 20,000 journals, including technical, medical, and social sciences titles. Scopus is

larger than PubMed and Web of Science (Falagas, Pitsouni, Malietzis, & Pappas, 2008) and

offers many relevant features that facilitate bibliometric analysis (e.g., author, country, and

affiliation contributions, citation analysis, and the “source type” function).

We searched all available literature, by “source title”, in Psychology of Music, Music Percep-

tion, and Musicae Scientiae. Using the Scopus “source type” function, we limited the search

to empirical and review articles only, excluding book chapters, conference papers, and edito-

rial

notes. We also excluded any document from 2018 because it was the year in which this study

was conducted. All available results were then exported to text files, including citation

information (i.e., authors, document title, year, source title, volume, issue, pages, citation

count, source and document type, and DOI), bibliographical information (i.e., affiliations,

serial identifiers, publisher, editor, language of original document, correspondence address,

and abbreviated source title), abstracts, as well as keywords. All data was retrieved on the

20th of April 2018 (see supplementary materials for the two main datasets used in this study).

In some situations, the same author might have more than one name, use different initials

in different publications (e.g., Sloboda, J. vs. Sloboda, J. A.), or have different name

spellings. This might generate inaccuracy and inconsistencies in the computational analysis

of authorship. There is not a clear solution to this problem. However, to reduce its negative

impact, we deleted the second initials from all authors’ names in the retrieved dataset,

including only the first surname and first initial. Moreover, it is important to note that our

data has a gap between 2002 and 2004 in the literature retrieved from Music Perception.

Scopus did not contain any documents from this source during these three years.

A.3 Results

A.2.2 Data analysis and visualization

Descriptive statistics and standard bibliometric indicators, including citation analysis, annual

growth of publications, authorship productivity, dominance, collaboration index, and country

productivity were used to produce an overview of the retrieved data. The application and

presentation of some of these indicators was based on the analysis reported in Sweileh et

al. (2017). In addition, we used the R package bibliometrix (Aria & Cuccurullo, 2017) to

analyze the most productive authors, countries, keywords, top cited articles and authors,

author dominance, index-h, and Lotka’s Law.

Visualization and Bibliometric maps were created using emphVOSviewer (Van Eck &

Waltman, 2010), which uses a unified framework for mapping and clustering (Waltman, Van

Eck, & Noyos, 2010). The software is mainly intended for analysis of bibliometric networks

and can create three types of visualizations: Network visualizations, overlay visualizations,

and density visualizations. In the network visualizations, items are represented by their label

and by a circle. The size of the circles is determined by the weight of the item. The color

of an item is determined by the cluster to which the item belongs. Lines between items

represent links and the stronger the link is, the wider the line. The distance between items in

the map, indicates the degree of relatedness between them. Furthermore, we used the R

package rworldmap (South, 2011) to generate a visualization of the world’s geographical

distribution of countries’ productivity.

A.3 Results

A.3.1 Retrieved Literature

A total of 2,089 documents were retrieved, covering a time period of 44 years (1973-2017)

beginning from the first publication of Psychology of Music in 1973.Table A.1 shows the

total number and type of articles retrieved as well as the average number of publications per

year within each of the three journals and within all of the journals in total. The majority of

documents were research articles (1,987; 95.12%), whereas review articles only represented

a minimal portion (102; 4.88%). Psychology of music was the journal with the largest

number of retrieved articles (934; 44.71%), followed by Music Perception (746; 35.71%),

and Musicae Scientiae (409; 19.58%). However, when taking into account the years that

each journal has been active, the average number of publications per year is comparable

across the three journals (20.76, 23.31, and 19.48, respectively). Table A.2 shows the top 20

contributions made by author, keywords, and countries. See Appendix A, in the paper

A.3 Results

published online, for the tables of the top 20 contributions made by author, keyword, and

country by decade, and Appendix B for the top 10 contributions by each journal.

Table A.1 Number and type of articles retrieved.

PoM: Psychology of Music; MP: Music Perception; MS: Musicae Scientiae.

Table A.2 Top 20 contributions of authors, keywords, and countries.

TP: total publications. *Country of corresponding author.

A.3.2 Growth in number of publications

The mean number of publications from 1973 to 2017 was 46.42 (SD= 35.56). The total

percentage of relative growth was 11%. The highest productivity was observed in 2016 with

a total of 135 publications (6.46%) and the lowest productivity was observed in 1975 with a

total of 9 publications (.43%). Figure A.1 shows the total number of publications in the three

journals over time. The total number of publications increased significantly over time, as

indicated by a simple linear regression, F(1,43)= 141.1, p< .001, with an R2 of .766.

Table A.3 shows the annual number of publications, annual growth rate (AGR), and relative

growth rate (RGR). The AGR indicates the percentage of change in the number of publications

A.3 Results

over one year. The AGR is calculated using the following equation: AGR = [(TP ending value

– TP beginning value)/ TP beginning value] *100, where TP is total number of publication.

The RGR indicates the growth rate relative to the total number of publications per year. The

RGR was calculated based on the following equation: RGR = [loge W2 - loge W1] / (T2 – T1),

where loge W2 is the log of the final number of publications after a specific period of interval;

logeW1 is the log of the initial number of publications; and T1-T2 is the unit difference

between the initial time and the final time.

Appendix C (in the paper published online) shows the annual number of publications, AGR,

and RGR in the three journals separately. In Psychology of Music, the average number of

publications from 1973 to 2017 was 20.76 (SD = 16.48), with a total relative growth rate of

9%. In Music Perception the mean number of publications from 1983 to 2017 was 23.31 (SD

= 8.31), with a total relative growth rate of 15%. In Musicae Scientiae, the average number

of publications from 1997 to 2017 was 19.48 (SD = 10.38), with a total relative growth rate

of 18%.

Figure A.1 Total number of publications per journal over time.

PoM: Psychology of Music; MP: Music Perception; MS: Musicae Scientiae.

A.3 Results

Table A.3 (left) Annual number of publications, AGR, and RGR.

Table A.4 (right) Summary of the citation analysis.

AGR: annual growth rate and RGR: relative growth rate. TC: total citations.

A.3.3 Citation analysis

Table A.4 shows the summary of the citation analysis of all three journals combined. Re-

trieved documents received a total of 33,771 citations, a mean of 16.17 (SD = 26.93) citations

per document, and median (Q1 – Q3) of 7 (2 – 20). While the highest number of citations per

document was in 2007, with 2,059 (M= 24.2, SD = 31.4) citations, the lowest was in 1975

with 25 citations (M= 2.8, SD = 2.9). Figure A.2 shows the average total number of citations

A.3 Results

over time. Across the entire time period, the average number of citations did not increase

significantly, as indicated by a simple linear regression, F(1,43) = .21, p= .65, R2 = .005.

However, the relationship between the average citations and year followed an inverted-U

shaped, as indicated by a statistically significant quadratic regression, F(2,42) = 52.65, p <

.001, R2 = .715.

Appendix D (in the paper published online) shows the summary of citation analysis in the

three journals separately. In Psychology of Music, the retrieved documents received a total

of 13,344 citations, a mean of 16.98 (SD = 26.12) citations per document, and median (Q1 –

Q3) of 8 (3 - 21). In Music Perception, the documents received a total of 17,069 citations, a

mean of 24.38 (SD = 33.25) citations per document, and median (Q1 – Q3) of 14 (5 - 29). In

Musicae Scientiae, the documents received a total of 3,358 citations, a mean of 10.17 (SD =

15.04) citations per document, and median (Q1 – Q3) of 5 (2 - 12).

Figure A.2 Average total citations per year over time.

The top 10 cited articles and authors in the retrieved literature are shown in Table A.5 (a)

and (b) respectively. The publication that received the highest amount of citations was

“Perception of Temporal Patterns” by Povel and Essens (1985), with a total of 364 citations

and an average of 11.03 citations per year. The author with the highest number of citations

was John Sloboda who received a total of 1,070 citations.

A.3 Results

Table A.5 (a) Top 10 cited articles in the retrieved literature and (b) top 10 cited authors in

the retrieved literature.

PoM: Psychology of Music; MP: Music Perception; MS: Musicae Scientiae. TC: total citations; TP: total

publications.

A.3.4 Authorship analysis: productivity, dominance, collaboration, and

Lotka’s Law

A total of 2,632 authors were covered in the retrieved literature, with a mean of 1.26 authors

per article and a mean of .79 articles per author. The mean number of co-author per article

was 2.08. Table A.6 shows the average authors per document, author productivity, and

collaboration index (CI). The mean number of authors per document increased significantly

over time, from a mean of 1.2 in the first period of 10 years (1973-1982) to a mean of

2.48 in the last period of 10 years (2008-2017), F(1,43) = 221.19, p< .001, R2 = .837. The

collaboration index (CI) for multi-authored papers (CI = number of authors in multi-authored

publications/number of multi-authored papers) increased significantly over time from 2.00 in

1974 (the first year with a multi-authored paper) to 2.98 in 2017, F(1,43) = 78.91, p < .001,

R2 = .653.

A.3 Results

Table A.6 Average authors per document, author productivity, and collaboration index.

Percentages in brackets. TA: total number of authors and CI: collaboration index.

Figure A.3 shows the number of single-authored and multi-authored publications over time.

While a total of 828 documents (39.67%) were single-authored publications, a total of 1,262

publications (60.41%) were multi-authored. Figure A.4 shows a network visualization map

of author collaborations. The relatedness of authors is determined based on their number of

co-authored publications. Authors with a minimum of 5 co-authorship publications and a

minimum of 100 total citations are visualized, resulting in a total of 49 authors.

A.3 Results

Figure A.3 Number of single-authored and multi-authored publications over time.

Figure A.4 Network visualization map of author collaborations.

The width of the line shows the strength of the collaboration. The size of the circle indicates the total number

of publications per author. The color of the circle indicates the cluster to which the author belongs.

A.3 Results

Table A.7 shows the authors with a minimum dominance factor of > .1. The dominance factor

was proposed by Kumar & Kumar (2008), indicating a ratio of the fraction of multi-authored

publications in which an author appears as first author (dominance factor 1 means that an

author is the first author in all of his or her multi-authored papers). The author with the

highest dominance factor (.47) was Tuomas Eerola, being the first author in 8 publications

out of 17 multi-authored publications.

Table A.7 Authors with a minimum dominance factor of > .1.

Figure A.5 depicts Lotka’s law coefficient for scientific productivity (Lotka, 1926), indicating

the theoretical distribution (red) and the estimated distribution based on the retrieved literature

(blue). Lotka’s law describes the frequency of publication by authors in any given field. It

assumes an inverse square law in which the number of authors making a certain number of

contributions is a fixed ratio to the number of authors publishing a single article, implying

that the theoretical beta coefficient of Lotka’s law nearly always equals 2. Using the function

lotka from the R package bibliometrix (Aria & Cuccurullo, 2017) we estimated the Beta

coefficient of the retrieved literature, which was 2.3 and had a goodness of fit equal to .94. A

Kolmogorov-Smirnoff two sample test indicated that there were no significant differences

between the observed and the theoretical Lotka distribution, p = .22.

A.3 Results

Figure A.5 Lotka’s law coefficient for scientific productivity (theoretical and estimated

distributions).

A.3.5 Country analysis: productivity, collaborations, and geographi-

cal distribution

The number of countries contributing to the retrieved literature was 49. Table A.8 displays

the countries with a minimum production of 5 publications, including their frequency, total

number of citations, and the number of single country publications as well as multiple country

publications. The USA and the UK had the highest total citations with 8,669 (25.67%) and

5,954 (17.63%) and a mean of 17.99 and 18.04 citations per publication, respectively.

Figure A.6 shows the geographical distribution of publications. The map was created using

the R package rworldmap. The map is color-coded using six categories (1 = 0-100, 2 = 101-

200, 3 = 201-300, 4 = 301-400, 5 = 4001, 500, and 6 = 5001-600 publications), where

countries in dark green color had the highest productivity and light yellow countries with the

lowest. Countries with no color indicate that there was no retrieved data from these areas.

Figure A.7 depicts a network visualization map of international collaborations. The related-

ness of countries is determined based on their number of co-authored publications. Countries

with a minimum of 10 international co-authorship publications and a minimum of 100 total

citations are visualized. As a result, 19 countries are visualized, clustering in 4 groups. Closer

circles indicate closer research collaboration between countries.

A.3 Results

Table A.8 Countries with a minimum productivity of five publications (country of

corresponding author).

TP: total publications, TC: total citations, SCP: single-country publication, MCP: multiple-country publication.

A.3 Results

Figure A.6 Geographical distribution of publications without correcting for country

population (top) and with the correction (bottom).

Countries colored dark blue had the highest productivity and countries colored light yellow had the lowest.

Countries with no color indicate that there was no retrieved data from these areas.

Figure A.7 Network visualization map of international collaborations.

The width of the line shows the strength of the collaboration. The size of the circle indicates the total number

of publications per country. The color of the circle indicates the cluster to which the country belongs.

A.4 Discussion

A.3.6 Conceptual Language

Figure A.8 shows an overlay visualization map of author keywords occurrences (i.e., key-

words listed by the authors on each publication). Only keywords that occurred a minimum

of 10 times were included, resulting in a total of 75 keywords. Overlay maps are similar

to network maps but they are colored based on a given score. The scores used in Figure 8

are based on the average publication year of each keyword. Dark blue represents the oldest

average year of publications and red the most recent. The interpretation of the maps is

the same as in the network visualization maps (i.e., closer circles indicate closer keyword

occurrence).

Figure A.8 Network visualization map of keyword occurrences.

The width of the line shows the strength of the co-occurrence between keywords. The size of the circle

indicates the total number of occurrences. The color of the circle indicates average year of publications.

A.4 Discussion

This study aimed to analyze, through visualization and bibliometric techniques, all published

literature from Psychology of Music, Music Perception, and Musicae Scientiae. Using all

available literature in Scopus, a total of 2,089 publications, 2,632 authors, and 49 countries

constituted the retrieved literature, covering a time span of 44 years (1973-2017). The major-

A.4 Discussion

ity of publications were empirical articles (95%), whereas review articles only represented

the 5% of the literature.

A.4.1 Comparing the three journals

Psychology of Music was the first journal to begin publishing in 1973. Second was Music

Perception in 1983 and third Musicae Scientiae in 1997. These differences in the active time

span of each journal explain why Psychology of Music has the largest amount of retrieved

articles (44%), followed by Music Perception (36%), and Musicae Scientiae (20%). However,

the average number of publications per year in the three journals is very similar (20.76, 23.31,

and 19.48, respectively).

Interestingly, Musicae Scientiae has the highest relative

growth rate

18%. Music Perception has a relative growth rate of 15% and Psychology of Music of 9%.

When looking at the average citations per document, Music Perception has the highest mean

citations per document (M = 24.38, SD = 33.25), followed by Psychology of Music (M

=16.98 (SD = 26.12) and Musicae Scientiae (M = 10.17, SD = 15.04). Nevertheless, this

pattern changes if we look at the average citations in the three most recent years (from 2015

to 2017), as calculated by SCImago’s SJR ranking. In this case, Psychology of Music

remains in the first place, but Musicae Scientiae moves forward to the second position and

Music Perception to the last. These results could inspire future research to investigate reasons

for

such differences. For example, examining how funding, publication costs, access, and

editorial teams might influence or predict productivity and citation outcomes.

A.4.2 Growth of publications

Our results show that from 1973 to 2017 there was an overall growth in the number of publi-

cations across all three journals. This may not be surprising as research article publications

have seen an overall 3% growth every year across all disciplines and there is some indication

that this growth has accelerated even more in recent years (Ware & Mabet, 2015). This

growth may also be due to an increase in the amount of researchers overall and an increase in

the number of journals publishing music psychology research (Ware & Mabe, 2015). From

our retrieved literature we found an overall growth rate of 11%, which is slightly higher than

the overall average of 3% (Ware & Mabe, 2015).

The growth of music psychology is not only represented by our results but might also be

evident in the amount of pop science articles published in recent years. For example, articles

written for Psychology Today such as "Musical Preferences and the Brain" (Greenburg,

2017), op-eds in the New York Times such as “Why Music Makes Our Brain Sing” (Zatorre

A.4 Discussion

& Salimporr, 2013), and popular books such as This Is Your Brain On Music (Levitin, 2006)

and

Musicophilia: Tales of Music and the Brain (Sacks, 2007). Growth of interest in music

psychology and its research, more specifically music and health research, may also be seen

in the formation of the UK because All-Party Parliamentary Group on Arts, Health and

Wellbeing (APPGAHW) in 2014, which aims to improve awareness of the benefits that the

arts can bring to health and wellbeing. This UK group uses the research findings from music

psychology, and other related arts disciplines, to help inform policies. Future research could

be done to investigate the subsequent effects of increases in publications on the number of

popular science publications and on governmental policies. Understanding this could give

better insight into the impact of music psychology research outside of the academic audience.

A.4.3 Citation analysis

The retrieved documents received a total of 33,771 citations, with a mean of 16.17 (SD =

26.93) citations per document. This is relatively small compared to other related disciplines

such as neuroscience, with 187 average citations per article, experimental psychology with

67, and clinical psychology with 68 (Patience, Patience, Blais, Bertrand, 2017). However,

compared to music research publications, which has an average of about seven citations per

article, it is relatively higher (Patience et al., 2017).

Across the entire time period, the average number of citations did not increase significantly.

However, we identified a significant inverted-U shaped relationship between year of publica-

tion and average number of citations, with its highest peak in 2007, which received 2,059

citations. This finding can be explained by the natural gap between year of publication

and year of first citation. Hancock and Price (2016) provided some evidence of this gap by

examining the first citation speed for articles in Psychology of Music from 1973 and 2012.

The authors found that the probability of an article receiving a first citation was .25 after 2

years, .50 after 4 years, and .75 after 7 years (Hancock & Price, 2016).

The publication that received the highest amount of citations was “Perception of Temporal

Patterns” by Povel and Essens (1985), with a total of 364 citations and an average of 11.03

citations per year. When looking at the top ten most cited articles (Table A.5a), we see that

four out of the ten are about music and emotion and three are about investigating the

temporal aspect of music. This may speak to the most cited areas or sub disciplines in the

field of music psychology within these three journals. The author with the highest number of

citations was John Anthony Sloboda who received a total of 1,070 citations. John Sloboda is

also known for his research in music and emotion, again emphasizing a key area of music

psychology research over the years.

A.4 Discussion

However, note that these results only cover articles published within three music psychology

specific journals. For instance, we are not capturing articles published in neuroscience or

general psychology journals that represent other sub disciplines within music psychology. It

is also important to mention that we only used the citation analysis provided in Scopus on

the 20th of April 2018. The content of this database is frequently updated, therefore, the

numbers reported here will likely change over time. Moreover, there are significant

differences between the number of citations indexed in Scopus and other databases, such

as Web of Knowledge and Google Scholar (Meho & Yang, 2007). While both Scopus and

Web of Knowledge index mostly refereed journal articles, Google Scholar indexes refereed

and non-refereed types of documents. In addition, citation counts in different databases rely

strongly on the subject matter of the researcher (Meho & Yang, 2007), some subjects being

more represented in one database than in another.

Although it was beyond the scope of this study, it would be interesting to carry out an analysis to

understand different factors which may predict the number of citations a publication might

receive. As predictors, one could use the total number of authors per document, gender of

the author, affiliation, country, funding body, research area, and/or journal of publication.

For instance, Patience et al. (2017) found that the citation rate correlates positively with the

number of funding agencies that finance the research. This is a thought-provoking element

we did not account for in the present study. The effect funding has on the dissemination and

impact of certain research is known, but not within the field of music psychology specifically.

A.4.4 Authorship analysis

A total of 2,632 authors were covered in the retrieved literature, with a mean of 1.26 authors

per article and a mean of .79 articles per author. The mean number of authors per document

increased significantly over time, from a mean of 1.2 in the first period of 10 years (1973-1982)

to a mean of 2.48 in the last period of 10 years (2008-2017). While a total of 828 documents

(39.67%) were single-authored publications, a total of 1,262 publications

(60.41%) were

multi-authored. Both the number of single-authored papers and multi-authored papers increased

significantly over time. However, the magnitude of this increase was higher

in the publications

with multiple authors. Finally, the collaboration index (CI) for multi- authored papers (i.e.,

CI = number of authors in multi-authored publications/number of multi-authored papers)

increased significantly over time, from 2.00 in 1974 (the first year with a multi-authored

paper) to 2.98 in 2017. We can speculate why this growth may be happening, including a

mixture of a spike in interest in music psychology, music psychology

A.4 Discussion

related journals, and programs training students in music psychology subsequently resulting

in more music psychology researchers.

This growth in the total number of authors and collaboration are not just a significant trend

in music psychology but is observed in general scientific literature. The Economist (2016)

found that in 34 million research papers published in peer-reviewed journals and conference

proceedings between 1996 and 2015, the average number of authors per paper grew from 3.2

to 4.4. Many factors could be responsible of this growth. One reason could be due to the fact

that research is becoming more multi- and interdisciplinary in general, which is particularly

true in the case of music psychology. Another reason may be due to authors wanting to “pad

their publication lists” and the increasing institutional pressures to “publish or perish” (The

Economist, 2016). Multi-authored papers help cut down the workload resulting in more

publications per author per year. Future research could investigate more systematically the

reason for this increase and try to understand how this might affect the impact or rigor of

published scientific research.

The visualization map also gives a good indication of the spread of collaboration happening

both internationally and within specific domains. For example, the blue cluster in the network

visualization (Figure A.5) includes individuals from a range of sub disciplines such as

everyday uses of music, music perception and music and memory and is mostly comprised of

UK researchers. The author dominance factor (i.e., a ratio of the fraction of multi-authored

publications in which an author appears as first author) is interesting in providing a different

type of information in addition to author productivity. This visualization helps to track how

collaborations across different domains and areas may be carried or created by certain

dominate individuals within the field.

Finally, when comparing our data set to Loka’s theoretical distribution (Lotka, 1926), we

found no significant differences between the observed and the theoretical distributions.

Although expected, this is a good indicator that the literature in music psychology conforms to

Lotka’s law. That is, the distribution of the number of authors and their scientific productivity

(i.e., number of publications) is highly asymmetric: While very few authors publish many

articles, the remaining authors publish very few.

A.4.5 Country analysis

When looking specifically at the international collaborations and distributions of publications

found that out of the total 49 countries contributing to the retrieved literature, the US and

the UK were the most productive cuntries, defined as having the highest number of

A.4 Discussion

publications (US = 23% and UK = 16%) and citations (US = 26% and UK = 18%). Note,

however, that our analysis did not account for population size and this variable is confounded

in the data. The collaboration network map shows this predominance of the UK and the USA

as well but also shows how more countries collaborate with the UK creating more

international collaborations than the US. This may have to do with the UK being within

the wider EU and thus fostering more collaboration between countries. This prominence

of research coming from the US and the UK is not specific to music psychology. However,

the full picture of nation productivity in music psychology looks different compared to the

general picture. The world’s most research-intensive nations, measured by field-weighted

citation impact are the UK, US, China, Japan, Germany, Italy, Canada and France (Kisjes,

2013). However, in our study the top 8 most productive countries were the USA, UK,

Australia, Canada, Germany, Finland, France, and the Netherlands. The productivity of these

countries may be related to certain funding opportunities available in these countries, number

of labs and number of teaching programs based in these countries. Future research could

investigate how funding affects the geographical distribution of music psychology. It is

important to think about which nation’s voices are being heard and which are the loudest

within the music psychology research. There is a limitation in knowledge if only a few

nations are represented. Working towards creating opportunities in other countries for music

psychology research and providing places for people to train could help disperse the

distribution beyond Europe and the US.

A.4.6 Main conceptual language

The keywords that researchers used to describe their articles and how often they co-occur

with others indicate the research trends and themes in music psychology. By selecting those

keywords that occurred a minimum of 10 times we obtained a total of 75 keywords (Figure

A.8). The keywords “music” and “emotion” have the highest number of co-occurrences as

well as connections with other keywords. This finding is in line with the general interest and

significant increase in research on music and emotion (Eerola & Vuoskoski, 2013;

Gabrielsson & Lindstrom, 2011; Juslin & Laukka, 2003; Västfjäll, 2002). While some

keywords connect very well with others (e.g., memory, performance, preference), others are

more disconnected (e.g., flow, cross-cultural, musical expertise). It is also interesting to see

how a close group of keywords represent research areas. For instance, a clear research area

is constituted by “timing”, “synchronization”, “rhythm”, and “meter”; another by “music

therapy”, “stress”, “depression”, “individual differences”, and “personality”. In addition, the

overlay map shows how keyword use changes over time. We can see that keywords such as

“synchronization” and “timing” both co-occur and are prominently used

A.4 Discussion

in the early 2000s, whereas keywords such as “self-regulation”, “flow”, and “emotion

regulation” appear more popular in recent publications. Overall, this network map allows

us to summarize and better understand the complex field of music psychology in a single

picture, but the applications of this visualization technique are far-reaching. We encourage

researchers to use this tool to define unexplored research areas within music psychology as

well as complement their literature reviews. Although this is the first published article that

uses VOSviewer (Van Eck & Waltman, 2010) to create visualization network maps within

music psychology, the software has been used in more than 500 publications since 2006

(http://www.vosviewer.com/publications ).

A.4.7 Limitations of the study

The present study has two main limitations. First, we only included three journals in our

analysis. This choice was based on the journals’ content and impact. The aim was to select

the most prominent journals that specifically look at music psychology research. Moreover,

we needed to use journals indexed in Scopus, as we used this database to retrieve the literature

(e.g., the journals of Psychomusicology: Music, Mind, and Brain and Music & Science are

yet not indexed in Scopus). This is an important limitation because high quality research on

music psychology is published in a wide range of journals from a wide variety of disciplines,

including experimental psychology, social psychology, clinical psychology, computer science,

marketing and advertising, personality, and neuroscience. Thus, our study only examines a

fraction of the total number of music psychology research publications and our conclusions

can then only be drawn from this fraction of literature. It also means that some authors that

do not appear as relevant in this dataset might actually be very influential in general.

The second main limitation is due to the fact that we used Scopus to retrieve the literature,

including the citation analysis. This limitation is inherent to any bibliometric study using

similar search strategies. Even though Scopus is the largest existing database (Falagas et al.,

2008), it is not a complete record of all published literature due to licensing. For example,

articles from Music Perception between 2002 and 2004 are missing in Scopus. In addition,

when performing databases searches, there is a potential for false positive and false negative

results; and the number of citations differ depending on the database (Meho & Yang, 2007).

Finally, some authors might have more than one name or different name spelling, which

might have caused inaccuracies in the result. Although no ideal solution exist to this problem,

reduced its potential negative impact at the minimal level by deleting the second initials from

all authors’ names in the retrieved dataset, including only the first surname and the first

A.5 Conclusion

initial. We hope that the limitations of the current study are justified by the benefits of using

large-scale computational bibliometric analysis.

A.5 Conclusion

The study reported here begins to investigate the general research trends, reach, and gaps

within the published literature in three prominent music psychology journals. Using bib-

liometric techniques to visualize and understanding the past and present of research in

music

psychology leads us to critical observations and conclusions opening many interesting

avenues

for future collaborations and research in the field.

More international collaboration, outside of Europe and the USA should be persued, allowing

for

different types of questions, methods and potential findings, steering our field away from

WEIRD (Westernized, educated, industrialized, rich, and democratic) populations

(Henrich,

Heine, & Norenzayan, 2010). Future studies should be done to investigate potential

predictors

of music psychology research citations. Understanding how the system around music

psychology research, it’s funding schemes, organizations and institutions, and the influence

of certain individuals and countries impact the dissemination and academic impact of music

psychology research could shed light on how the system is working and potential ways to

improve it. Finally, future research should be persued investigating the wider impact of music

psychology research to the general public and policies. Scientific communication and

research impact is only becoming more important. Using similar computational large- scale

analysis allows for these questions to be more objectively addressed.

Music psychology is still a relatively young field. Taking the time to objectively look back

and reflect on how the field has progressed, which this study has only just begun to do, helps

push the field forward in new and exciting directions. More research should be done using

similar methods giving insight to the past, present and future of music psychology research.

A.6 References

Aria, M., & Cuccurullo, C. (2017). bibliometrix: An R-tool for comprehensive science

mapping analysis. Journal of Informetrics, 11(4), 959-975.

Bellis, N. De. (2009). Bibliometrics and citation analysis: from the science citation index to

cybermetrics. Scarecrow Press.

A.6 References

Blázquez-Ruiz, J., Guerrero-Bote, V. P., & Moya-Anegón, F. (2016). New scientometric-

based knowledge map of food science research (2003 to 2014).

Comprehensive Reviews

Food Science and Food Safety, 15(6), 1040–1055.

Blažun, H., Kokol, P., & Vošner, J. (2015). Research literature production on nursing

competences from 1981 till 2012: A bibliometric snapshot. Nurse Education Today,

35(5), 673–679.

Chen, S., Arsenault, C., Gingras, Y., & Larivière, V. (2014). Exploring the interdisciplinary

evolution of a discipline: the case of Biochemistry and Molecular Biology. Scientomet-

rics, 102(2), 1307–1323.

Diaz, F., Music, J. S.-J. of R. in, & 2014, undefined. (n.d.). Music and affective phenomena:

A 20-year content and bibliometric analysis of research in three eminent journals.

Journal of Research in Music Education, 62(1), 66-77.

Eerola, T., & Vuoskoski, J. K. (2013). A review of music and emotion studies: approaches,

emotion models, and stimuli. Music Perception: An Interdisciplinary Journal, 30(3),

307-340.

Falagas, M. E., Pitsouni, E. I., Malietzis, G. A., & Pappas, G. (2008). Comparison of

PubMed, Scopus, web of science, and Google scholar: strengths and weaknesses. The

FASEB journal, 22(2), 338-342.

Gabrielsson, A., & Lindström, E. (2001). The influence of musical structure on emotional

expression. In P.N. Juslin, & J.A. Sloboda (Eds., pp. 223-248), Music and Emotion:

Theory and Research. New York: Oxford University Press.

Hallam, s., Cross, I., & Thaut, M. (2011). Oxford handbook of music psychology (2nd ed.).

Oxford, UK: Oxford University Press.

Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?.

Behavioral and Brain Sciences, 33(2-3), 61-83.

Juslin, P. N., & Laukka, P. (2003). Communication of emotions in vocal expression and

music performance: Different channels, same code?. Psychological Bulletin, 129(5),

770.

Laengle,

S.,

Merigó,

M.,

Miranda,

J.,

Słowin´ski,

R.,

Bomze,

I.,

Borgonovo,

E.,

Teunter,

R. (2017). Forty years of the European Journal of Operational Research: A bibliometric

overview. European Journal of Operational Research, 262(3), 803–816.

Levitin, D. J. (2006). This is your brain on music: The science of a human obsession.

Penguin.

Meho, L. I., & Yang, K. (2007). Impact of data sources on citation counts and rankings of

LIS faculty: Web of Science versus Scopus and Google Scholar. Journal of the

American Society for Information Science and Technology, 58(13), 2105-2125.

A.6 References

Mryglod, O., Holovatch, Y., Kenna, R., & Berche, B. (2016). Quantifying the evolution

of a scientific topic: reaction of the academic community to the Chornobyl disaster.

Scientometrics, 106(3), 1151–1166.

Patience, G. S., Patience, C. A., Blais, B., & Bertrand, F. (2017). Citation analysis of

scientific categories. Heliyon, 3(5), e00300.

Povel, D.-J., & Essens, P. (1985). Perception of temporal patterns. Music Perception, 2(4),

411–440.

Sacks, O. (2010). Musicophilia: Tales of music and the brain. Vintage Canada.

Sweileh, W. M. (2017). Global research trends of World Health Organization’s top eight

emerging pathogens. Globalization and Health, 13(1), 9.

Sweileh, W. M., Al-Jabi, S. W., AbuTaha, A. S., Zyoud, S. H., Anayah, F. M. A., & Sawalha,

A. F. (2017). Bibliometric analysis of worldwide scientific literature in mobile - health:

2006-2016. BMC Medical Informatics and Decision Making, 17(1), 1–12.

Sweileh, W. M., Al-Jabi, S. W., Sawalha, A. F., & Zyoud, S. H. (2016a). Bibliometric profile

the global scientific research on autism spectrum disorders. SpringerPlus, 5(1).

Sweileh, W. M., Shraim, N. Y., Al-Jabi, S. W., Sawalha, A. F., Rahhal, B., Khayyat, R. A., &

Zyoud, S. H. (2016). Assessing worldwide research activity on probiotics in pediatrics

using Scopus database: 1994–2014. World Allergy Organization Journal, 9(1), 25.

Tan, S. L., Pfordresher, P., & Harré, R. (2017). Psychology of music: From sound to

significance. Routledge.

Thaut, M. (2016). History and research. In S. Hallam, I. Cross, & M. Thaut (Eds.), Oxford

handbook of music psychology (2nd ed.). Oxford, UK: Oxford University Press.

Thompson, W.F. (2009). Music Thought & Feeling: Understanding the Psychology of Music.

New York: Oxford University Press.

Tirovolas, A. K., & Levitin, D. J. (2011). Music perception and cognition research from

1983 TO 2010: A categorical and blbliometric analysis of empirical articles in music

perception. Music Perception, 29(1), 23–36.

van Eck, N. J., & Waltman, L. (2010). Software survey: VOSviewer, a computer program for

bibliometric mapping. Scientometrics, 84(2), 523–538.

Västfjäll, D. (2001). Emotion induction through music: A review of the musical mood

induction procedure. Musicae Scientiae, 5, 173-211.

Waltman, L., Van Eck, N. J., & Noyons, E. C. (2010). A unified approach to mapping and

clustering of bibliometric networks. Journal of Informetrics, 4(4), 629-635.

Appendix B

The Behavioural Economics of Music:

Systematic review (S2)

The following paper has not yet been accepted to a peer-reviewed journal. The text presented

here is the most updated version of the manuscript as written by the time in which this thesis

was published (August 2021). For presentation in this thesis, the appendices of the paper

have been removed. Moreover, there may be minor modifications in the text to guarantee a

consistent typographic style throughout the thesis, such as the position of figures and tables.

Author contribution

I conceived of the initial idea of the paper, whereas Dr. Nikhil Masters (School of Social

Sciences, University of Manchester) played a major role in developing it further. Prof. Dr.

Daniel Müllensiefen (Goldsmiths, University of London) and Prof. Dr. Jochen Steffens

(Technische Universität Berlin) supervised the project at all stages. I was responsible for

most of the systematic literature review, whereas Prof. Dr. Jochen Steffens and Dr. Nikhil

Masters helped in the systematic screening processes. All other aspects of the paper were

done collaboratively with all authors. The paper was written together with Dr. Nikhil

Masters, with whom I share the first authorship.

The Behavioural Economics of Music: Systematic review

and future firections

In this paper, we conduct a systematic literature review to examine how behavioural eco-

nomics has been utilised to study judgements and decision making related to music. Using

a robust search strategy, we identified a total of 33 studies within four distinctive BEM areas

that readily apply to music research: heuristics and biases, social decision making,

behavioural time preferences, and dual process theory. We organised the discussion of this

literature around these BEM areas, which allowed us to demonstrate the value of behavioural

economics for music research and deduce further suggestions for future research. We hope

this paper contributes to the establishment of the Behavioural Economics of Music (BEM), an

interdisciplinary field that stimulates research on the intersection between music, psychology,

economics, and other disciplines. The BEM can be used as a toolkit to generate new ideas

and accelerate progress in any area concerned with human behaviours related to music.

Keywords: music psychology; behavioural economics; musical behaviour; decision making.

B.1 Introduction

This paper discusses the role of behavioural economics in understanding decision making

related to music. Specifically, we conduct a systematic literature review with the primary

goal of identifying studies that have utilised insights from behavioural economics to study

music. Our second goal is to use the results from our search to provide meaningful direction

for future research in this area. Based on these two objectives, we propose the Behavioural

Economics of Music (BEM), a unified research program that promotes the study of music

research using the tools of behavioural economics.

Our motivation flows from an extensive literature in music psychology – a branch of psy-

chology and musicology that aims to understand the psychological processes by which

music is perceived, processed, responded to, created, and integrated in everyday life (see

Hallam, Cross, & Thaut, 2016; Deutsch, 2013; Tan, Pfordresher, & Harré, 2017, for reviews).

Decision making is inherent to many of these processes and constitutes the basis of musical

behaviour. This includes, amongst others, music composition and improvisation choices,

performance evaluation, music preferences and listening behaviour, music consumption and

the market, the role of music in branding and advertising, and music use in health, such as

music therapy. Research on these areas has been crucial in our understanding of the

cognitive processes that drive music-related decision making. For example, studies have

identified the neural mechanisms that support decision making in music improvisation by

scanning musicians’ brains while improvising (see Beaty, 2015, for a review). When

looking

at music preferences and listening behaviour, researchers have found that individuals

listen to

music and choose different pieces to satisfy a number of psychological needs, from distraction

and motivation, to emotional regulation and stress reduction (e.g., Greb et al.,

2018;

Linnemann et al., 2018; Saarikallio & Erkkilä, 2007). In clinical contexts, practitioners

rely on

psychological-based models, such as the transformational design model (TDM), to

successfully design and implement music interventions to support patients’ health needs (see

MacDonald et al., 2013, for a review).

Independent from music psychology, music-related decision making has also been studied

through the lens of economics (see Byun, 2016; Connolly & Krueger, 2005; Tschmuck, 2007,

for general overviews). In particular, research from economics has been able to improve

understanding of decision making in music markets by successfully applying theoretical tools

from neoclassical economics such as producer theory, consumer theory and game theory.

Examples include behaviour of firms in the music industry (Burke, 1996; Ko & Lau, 2016;

Sweeting, 2013), explaining large scale music success (Elliott & Simmons, 2011; Hendricks

& Sorensen, 2009; Fox & Kochanowski, 2007; Strobl & Tucker, 2000), copyright and music

B.1 Introduction

piracy (see Oberholzer-Gee & Strumpf, 2009; Varian, 2005, for reviews), and the economics

of live music events (Decrop & Derbaix, 2014; Hiller, 2016; Holt, 2010; Krueger, 2005;

Larsen & Hussels, 2011).

In recent decades, there has been a burst of interest in the field of behavioural economics (see

Angner and Loewenstein, 2012; Dhami, 2016; Thaler, 2016, for overviews). In particular,

some researchers have begun to utilise tools from behavioural economics to study music-

related decision making. Behavioural economics, a subdiscipline of standard economics,

aims to increase its explanatory power by relaxing the rationality assumptions of homo eco-

nomicus, incorporating insights from an array of disciplines, such as psychology, sociology,

anthropology, biology, and neuroscience. Such an interdisciplinary approach seems ideal to

study the multi-faceted nature of music, adding value to the existing body of research in both

music psychology and economics.

Researchers from both economics and psychology stand to gain from utilising the behavioural

economics toolkit. Economists studying music are able to pursue a more evidence-based

approach in which individuals may not always be fully rational, self-interested utility max-

imisers. For example, consider how a music listener makes a choice about which song to

play next on their playlist. Such an individual may use mental shortcuts or heuristics to come

to a decision quickly rather than spend hours carefully considering all the alternatives.

Likewise, instead of setting a fixed price for their music, a music artist may instead ask for

voluntary donations from their fans as a ‘fairer’ way to distribute their music. For music

psychologists, since behavioural economics still maintains its economic identity, they have

access to a body of behavioural economic theory, mathematically rigorous, underpinned by

economic principles (see Dhami, 2016). Incorporating such theory may prove significant for

addressing key issues in music research that have eluded researchers so far. For instance, as

Greasly and Lamont (2016) remark, despite the wide range of psychological approaches that

have been used to investigate music preferences, there is currently no unified theory that can

accurately predict such preferences. Above all, we stress the synergistic benefits that music

researchers can obtain by utilising behavioural economics.

This paper contributes to the literature in two main ways. Our first contribution is to provide

an up-to-date account of studies using behavioural economics for research on music-related

decision making. Using a robust search strategy that is representative of the behavioural

economics literature, we find 33 studies identified within four distinct research areas of

behavioural economics – heuristics and biases, social decision-making behavioural time

preferences, and dual-process theory. We organise our discussion around these areas, enabling

the reader to gain a better understanding of where these studies fit in the broader context of

B.2 Methods

the behavioural economics literature. Our second contribution is to demonstrate the potential

of our proposed BEM research program by providing guidance for future work. Our focus

is on issues important to music researchers from psychology and economics where we see

clear benefit from using behavioural economics. We note that research in music-related

decision making is still relatively young, with no specific research field dedicated to this area

exclusively. Thus, we see our work as the first steps in creating this.

The remainder of this paper is organised as follows. Section 2 outlines the methods for

conducting the systematic literature review. Section 3 provides an overview of the results

from the systematic search.

Section 4 discusses the retrieved literature in more detail.

Section

5 discusses future directions within the BEM research program. Section 6 concludes.

B.2 Methods

This section outlines the methods employed for the systematic literature review. In section

B.2.1, we present our procedure for selecting the behavioural economics keywords to be used

in the systematic search. In section B.2.2, we give details of each stage of the systematic

review.

B.2.1 Behavioural economics keywords

An important requirement of the systematic literature review was to select a list of keywords

that is representative of behavioural economics. This procedure consisted of three steps,

summarised in Figure B.1: (i) extraction of keywords that appeared in titles, headings,

and sub-headings of prominent textbooks in behavioural economics and decision-making

published within the last 10 years, resulting in 585 keywords across all books (Angner, 2012;

Ball & Thompson, 2017; Cartwright, 2011; Dawes & Hastie, 2010; Dhami, 2016;

Holyoak

& Morrison, 2012; Ogaski & Tanaka, 2017; Wilkinson & Klaes, 2017); (ii) assessing

keyword

eligibility by selecting only those keywords that were duplicated in at least two of the

textbooks, leaving 69 keywords; (iii) creating a comprehensive list by including alternative

spellings (e.g., behaviour vs. behaviour) and synonyms (e.g., mental accounting vs.

psychological accounting), adding 46 extra keywords. Thus, the final list comprised a total

of 115 keywords. Three behavioural economists at the faculty level independently evaluated

the procedure using a 10-point scale (1 = not at all; 10 = very much) in terms of objectivity

(M = 7.7; SD = 1.7), adequacy (M = 8.0; SD = 0), and exhaustiveness (M = 7.5; SD = 0.5),

providing reassurance that overall our keyword strategy is effective.

B.2 Methods

Figure B.1 Behavioural economics keywords selection.

K = number of keywords at each step in the selection procedure.

B.2.2 Systematic literature review

Here we describe the methods undertaken for the systematic literature review to identify

published studies that have utilised behavioural economics to study music-related decision

making. Following an established protocol, we applied the methodology outlined by the

Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) (Moher,

Liberati, Tetzlaff, & Altman, 2009). The systematic review consisted of four stages: (i)

identification of studies through a database search, (ii) an initial systematic screening based

on titles and abstracts only, (iii) a second systematic screening based on full-text, and (iv)

coding of the final set of included studies. Figure B.2 summarises the outcome of each stage

using a PRISMA flow diagram.

In the identification stage, a database search was conducted using the following syntax: the

list of 115 behavioural economics keywords connected with the keyword "music" (including

the same word with different endings; i.e., musical, musicians, musicality, musicianship).

The database search was undertaken on the 6th of June 2018 using Scopus, Web of Science,

PsycINFO, Academic Search Complete, Business Source Complete, and the 10 first pages of

Google Scholar. The search parameters were the same in all databases: to identify papers

that used at least one of the behavioural economics keywords and the keyword “music” in the

title, abstract, or authors’ keywords; and to search only in peer-reviewed journals published

in English (excluding book chapters, book reviews, conference papers, and editorial notes). A

total of 338 studies were identified in this stage, which after duplicate studies were removed,

resulted in 202 studies ready for screening.

In both systematic screenings, we used the following inclusion criteria to determine whether

a study was included: (i) the study is written in English and published in a peer-reviewed

journal, (ii) the study examines judgements and decision making related to music, and (iii)

B.2 Methods

the study uses behavioural economics to study music-related decision making. In the first

systematic screening, two reviewers, a music psychologist (MA) and a behavioural economist

(NM), independently screened each study based only on the abstract. A third reviewer (JS)

screened those cases in which there were conflicts. This resulted in 119 studies proceeding to the

second screening stage. In the second systematic screening, two reviewers (MAT and JS)

independently assessed the studies based on full-text, with a third reviewer (NM) resolving

those cases with conflicts. The final number of studies included for the systematic review

was 35.

The final set of studies were independently coded by two reviewers (MAT and JS) along the

following attributes: research area within behavioural economics, research area within music,

academic discipline, and methods used (i.e., experimental, field-data, survey, theoretical). A

third researcher (NM) resolved any conflicts in the coding.

Figure B.2 PRISMA flow diagram.

B.3 Overview of results 79

B.3 Overview of results

From the systematic literature review we found 33 studies that utilised behavioural eco-

nomics to study music-related decision making. In particular, we identify four areas within

behavioural economics that these studies can be categorised (henceforth known as BEM

areas): heuristics and biases (n = 16), social decision-making (n = 9), behavioural time

preferences (n = 4), and dual-process theory (n = 4). We organise the discussion of these

studies around these BEM areas (see next section).

Within music, the most prominent areas are music preferences (n = 9) and music consumption

= 9), followed by music performance (n = 5), piracy (n = 4), music memory (n = 3), music

perception (n = 1), music in advertising (n = 1), and music in health (n = 1). The sample of

studies is multi-disciplinary, with the majority from psychology (including music psychology)

(n = 15), economics (n = 10), neuroscience (n = 5), business (n = 2), and health (n = 1). The

majority of studies are empirical (n = 28) with the most of these being experimental (n = 22)

and the rest survey, field-data, theoretical and mixed methods. Publication dates indicate that

this

literature is relatively recent (i.e., 27 out of the 33 studies were published in the last 10 years).

B.4 BEM Discussion

In this section we provide a discussion of the studies identified in the systematic review,

giving particular focus to their contribution from behavioural economics. We organise this

literature around the BEM areas outlined in the previous section.

B.4.1 Heuristics and biases

When making musical judgments and decisions, people are limited by their mental capacity,

available information, and time. For example, composers are influenced by their current emo-

tional state and memory when creating a new piece of music, whereas music performances

are bounded to the musicians’ cognitive and body capacity. Similarly, music preferences and

choices are influenced by the information available to the listener, such as the popularity of

the music or the prestige of the artist. This human condition is known as bounded rationality

(Simon, 1955, 1982) and affects decision making in any domain. Extensive evidence in

behavioural economics shows that to make efficient judgments and decisions under bounded

rationality, people rely on cognitive biases and heuristics. Heuristics are mental shortcuts

used by individuals to simplify complex decisions into easier to calculate operations (Tversky

B.4 BEM Discussion

& Kahneman, 1971, 1974), allowing people to make decisions quickly, in terms of computa-

tion time, and efficiently, in the use of information. Heuristics represent a departure from the

predictions of neoclassical economics, which assume that individuals are fully rational utility

maximisers who have limitless cognitive abilities and follow the laws of probability and

statistics. Instead, research has shown that individuals use information selectively, which may

lead to individuals systematically making sub-optimal decisions, known as cognitive biases

(see Dhami, 2016; Hastie & Dawes, 2010, for reviews). From the 33 studies identified in the

systematic review, 15 apply heuristics and biases to music related decisions, making this the

most prominent BEM area in our review. We organise the discussion of these studies into the

following sub-sections – judgement heuristics, processing fluency and framing effects.

Judgement heuristics

This section focuses on studies from the review that investigated the use of judgement heuris-

tics when making music-related decisions. Specifically, we discuss the following heuristics –

the

availability heuristic, the peak-end rule, the affect heuristic, and the representativeness

heuristic.

A number of the studies provide evidence that people apply heuristics when making musical

judgements recalled from memory, such as the availability heuristic and the peak-end rule.

The availability heuristic describes the tendency for individuals to judge the likelihood of an

event by the ease with which similar events can be brought to the mind (Tversky &

Kahneman, 1974). Vuvan, Podolak & Schmuckler (2014) examined whether the availability

heuristic can explain tonal expectancies in music memory. Participants were presented with

melodies followed by a test tone and subsequently asked to indicate whether the test tone

was present in the melody or not. The test tones were manipulated so they could either be

highly expected to be contained in the melody (the tone was related to the tonality or scale

of the melody), moderately expected or unexpected. The results indicated that participants

tended to falsely recall that the test tone was in the melody more frequently when it was

highly expected to be in the melody, consistent with the idea that such tones are more easily

‘available’ to the mind.

Rozin, Rozin & Goldberg (2004) investigated how listeners make affective judgements

about past music experiences. Participants listened to selections of music while providing

moment-to-moment emotional intensity ratings. The authors found that in line with the

peak-end rule (Fredrickson & Kahneman, 1993), participants judged the emotional intensity

of the music on how they felt at its the most intense point (the peak) and at its end, rather

than the total sum or average of every moment of the experience. Schäfer, Zimmermann &

B.4 BEM Discussion

Settelmeyer (2014) build upon this work and to account to even richer temporal profile of

emotional intensity. They find that although the average of all experienced moments was a

strong predictor of overall emotional intensity, the peaks and end play a significant role in

the evaluation of the musical experience.

Affective judgements of music have also been shown to be influenced by information pre-

sented with a song, even when identical songs have been played. Anglada-Tort, Steffans

& Müllensiefen (2018) examined whether judgements differed about the aesthetic value of

music (e.g., liking, beautiful, inspiring), when the emotionality of the song titles had been

manipulated to evoke feelings of positive, negative or neutral affect. They hypothesised that

participants would apply the affect heuristic – the tendency to rely on good/bad feelings ex-

perienced in relation to a stimulus (Slovic, Finucane, Peters & MacGregor, 2002). Consistent

with this, the negative titles were indeed found to have lower rating of aesthetic value.

A further characteristic of the affect heuristic is that it has been shown to influence both

the perceived benefits of an activity and risk perceptions from that activity, and that these

judgements are negatively correlated with each other, rather than being independent (Fin-

ucane, Alhakami, Slovic & Johnson, 2000). Watson, Zizzo & Fleming (2017) investigate

whether attitudes and behaviour towards music piracy can be explained by the affect heuristic.

Using an online longitudinal survey, the authors find that the perceived benefits of illegal

music-file sharing (e.g., financial benefits, ease of access) were negatively related to the

perceived risks (e.g., lawsuits against individuals), providing support for the affect heuristic.

Furthermore, perceived benefit rather than legal risk was found to be a better predictor of

actual file sharing behaviour, implying that a more effective route to tackling music piracy

ought to be providing services that give consumers the benefits of file sharing, rather than

upholding legal enforcement.

Lonsdale & North (2011) investigated whether music stereotypes (i.e., how people judge the

likely music taste of others) can be explained through the representativeness heuristic. The

representativeness heuristic is the tendency to judge the probability that a sample belongs to

a population by looking at the degree to which that sample resembles the population

(Tversky

& Kahneman, 1974).

The authors found that when asked to evaluate the music taste

of fictional

individuals, participants exhibited a common bias towards the music stereotype for that

individual, e.g., an individual described to engage in anti-social behaviour was more likely to

be attributed to liking hip-hop music. Importantly, participants’ judgements were highly

correlated with the perceived similarity of that individual to stereotypical music fans, rather

than the base-rate probability estimates for being a fan of that genre, giving increased support

for the representativeness heuristic in explaining this behaviour.

B.4 BEM Discussion

Processing fluency

Processing fluency refers to the subjective experience of ease with which people process

information. A key observation is that more fluent stimuli are often perceived as being

familiar and aesthetically pleasing compared to less-fluent stimuli (see Reber, Schwarz, &

Winkielman, 2004, for a review). In particular, factors thought to increase fluency such as

repeated exposure to a stimulus (Zajonc, 1968) and the amount of information represented

in a stimulus (Checkosky & Whitlock, 1973) can lead to more favourable evaluations. In

this section, we discuss studies identified in the systematic review that examined processing

fluency related to music.

Witvliet and Vrana (2007) investigated how repeated playing of music stimuli that varied by

emotional valence influenced participant liking of the music. The results indicated a polarised

response – repeated exposure to positive music led to increased liking for that music, whilst

repetition of the negative music led to increased disliking for that music. The authors also

measured physiological responses and find that certain facial muscles decreased in reactivity

after repeated exposure, providing evidence of increased ease of processing. Anglada-Tort

and Müllensiefen (2017) examined fluency effects through the repeated recording illusion –

the phenomenon which listeners perceive music stimuli to be different, whilst in fact they are

identical. The results also indicated a differential response with participants’ preferences for

music increasing with repeated listening only for familiar pop music and not for a relatively

unknown classical piece. The results from both studies suggest that whilst repeated exposure

appears to influence music preferences, factors associated with the listener experience such

as emotional connection and familiarity with the music seems important to uncover more

nuanced patterns in this relationship.

A series of studies have used properties of music stimuli as a way of investigating processing

fluency, such as altering the linguistic properties of music (linguistic fluency). Nunes,

Ordanini & Valsesia (2015) manipulated the amount of repetition in the lyrics of identical

songs in the lab and found that this was related to increased perceived familiarity. The authors

then examined the effect of song repetitiveness on a song’s popularity in the marketplace

using field data from the US singles chart. The results indicated that more repetitive songs

were more likely to be a number one hit as well as being faster climbers to the top of the

chart. Anglada-Tort et al. (2018) found that when manipulating the linguistic fluency of

artist names and song titles in a foreign language, identical music excerpts presented with

easy-to-pronounce names were preferred in terms of subjective value and aesthetic quality,

compared to excerpts presented with difficult-to-pronounce names. These results even held

B.4 BEM Discussion

for participants with high levels of music training, indicating that susceptibility to the effects

of processing fluency is not offset by increased knowledge of music.

Seror III and Neil (2003) examined whether a single target note could be detected within

a harmonic interval of several notes played simultaneously, for which the interval was either

consonant or dissonant. Consonant intervals (e.g., a perfect fifth) are associated with feeling

of pleasantness and agreeableness, whilst dissonant intervals (e.g., a tritone) with

unpleasantness and harshness. The authors hypothesised that since consonant intervals are

more likely to be fluent than dissonant intervals, the ability to identify the target note should

be easier. The results confirmed this hypothesis, showing faster and more accurate pitch

discrimination in consonant intervals.

Krishnan, Kellaris & Aurand (2012) examined the role of processing fluency in auditory

branding, a branch of branding that uses short periods of music or audio logos to convey core

brand values and prime brand recognition.

The authors investigated how the number of tones

(3,

6, 9 tones) in an audio logo affects participants’ willingness-to-pay (WTP) for the brand’s

associated product. They found that WTP varied with the number of tones and importantly,

that this variation was mediated by participants evaluations of fluency.

Finally, Huron (2013) provides a proposal of strategies that musicians can take in order to

maximise the overall hedonic effect of their performances. Specifically, the author applies

the findings that increased fluency through repeated exposure is likely to lead to favourable

evaluations amongst listeners; as well as the counter-effect of habituation – that successive

repetitions may become less novel to listeners and therefore lead to unresponsiveness. Three

main strategies are outlined (i) the trance strategy, which involves high levels of repetition

to induce positive feelings from increased fluency; (ii) the variation strategy, which allows

for fluency through repetition but with slight modifications to curb habituation; and (iii) the

rondo strategy, which involves reduced repetition and relies on the introduction of new music

material. Such an approach provides an insight into how processing fluency can be applied

by music performers.

Framing and loss aversion

Framing represents the systematic change in an individual’s decision when presented with

a choice in informationally equivalent ways (Tversky & Kahneman, 1981; see Kühberger,

1998, for a meta-analysis). When framing either highlights the positive or negative aspects

of the same decision, humans exhibit loss aversion. Research on loss aversion consistently

shows that humans prefer avoiding losses to acquire equivalent gains (Kahneman & Tversky,

B.4 BEM Discussion

1979). For example, people are more willing to take risks to avoid a loss than to make a gain

(Schindler & Pfattheicvher, 2016) and loss aversion can explain why penalty frames are

sometimes more effective than reward framesd in motivating people (Gächter et al., 2009).

In this section, we discuss studies identified in the systematic review that examined framing

effects in music decision-making.

A key finding in the literature is the observation of prestige-effects when evaluating musical

performance. Anglada-Tort and Müllensiefen (2017) found that listeners evaluated identical

recordings more positively (e.g., liking, quality, pitch and rhythm accuracy) when the music

was framed as being performed by a professional musician rather than performed by a

less skilled musician. Aydogan et al. (2018) further investigated the framing of musician

status on music evaluation by conducting a neuroimaging study. In addition to replicating

prestige-effects in music performance evolution, functional magnetic resonance imaging

(fMRI) indicated that higher activation in the ventromedial prefrontal cortex (vmPFC), a

region shown to play a key role in subjective value, is able to predict the magnitude of

this bias. Interestingly, for participants who preferred the student performance, increased

activation was observed in the dorsolateral prefrontal cortex (dlPFC), a region related to

cognitive control and deliberative effortful thinking. This suggests that these participants

were able to suppress the framing bias by exerting cognitive control.

Another area in which framing effects have been applied in music settings is adolescent

behaviour. North and Hargreaves (2005) investigated whether simply labelling music as

being harmful to young people lead to perceptions that it is harmful. Participants were

presented with identical pop songs either framed as “suicide-inducing” or “life-affirming”.

Evaluations showed that indeed the same piece of music were perceived as they were framed.

Bruijn, Spaans, Jansen & van’t Riet (2015) undertook a framing intervention study to

behaviours surrounding hearing loss prevention amongst adolescents. Young people recruited

from schools initially provided information on

their music listening

behaviour

and intentions

to listen to music at low volumes. Two weeks later, they were then asked the same questions

after

they had been exposed to persuasive messages about hearing loss, framed as a gain- frame

(positive consequences of listening to music at a reduced volume) vs. a loss-frame (negative

consequences of not doing so). The results indicated that the loss-frame was an effective

strategy to increase intentions and are consistent with loss aversion (Kahneman & Tversky,

1979), whereby risks framed as losses lead to behaviours to avoid this loss more

proportionately than when framed as gains.

Finally, in the context of online music steaming services, Li and Cheng (2014) examined the

factors that influence consumer intentions to switch from an advertising revenue model (free

B.4 BEM Discussion

content but subject to advertising) to a pay-for-content model (higher quality content with

no advertising but consumers pay a subscription). Drawing upon status quo bias theory i.e.,

the tendency to stay with the current option (Samuelson & Zeckhauser, 1988), the authors

identify loss aversion as a channel that affects switching intentions. Specifically, consumers

were concerned about the perceived sacrifices of leaving the current plan including the

monetary cost, the time and effort to switch and the risk that the new plan would not be

enjoyable.

Summary

The studies identified in this section suggest that when making judgements and decisions,

listeners are limited by their mental capacity (e.g., memory and emotion), time, and in-

formation available (e.g., song titles or descriptions about the performer). Consequently,

listeners rely on cognitive biases and heuristics that do not depend on the music stimuli

themselves. In particular, we identified four judgmental heuristics that readily apply to music

decision-making, allowing people to simplify complex decisions into easier-to-calculate

operations - i.e., the availability heuristic, peak-end-rule, the representativeness heuristic,

and the affect heuristic. Moreover, we found that fluency manipulations in music, such as

repetition and consonance, can influence music perception and, in turn, affect preferential

judgments.

Overall, all studies supported the fluency hypothesis, showing higher preferences

for

easier-to-process music compared to less-fluent stimuli. Finally, in line with framing,

several

studies found that contextual information often presented with music (e.g. song titles,

labels, or

the social status of the artist) has a significant impact on how listeners respond to it and

develop music preferences. Framing interventions can also be successfully applied to

prevent hearing loss amongst adolescents.

B.4.2 Social decision making

In 2007, the critically acclaimed band Radiohead surprised the music industry by offering

their new album “In Rainbows” as a digital download using a pay-what-you-want (PWYW)

agreement. Essentially, this meant that fans could pay as much as they liked for the album,

including a zero option. Although at odds with neoclassical economic theory, in which

consumers would download the album for free, fans actually made voluntary payments for

the album. One possible explanation for such generous payments under PWYW is that

individuals exhibit social preferences, i.e., they care about the preferences of others (see

Fehr & Schmidt, 2006 for a review). Another possibility is that individuals’ decisions are

highly influenced by the choices of their peers, also known as peer effects (Banerjee, 1992;

B.4 BEM Discussion

Bikhchandani, Hirschleifer & Welch, 1992). The example above suggets that music decision-

making does not happen in a vacuum, but it is influenced by the social world around us. In

this section, we discuss studies identified in the review that fall within the broader umbrella

of social decision making, including social preferences and peer effects.

Social preferences

We found a number of studies that examined the possible motivations behind consumer

behaviour under PWYW. Regner and Barria (2009) investigated whether voluntary payments

could be observed using field data for customers from an independent record label. Customers

were able to set the price they wanted to pay for an album within the range of 5−18, with

the label giving a recommendation of $8. The data indicated that around 85% of customers

chose to make payments that exceeded the minimum required payment, with the average

payment above the recommendation. The authors conjectured that since the label offers an

extensive try-before-you-buy service, reciprocity may be driving the generous payments.

More formally, they set up a behavioural game theory model to show how concerns for

reciprocity can switch behaviour from a selfish outcome, which customers simply offer

the minimum, to the more generous outcome, as observed in the data. In a follow up study,

Regner (2015) tested this theoretical prediction by testing the relationship between an

individual’s payments and survey data examining their motivations behind their decision. The

results indicated that reciprocity was indeed a driver for these payments.

In a controlled experiment, Waskow et al. (2016) compared WTP payments for albums

under PWYW vs. a traditional fixed-price. Although as expected, average WTP was lower

in the PWYW condition, these payments were significantly greater than zero. In addition,

the authors investigated whether these behavioural differences could be explained at the

neural level. Neuroimaging data revealed significant differences between the two conditions

with WTP only being related to neural activity in the fixed-price condition. Specifically,

correlations were found in the frontal brain regions - orbitofrontal cortex (OFC), medial

prefrontal cortex (mPFC), anterior cingulate cortex (ACC), areas linked to reward processing. No

such relationships were found in the PWYW condition suggesting a different mechanism

the neural level.

Harbi, Grolleau & Bekir (2014) examined the feasibility of a PWYW pricing strategy

from the point of the view of the artist. The authors propose a theoretical microeconomic

model to compare profitability under PWYW vs. a fixed-price scenario with/without piracy.

Importantly in the model, consumers gain procedural utility from buying music i.e., they care

not

only about their satisfaction from consuming music but also the conditions in which the

B.4 BEM Discussion

music is made, including the welfare of the artist. As a result, PWYW can be profitable for

the artist as it reduces piracy, promoting positive voluntary payments as well as increasing the

demand and prices for live performance through increased network size. A PWYW strategy

can also increase an artist’s profit share relative to the record label and may be useful as

bargaining tool when negotiating contracts.

Hashim, Kannan, Maximiano & Ulmer (2014) investigated piracy behaviour in a public

goods experiment framed in the context of music consumption. In this game, adolescent

participants decided whether to buy songs or to download them for free. If participants

are purely self-interested, then the game-theoretic prediction is maximum piracy, whereas

if they exhibit social preferences and care about the record label, this would lead to the

purchase of songs. In particular, the authors examined the effectiveness of different sources

of advice to reduce piracy, with the strength of the relationship between the participant and

the source of advice (i.e., the social tie) being manipulated. The results showed that advice

from parents, who had the highest social ties with the participants, was the most effective in

leading participants to pay for music.

Finally, Sonnabend (2016) looked at pricing decisions of music artists in the live music

industry. In a theoretical model, fans are concerned about the fairness of prices of live gigs,

such that if prices are above a reference price, they do not buy tickets. The model shows that

such concerns can be enough for the artist to keep prices rigid, even at times when there is

higher demand, for example on the weekends. Higher prices due to increased costs borne by

the artist are perceived as fair and are therefore tolerated.

Peer Effects

Berlin, Bernard & Fürst (2015) examined the effect of music ratings evaluated by peers on

teenage consumption choices of songs by bestsellers vs. new artists. The results indicated

that participants tended to imitate their peers with more listening time devoted to bestsellers

rather than new artists, thereby strengthening the superstar effect, in which relatively small

numbers of music artists dominate the industry.

Berns, Capra, Moore & Noussair (2010) found that information about song popularity through

website downloads led participants to change likability ratings of songs in the direction of the

popularity. In order to distinguish between competing mechanisms for how this information

affected music consumption, the authors analysed neuroimaging data for the participants.

The fMRI data supported the hypothesis that participants were motivated to change their

rating due to a desire for conformity rather than a change in the intrinsic value of the music.

B.4 BEM Discussion

Specifically, correlations between the tendency to change rating and neural activity were

found in the bilateral anterior insula and ACC, regions associated with negative feeling states,

suggesting that for these participants, the mismatch between their ratings and others’ ratings

may have led to cognitive/emotional dissonance that had to be resolved. In a follow up study,

Berns and Moore (2012) found that the same neuroimaging data were able to predict future

sales of these songs, with activity within the ventral striatum (associated with reward) being

correlated with future commercial success. We therefore see that in addition to the effect of

social information driving changes in music consumption at the individual level, the brain

responses of individuals are able to predict future commercial success at the population level,

insights that could be useful for neuromarketing and hit song science.

Summary

The studies outlined above suggest that music decision-making does not happen in a social

vacuum, but instead is largely influenced by social preferences and information. We found

evidence showing that PWYW is a viable strategy to sell music, offering several advantages

compared to traditional fixed-price strategies. Social preferences, such as reciprocity and guilt,

are important to understand consumers’ motivation to engage in different revenue models

for music consumption, including the success of PWYW. In addition, social preferences can

help better understand pricing strategies in the concert industry. Another important aspect

of social decision-making is social influence, such as peer effects. Several studies suggest

that peer effects can play a determinant role in music preferences and choices, which in turn

can influence the music market and determine outcomes such as the next successful artist or

hit song. We also found that interventions based on social information can be successfully

applied to reduce music piracy amongst adolescents.

B.4.3 Behavioural Time Preferences

Many decisions in music have a time dimension. For example, individuals have to consider

their future selves and preferences when buying music online, creating a playlist for a holiday

trip, or deciding to take music lessons. A significant amount of research in behavioural

economics has been devoted to decisions that have a time dimension (see Dhami, 2016, for

s review). A central aspect of this research is that individuals exhibit present-biased time

preferences, i.e., they have a strong preference for immediate gratification (O’ Donoghue &

Rabin, 1999). This desire can be so strong that it can lead the individual to alter a previously

made decision at a later point in time. In such cases, the standard exponential discounted

utility model is insufficient to capture these patterns of time inconsistency, and instead a

B.4 BEM Discussion

hyperbolic function is a more accurate representation. In this section, we discuss studies

identified in the systematic review that have applied behavioural time preferences to music

decision making.

Gans (2014) utilised behavioural time preferences to address an ongoing question in the

music industry: why are artists still entering the music industry if revenue from selling music

has decreased as a result of digital technology and piracy? He proposes a theoretical

microeconomic model, whereby artists face a dynamic trade-off between fame (the reward of

being supported by fans) and fortune (the revenue generated from music sales). So, although

artists may initially choose fame to build up a fan-base, they may choose to ‘sell-out’ in

the future to focus on financial rewards. Crucially, the model allows for time inconsistent

preferences such that when starting out, these artists under-weigh the idea that they will sell-

out in the future, and therefore are not deterred by the threat of lost revenue in the future due

to piracy when starting their careers.

De Bruijn et al. (2015) carried out a framing intervention study to examine behaviour sur-

rounding hearing loss prevention amongst adolescents (previously discussed in section 4.1.3).

In addition to finding that persuasive messaging framed as losses increased student intentions

to listen to music at low volumes, the study also investigated whether the temporal framing

of consequences (short vs. long term) would affect behaviour. The results indicated that

only

messages containing short-term consequences of loud music were effective in changing

listening

intentions. This suggest that the young people in the sample were susceptible to present bias

by overweighting immediate negative consequences and underweighting the long-term

consequences.

Several studies have examined time preferences for music consumption in term of its hedonic

value. i.e., based on the multisensory and emotional aspects of one’s experience. Charlton

and Fantino (2008) measured the discount rate of various commodities including music by

asking participants to choose between a given quantity of the commodity today vs. $100

after some delay. The authors found that time preferences for music fitted a hyperbolic

function well, indicating that the participants placed high value on immediate consumption

similar to other primary reinforcers such as food and drink. Kahneman & Snell (1999)

investigated people’s ability to forecast their future hedonic experiences from listening to

music. Participants listened to the same piece of music for 7 consecutive days after they had

given their predictions about how they would like the music at the beginning and end of the

week. In line with time inconsistency, the results indicated that the participants were poor at

hedonic forecasting, over-estimating the effect of repetition in reducing their future liking for

the

music. Finally, Kahnx, Ratner & Kahneman (1997) examined how individuals decide

B.4 BEM Discussion

which songs to play over a given period of time, such as when creating a playlist. When

making repeated choices between a liked song and less-preferred songs, the findings showed

that listeners do not always choose a song that maximised their enjoyment but instead opted

for less-preferred music in order to seek variety.

Summary

Overall, these studies show that behavioural time preferences can give a deeper insight into

how music is valued and consumed over time. We found evidence that time preferences when

consuming music follow a hyperbolic function and, therefore, consumers disproportionally

prioritise immediate benefits over future gains. This has implications for how consumers

demand music, particularly with the emergence of music streaming platforms providing music

instantaneously. Another important concept is time inconsistency, which was successfully

applied to model how artists’ decisions may shape the music market. Finally, we gained

valuable insights from studies focusing on time preferences for music consumption in terms

of its hedonic value. In particular, listeners’ ability to predict pleasure in their future music

consumption is rather low and when choosing music repeatedly over time, listeners do not

always choose music that maximizes their pleasure but instead seek variety, choosing to

listen to music that is less preferred.

B.4.4 Dual process theory

Dual process theories posit that there are two different modes of processing – an emotional

system and a cognitive system. The emotional system (system 1) is seen as fast, automatic

and unconscious, whilst the cognitive system (system 2) is seen as slow, deliberative and

conscious (see Evans, 2008; Frankish & Evans, 2009, for reviews). Within music, exploring

the interaction between emotional and cognitive processes could prove particularly insightful

when analysing musical performance, e.g., the extent to which performers rely on conscious

vs. unconscious decisions and the factors that can affect this. In this section, we review studies

identified that have examined dual process theories with regard to musical performance.

Bangert, Fabian, Shubert and Yeadon (2014) undertook a case study in which performance

data and retrospective accounts were taken from an expert cellist performing a piece of

familiar music, with the goal of understanding which music decisions were intuitive and

which were deliberate. The results indicated that out of the 134 music decisions made, 65%

were categorised as deliberate. In a second study, Bangert, Schubert, and Fabian (2015)

applied the same method using a small sample of professional violinists, but with an

unfamiliar piece of music. This time, 82% of music decisions were categorised as intuitive.

B.4 BEM Discussion

These results suggest that familiarity of the music plays a key role in determining whether

performers apply system 1 or system 2 processing.

Another factor thought to be important in the interaction between system 1 and system 2

processing is expertise of the performer. Bangert, Schubert, and Fabian (2014) outline the

‘spiral’ model – a visual representation to describe how the gaining of expertise can affect

the relative proportions of intuitive and deliberate decision-making of a performer. When

a novice starts, they rely on intuition stemming from less developed knowledge (immature

intuition) and performance is likely to contain guesses and mistakes. After practising, the

performer increases their knowledge and moves towards a process of greater deliberation

based upon more informed decisions. With even more practise, the performer returns to

intuitive processing, but this now comes from highly developed knowledge, so that deliberate

decisions have become automatic (mature intuition). As the performer encounters new music

problems, each iteration between intuitive and deliberate decision-making contains less

mistakes with the performer having greater control, represented as upward movements along

a continually narrowing spiral. Rosen et al. (2016) tested whether expertise can moderate the

effect of increased system 2 processing on the quality of jazz improvisations. In a neuro study,

transcranial direct current stimulation (tDCS) was applied to the dlPFC (a region related

to deliberative thinking) of jazz pianists of varying expertise. Results indicated that brain

stimulation increased the performance quality for less experienced musicians, but hindered

the performance quality for expert musicians. These finding are consistent with the model

above suggesting that novices benefit from increased top-down control, where as experts

would benefit from increased system 1 processing.

Summary

The studies identified in this section show that exploring the interaction between System-1

and System-2 processes in the brain can help increase our understanding of how musicians

make decisions while performing, as well as the what is the role that music expertise. While

the studies by Bangert and colleagues provide a useful starting point to examine dual-process

theory in music performance using qualitative methods, the study by Rosen et al. (2016)

highlights the benefits of using neurostimulation techniques to examine more systematically

the two systems of processing in music making and creativity.

B.5 Future Directions

B.5 Future directions

Having demonstrated the value of behavioural economics to music-decision research using

the current literature, we now turn to our second objective of exploring how behavioural

economics can inform future research. Again, to guide our discussion we use the BEM areas

categorised from the systematic review. In the final part of this section, we introduce the

BEM research programme and show how it can be applied practically in the context of a

real-world music issue.

Heuristics in Music Performance

The area of heuristics and biases was found to be the most prevalent research topic within the

review, focusing almost exclusively on issues related to music preferences and consumption.

Surprisingly, however, we found no studies that applied heuristics to music composition and

improvisation choices. Given the highly demanding nature of music improvisation, including

both cognitive and bodily limitations (Ashley, 2016), it seems likely that musicians rely on

fast and frugal heuristics to simplify complex decisions while improvising. A recent study

by Beaty et al. (2020) provides some initial evidence for this, demonstrating that eminent

jazz musicians tend to start their solo improvisations with music sequences that are

melodically simpler before creating more complex sequences, the so-called ‘easy-first’ bias.

We hypothesise that musicians may be relying on other heuristics in their performances, some

which have not yet been studied in the music domain at all. One such candidate is the

anchoring heuristic (Tversky & Kahneman, 1974), whereby musicians during a performance

rely

on initial musical choices (the anchor) to inform future decisions. At this point, we note the

link between heuristics and the two-systems view from dual process theory (Kahneman,

2011). If performers are using heuristics and applying System 1 processing to save on

cognitive effort, then this may be observable at the neural level. In this regard, we encourage

further use of the research methods employed in the field of improvisation neuroscience (see

Beaty, 2015, for a review), in order to gain a deeper understanding of the neural processes

that underpin heuristics in music composition and performance.

Cognitive Biases in Music Consumption

A potentially fruitful application of cognitive bias research is choice overload. The dramatic

increase in recent years of music streaming services (e.g., Spotify, YouTube) provides lis-

teners with a large assortment of songs instantly. However, listeners may not necessarily be

benefiting from this vast amount of choice. In fact, much evidence indicates that providing

individuals with variety can lead to negative outcomes including choice deferral, choice rever-

B.5 Future Directions

sal, reverting to the default option, and overall lower satisfaction (see Chernev, Böckenholt,

& Goodman, 2015, for a review). Despite voluminous research on choice overload, there

has been little application to music listening behaviour. One promising avenue of research

is provided by Ferweda, Yang, Schedl, & Tkalcic (2019), who found that music expertise

moderates the relationship between how music is organised on streaming platforms (e.g.,

mood, genre, activity) and preferences for choice set size. For example, when presented with

music organised by mood, participants with more music expertise preferred a system with

fewer choices; but when presented with music organised by genre, these participants

preferred a system with more choices. We see scope to build on this work to understand more

fully how individual differences influence music taxonomy choices, minimising the adverse

effects of choice overload and improving the user experience from music streaming services.

Social Decision Making and Piracy

As discussed in the review, social preferences can lead individuals to make voluntary pay-

ments for music, whilst peer effects can influence an individual’s consumption choices. We

see benefit in combining these areas of research in order to examine ways to reduce music

piracy. For example, a robust finding in laboratory experiments is that many people are

“conditional co-operators”, whose pro-social behaviour is sensitive to observations of others’

behaviour (see Chaudhury, 2011, for a review). Exploring conditional cooperation in a music

setting may therefore hold the key to increasing compliance associated with music payments.

In addition, to further understand the dynamics of music consumption and emergence of

social norms in a more naturalistic environment, we encourage the use of social network

experiments (see Hawkins, Goodman, & Goldstone, 2019, for a review).

Such methods could

used to model the complex cognitive processes involved in music consumption, such as

learning, social coordination, and cultural transmission.

Behavioural Game Theory and Hit Song Science

We see a great amount of potential in applying behavioural game theory further to music-

decision research. One area of advance is hit song science, a field aimed at predicting song

success before market release. Traditionally, attempts at predicting song popularity have only

considered the intrinsic properties of the music itself, but given that social factors have a

substantial influence on market outcomes, this approach has been unsuccessful (see Pachet,

2012). One novel way to model social influence is to use Level-k and cognitive hierarchy

models (Camerer, Ho, & Chong, 2004; Stahl & Wilson, 1994, 1995). In such models, there

is a hierarchy about what players believe about the actions of other players. For example, a

B.5 Future Directions

well-known application for which Level-k modelling has made accurate predictions about

behaviour is the beauty contest (Keynes, 1936; Nagel, 1995). Here, we consider a music

adaptation, where individuals have to guess the song which corresponds to the average

preference of the competition. Whilst Level-0 individuals may simply choose their favourite

song, Level-1 players will choose the song that they believe the majority of Level-0 players

will choose, and Level-2 players will choose a song incorporating their beliefs about Level-1

players, and so on. Understanding how individuals form beliefs in music prediction markets

and how these beliefs are affected by others could give greater insight into how songs become

popular, potentially increasing prediction accuracy in hit song science.

Behavioural Time Preferences in Music Education

The review indicated that present-biased preferences can lead an individual to reverse a

previously made decision at a later point in time. Since time-inconsistent preferences are

often detrimental to the long-term interest of the individual, a substantial amount of the

literature has been focused on self-control (see Steel 2007, for a review). An area of music

research where such insights could prove to be beneficial is motivation in music education.

Although learning a musical instrument can be a personally satisfying and meaningful activity,

requires considerable effort in the form of regular practice. For many music students, this

may be difficult due to a lack of intrinsic motivation, belief in their competence, or reaction

to the learning environment (see Renwick & Reeve, 2012, for a review). Therefore, for

impatient students, the short-term temptation of not practising may be more desirable than

the long-term goal. Here we offer two proposals. First, we recommend that music education

researchers incorporate theoretical frameworks used in behavioural economics to model

time preferences, e.g., procrastination models to measure the extent to which individuals are

present-biased as well as how aware they are of their self-control problems (O’Donoghue

& Rabin, 1999). This could allow music researchers to gain a better understanding of how

individuals vary in their music-related motivation, and to investigate whether such model

parameters can be reliable trait markers of the efficacy of music learning (see Peters & Büchel,

2011, for a review). Second, to help improve motivation in music practice, we propose the

application of interventions successfully applied in other domains of self-control, such as

episodic future thinking and pre-commitment devices (e.g., Ariely & Wertenbroch, 2002;

Koffarnus, Jarmolowicz, Mueller, & Bickel, 2014).

B.5 Future directions

The BEM Research Programme

From our discussion of the extant literature and explorations of future work, we have

demonstrated the benefits of applying behavioural economics to music-decision research.

Specifically, we have shown that by relaxing the rationality assumptions of homo economicus

and drawing from interdisciplinary insights, researchers are able to follow an approach that

is both empirically supported and wider in scope. Furthermore, incorporating models from

behavioural economic theory, based upon principles of optimisation and underpinned by

axiomatic foundations, provides an internally consistent framework to work within. Here,

we propose the BEM research programme – an integrated research agenda that utilises the

behavioural economics toolkit in music-decision making. We emphasise that this programme

is not limited to using the BEM research areas only discussed in this review nor are they

required to be applied in isolation of each other. Instead, we see the BEM as a holistic

framework to study music-related behaviour. As an example, below we show how several

areas of the BEM can be applied practically to address a real-world music issue.

One area of concern amongst classical music organisations is the lack of socio-economic

diversity in the audience for classical music concerts, especially from young people and

ethnic minorities (see Chan, Goldthorpe, Keaney, Oskala, 2008; DiMaggio & Mukhtar,

2004; Kolb, 2001). Barriers to attendance often cited include perceived lack of knowledge,

feeling of not belonging to the community, and the desire for more social interaction at

concerts (Dearn and Price, 2016; Dobson & Pitts, 2011; Kolb, 2000). Applying audience

development strategies based on a combination of areas within behavioural economics

may help to reduce these barriers. These include measures aimed to change social norms

associated with classical concerts (e.g., more accessible music venues, relaxation of dress

code, promotion of a social community through increased performer/audience interactions);

framed advertising to actively challenge music stereotyping appealing to minority groups;

and the use of social networks to encourage positive peer effects. Furthermore, whilst there

has been some limited discussion about the perceived risk associated with the decision to

attend concerts (Baker, 2000; Price, 2017), this area could be developed by applying theories

reference-dependent

utility

(Kahneman

and

Tversky,

1979;

Ko˝szegi

Rabin,

2006).

Here,

risk attitudes of attenders can be captured relative to a reference point, e.g., expectations of

enjoyment, and could be particularly useful to model behaviour of new-attenders who may

differ in their expectations to those who attend concerts regularly.

We hope that from the

examples given here, this discussion has provided valuable stimulation

of ideas about the future

potential of applying behavioural economics to music research.

B.7 References

B.6 Conclusion

Departing from historically disconnected research programmes in psychology and economics,

this paper has discussed the benefits of using behavioural economics in music-decision

research. Our contributions to the literature are two-fold. First, through a systematic literature

review, we identified 33 studies that applied behavioural economics to music-related decision

making, categorised within four research areas – heuristics and biases, social decision making,

behavioural time preferences, and dual-process theory. These studies utilised theoretical and

empirical tools of behavioural economics, covering a wide area of music research, including

music consumption and piracy, music preferences, music performance, music perception

and memory, and music and health. Second, based upon the findings of our set of identified

studies, we discussed how behavioural economics can help develop new avenues of research.

Based on these objectives, we proposed the Behavioural Economics of Music (BEM), an

interdisciplinary research programme positioned at the intersection of music, psychology, and

economics. We are truly excited about such a programme, and we hope that this discussion

has stimulated interest in the potential of using behavioural economics to address key issues

in music-decision research. Finally, we note that research in music-related decision making

is still relatively young, with no specific research field dedicated to this area exclusively. Our

proposal of the BEM provides the first steps towards this.

B.7 References

Anglada-Tort, M., & Müllensiefen, D. (2017). The repeated recording illusion: the effects of

extrinsic and individual difference factors on musical judgments. Music Perception: An

Interdisciplinary Journal, 35(1), 94–117.

Anglada-Tort, M., Steffens, J., & Müllensiefen, D. (2018). Names and Titles Matter: The Im-

pact of Linguistic Fluency and the Affect Heuristic on Aesthetic and Value Judgements

of Music. Psychology of Aesthetics, Creativity, and the Arts, 13(3), 277-292.

B.7 References

Angner, E. (2012). A course in behavioral economics. Macmillan International Higher

Education.

Austin, J. R., Renwick, J. M., & McPherson, G. E. (2006). Developing motivation. In G.

E. McPherson (Ed.), The child as musician: A handbook of musical development (pp.

213–238). Oxford, UK: Oxford University Press.

Ball, L. J., & Thompson, V. A. (Eds.). (2017). International handbook of thinking and

reasoning. Routledge.

Bangert, D., Schubert, E., & Fabian, D. (2015). Practice thoughts and performance action:

Observing processes of musical decision-making. Music Performance Research, 7,

27–46.

Bangert, D., Schubert, E., & Fabian, D. (2014). A spiral model of musical decision-making.

Frontiers in psychology, 5, 320

Berlin, N., Bernard, A., & Fürst, G. (2015). Time spent on new songs: word-of-mouth and

price effects on teenager consumption. Journal of Cultural Economics, 39(2), 205–218.

Berns, G. S., & Moore, S. E. (2012). A neural predictor of cultural popularity. Journal of

Consumer Psychology, 22(1), 154–160.

Berns, G. S., Capra, C. M., Moore, S., & Noussair, C. (2010). Neural mechanisms of the

influence of popularity on adolescent ratings of music. NeuroImage, 49(3), 2687–2696.

Burke, A. E. (1996). How effective are international copyright conventions in the music

industry? Journal of Cultural Economics, 20(1), 51–66.

Byun, C. H. C. (2016). The economics of the popular music industry. Springer.

Cameron, S. (2015). Music in the marketplace: a social economics approach. Routledge.

Cameron, S. (2016). Past, present and future: music economics at the crossroads. Journal of

Cultural Economics, 40(1), 1–12.

Cartwright, E. (2018). Behavioral Economics. Routledge.

Charlton, S. R., & Fantino, E. (2008). Commodity specific rates of temporal discounting:

Does metabolic function underlie differences in rates of discounting? Behavioural

Processes, 77(3), 334–342.

Dawes, R., & Hastie, R. (2010). Rational choice in an uncertain world: The psychology of

judgment and decision making. Sage Publications.

De Bruijn, G. J., Spaans, P., Jansen, B., & Van’t Riet, J. (2016). Testing the effects of a mes-

sage framing intervention on intentions towards hearing loss prevention in adolescents.

Health Education Research, 31(2), 161–170.

Decrop, A., & Derbaix, M. (2014). Artist-Related Determinants of Music Concert Prices.

Psychology and Marketing, 31(8), 660–669.

Deutsch, D. (2013). Psychology of music. Elsevier.

B.7 References

Dhami, S. (2016). The foundations of behavioral economic analysis. Oxford University

Press.

Elliott, C., & Simmons, R. (2011). Factors determining UK album success.Applied Eco-

nomics, 43(30), 4699–4705.

Gans, J. S. (2015). “Selling Out” and the impact of music piracy on artist entry. Information

Economics and Policy, 32, 58–64.

Greb, F., Schlotz, W., & Steffens, J. (2018). Personal and situational influences on the

functions of music listening. Psychology of Music, 46(6), 763–794.

Hallam, S., Cross, I., & Thaut, M. (2016). Oxford handbook of music psychology (second

edition). Oxford University Press.

Hantula, D. A., Brockman, D. D. C., & Smith, C. L. (2008). Online shopping as foraging:

The effects of increasing delays on purchasing and patch residence. IEEE Transactions

on Professional Communication, 51(2), 147–154.

Hantula, D. A., & Bryant, K. (2005). Delay discounting determines delivery fees in an e-

commerce simulation: A behavioral economic perspective. Psychology and Marketing,

22(2), 153–161.

Hashim, M. J., Kannan, K. N., Maximiano, S., & Ulmer, J. R. (2014). Digital Piracy, Teens,

and the Source of Advice: An Experimental Study. Journal of Management Information

Systems, 31(2), 211–244.

Hendricks, K., & Sorensen, A. (2009). Information and the Skewness of Music Sales.

Journal of Political Economy, 117(2), 324–369.

Hiller, R. S. (2016). The importance of quality: How music festivals achieved commercial

success. Journal of Cultural Economics, 40(3), 309–334.

Holt, F. (2010). The economy of live music in the digital age. European Journal of Cultural

Studies, 13(2), 243–261.

Holyoak, K. J., & Morrison, R. G. (2012). The Oxford Handbook of Thinking and Reasoning.

Oxford University Press.

Huron, D. (2013). A Psychological Approach to Musical Form: The Habituation–Fluency

Theory of Repetition. Current Musicology, 96(96), 7–35.

IFPI. (2019). International Federation of the Phonographic Industry (IFPI) Global Music

Report 2018. Retrieved from http://www.ifpi.org/downloads/GMR2019.pdf

Impett, J. (2016). Making a mark: the psychology of composition. In S. Hallam, I. Cross, &

M. Thaut (Eds.), The oxford handbook of music psychology. Oxford University Press.

Jabbar, H. (2011). The behavioral economics of education: New directions for research.

Educational Researcher, 40(9), 446–453.

B.7 References

Kahn, B., Ratner, R., & Kahneman, D. (1997). Patterns of Hedonic Consumption Over Time.

Marketing Letters, 8(1), 85–96.

Kahneman, D. (2003). Maps of bounded rationality: Psychology for behavioral economics.

American Economic Review, 93(5), 1449–1475.

Kahneman, D. (2011). Thinking , Fast and Slow. Macmillan.

Kahneman, D., & Snell, J. (1992). Predicting a changing taste: Do people know what they

will like? Journal of Behavioral Decision Making, 5(3), 187–200.

Ko, T. H., & Lau, H. Y. K. (2016). A Decision Support Framework for Optimal Pricing and

Advertising of Digital Music as Durable Goods. IFAC-PapersOnLine, 49(12), 277–282.

Koffarnus, M. N., Jarmolowicz, D. P., Mueller, E. T., & Bickel, W. K. (2013). Changing

delay discounting in the light of the competing neurobehavioral decision systems theory:

a review. Journal of the experimental analysis of behavior, 99(1), 32-57.

Krishnan, V., Kellaris, J. J., & Aurand, T. W. (2012). Sonic logos: Can sound influence

willingness to pay? Journal of Product and Brand Management, 21(4), 275–284.

Krueger, A. B. (2005). The economics of real superstars: The market for rock concerts in the

material world. Journal of Labor Economics, 23(1), 1–30.

Lamont, A., & Greasley, A. (2016). Musical preferences. In S. Hallam, I. Cross, & M. Thaut

(Eds.), The oxford handbook of music psychology. Oxford University Press.

Lamont, A., Greasley, A., & Sloboda, J. (2016). Choosing to Hear Music. In S. Hallam,

I. Cross, & M. Thaut (Eds.), The oxford handbook of music psychology (pp. 1–20).

Oxford University Press.

Li, Z., & Chen, Y. (2014). From Free To Fee: Exploring the Antecedents of Consumer.

Journal of Electronic Commerce Research, 15(4), 281–300.

Liebowitz, S. J. (2004). Will MP3 downloads annihilate the record industry? The evidence

so far. Advances in the Study of Entrepreneurship, Innovation, and Economic Growth,

15, 229–260.

Liebowitz, S. J. (2006). File sharing: Creative destruction or just plain destruction? Journal

of Law and Economics, 49(1), 1–28.

Linnemann, A., Wenzel, M., Grammes, J., Kubiak, T., & Nater, U. M. (2018). Music

listening and stress in daily life—a matter of timing. International Journal of Behavioral

Medicine, 25(2), 223–230.

Lonsdale, A. J., & North, A. C. (2012). Musical taste and the representativeness heuristic.

Psychology of Music, 40(2), 131–142.

Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for

systematic reviews and meta-analyses: the PRISMA statement. Journal of Clinical

Epidemiology, 62(10), 1006–1012.

100

B.7 References

Mortimer, J. H., Nosko, C., & Sorensen, A. (2012). Supply responses to digital distribution:

Recorded music and live performances. Information Economics and Policy, 24(1), 3–

14.

Nunes, J. C., Ordanini, A., & Valsesia, F. (2014). The power of repetition: Repetitive lyrics

in a song increase processing fluency and drive market success. Journal of Consumer

Psychology, 25(2), 187–199.

North, A. C., & Hargreaves, D. J. (2008). The social and applied psychology of music. OUP

Oxford.

Oberholzer-Gee, F., & Strumpf, K. (2009). File sharing and copyright. Innovation Policy

and the Economy, 10, 19–55

Ogaski, & Tanaka. (2017). Behavioral Economics: toward a new economics by integration

with traditional economics. Springer.

Palazzi, A., Wagner Fritzen, B., & Gauer, G. (2019). Music-induced emotion effects on

decision-making. Psychology of Music, 47(5), 621–643.

Rayna, T., & Striukova, L. (2009). Monometapoly or the Economics of the Music Industry.

Prometheus, 27(3), 211–222.

Regner, T., & Barria, J. A. (2009). Do consumers pay voluntarily? The case of online music.

Journal of Economic Behavior and Organization, 71(2), 395–406.

Regner, T. (2015). Why consumers pay voluntarily: Evidence from online music. Journal of

Behavioral and Experimental Economics, 57, 205–214.

Renwick, J. M., & Reeve, J. (2012). Supporting motivation in music education. In G. E.

McPherson & G. F. Welch (Eds.), Oxford Handbook of Music Education (Vol. 1, pp.

143–162). New York: Oxford University Press.

Reynolds, B. (2006). A review of delay-discounting research with humans: relations to drug

use and gambling. Behavioural pharmacology, 17(8), 651-667.

Schäfer, T., Zimmermann, D., & Sedlmeier, P. (2014). How we remember the emotional

intensity of past musical experiences. Frontiers in Psychology, 5(AUG), 1–10.

Simon, H. A. (1955). A Behavioral Model of Rational Choice. The Quarterly Journal of

Economics, 69(1), 99–118.

Simon, H. A. (1982). Models of Bounded Rationality. MIT press.

Slovic, P., Finucane, M., Peters, E., & MacGregor, D. G. (2002). Rational actors or rational

fools: Implications of the effects heuristic for behavioral economics. Journal of Socio-

Economics, 31(4), 329–342.

Sonnabend, H. (2016). Fairness constraints on profit-seeking: evidence from the German

club concert industry. Journal of Cultural Economics, 40(4), 529–545.

101

B.7 References

Strobl, E. A., & Tucker, C. (2000). The dynamics of chart success in the U.K. pre-recorded

popular music industry. Journal of Cultural Economics, 24(2), 113–134.

Sweeting, A. (2013). Dynamic Product Positioning in Differentiated Product Markets: The

Effect of Fees for Musical Performance Rights on the Commercial Radio Industry.

Econometrica, 81(5), 1763–1803.

Tahler, R. (2015). Misbehaving: The Making of Behavioral Economics. WW Norton.

Tan, S., Pfordresher, P., & Harré, R. (2017). Psychology of music: From sound to significance.

Routledge.

Tschmuck, P. (2017). The economics of music. Agenda Publishing.

Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and

probability. Cognitive Psychology, 5(2), 207–232.

Tversky, A., & Kahneman, D. (1974). Judgment under Uncertainty: Heuristics and Biases.

Science, 185(4157), 1124–31. D

Varian, H. R. (2005). Copying and copyright. Journal of Economic Perspectives, 19(2),

121–138.

Vuvan, D. T., Podolak, O. M., & Schmuckler, M. A. (2014). Memory for musical tones: The

impact of tonality and the creation of false memories. Frontiers in Psychology, 5(JUN),

1–

18.

Waskow, S., Markett, S., Montag, C., Weber, B., Trautner, P., Kramarz, V., & Reuter, M.

(2016). Pay What You Want! A Pilot Study on Neural Correlates of Voluntary Payments

for Music. Frontiers in Psychology, 7.

Wilkinson, N., & Klaes, M. (2017). An introduction to behavioral economics. Macmillan

International Higher Education.

Witvliet, C. V. O., & Vrana, S. R. (2007). Play it again Sam: Repeated exposure to emo-

tionally evocative music polarises liking and smiling responses, and influences other

affective reports, facial EMG, and heart rate. Cognition and Emotion, 21(1), 3–25.

Appendix C

The repeated recording illusion (S3)

This is an Accepted Manuscript of an article published by UC Press in Music Perception on

24th February 2017, available online: https://doi.org/10.1525/mp.2017.35.1.94. The paper

is not the copy of the record and may not exactly replicate the authoritative document

published in the journal. For presentation in this thesis, the appendices of the paper have

been removed and the passages referring to each Appendix in the text modified to indicate

where to find the materials online. Moreover, there may be minor modifications in the text

to guarantee a consistent typographic style throughout the thesis, such as the position of

figures and tables. Please do not copy or cite without author’s permission.

Citation

Anglada-Tort, M., & Müllensiefen, D. (2017). The repeated recording illusion: the effects

of extrinsic and individual difference factors on musical judgments. Music Perception: An

Interdisciplinary Journal, 35(1), 94-117. DOI: https://doi.org/10.1525/mp.2017.35.1.94

Author contribution

The experiment presented in the paper was conducted during my master thesis in the MSc

in Music, Mind, and Brain, at Goldsmiths, University of London (2015-2016). The paper

was written after completing my masters and published during the first year of my PhD at

Technicshce Universität Berlin. Prof. Dr. Daniel Müllensiefen (Goldsmiths, University of

London) supervised this work at all stages.

The repeated recording illusion: the effects of extrinsic and

individual difference factors on musical judgments

The repeated recording illusion refers to the phenomenon in which listeners are under the

impression that they hear different musical stimuli while they are in fact identical. This

phenomenon has not yet been studied systematically. Thus, the present paper aims to

construct an experimental paradigm to enable the systematic measurement of the repeated

recording illusion, investigating individual difference factors that contribute to it as well as

extrinsic factors responsible for differences in musical judgements when the acoustic input

remains the same. Seventy-two participants were misled to think that they had heard three

different musical performances of an original piece when in fact they were exposed to the

same repeated recording. Each time, the recording was accompanied by a different text

suggesting a low, medium or high prestige of the performer. 75 % of the participants were

under the impression of hearing different musical performances. High levels of neuroticism

and openness made it significantly more likely that an individual would fall for the illusion.

Musicians were not any more or any less susceptible to the illusion than non-musicians.

For participants who fell for the illusion, the explicit prestige texts influenced evaluations of

the music significantly. In addition, the mere repetition of the stimulus showed a partial

effect. These results suggest that musical judgements are sometimes not based on musical

cues and features but are influenced by factors that do not depend on the music itself. The

repeated recording illusion can constitute a paradigm for investigating psychological biases

and individual differences in aesthetic and musical judgements because the illusion allows

for the study of their effects while the music remains the same. Results are interpreted within

Tversky and Kahneman’s framework of judgements and decision-making.

Keywords: aesthetics, individual differences, explicit information, music performance, judge-

ments and preferences.

104

C.1 Introduction

In 1977, the German radio station WDR 3 conducted an audience participation experiment

during a live programme (see the description in Behne, 1987). The radio broadcaster misled

the audience to think that they would hear three different performances of the same excerpt

of Bruckner Symphony No. 4, providing brief information about three different conductors

(Karl Böhm, Leonard Bernstein, and Herbert von Karajan) just before each recording was

played. However, the radio broadcaster played the same recording three times. The radio

station received 536 calls. 81.7 % of the callers were misled and reported differences between

the identical music recordings. Only the remaining 18.3 % of the listeners who called in

reported that there were no differences between the three performances. Nevertheless, we

note that the audience participation experiment had several shortcomings, such as a lack of

control over experimental conditions and a potential sampling bias for those listeners who

believed they had heard different musical performances to call the radio station. Therefore,

one of the main motivations of the present paper was the replication of this phenomenon in

an experimental setting.

We will refer to this phenomenon, where listeners are under the impression that they hear

different musical performances while in fact they are identical, as the repeated recording

illusion. Duerksen (1972) was amongst the first academic studies to use a similar approach.

He played two tape recordings of an identical piano performance to music major and non-

music major students. Participants were told that one performance was by an eminent

professional pianist and the other one by a student. Both groups rated technical and musical

characteristics of the music recording consistently lower when told the performance was by a

student than when told it was by a professional. However, Duerksen (1972) merely attributed

the

findings to an effect of expectations and did not investigate whether participants believed that

they had heard the same or different musical performances.

There are a number of studies that used similar experimental paradigms, presenting partici-

pants with identical recordings in succession (Behne & Wöllner, 2011; Cavitt, 1997, 2002;

Elliott, 1995; Griffiths, 2008; Juchniewicz, 2008; Radocy, 1976; Silvey, 2009). The main

purpose of these studies was to investigate non-musical factors that influence evaluations

of musical performances, such as the effect of expectations (Cavitt, 1997, 2002; Duerksen,

1972), authority (Radocy, 1976), musicians’ body movements (Behne & Wöllner, 2011;

Juchniewicz, 2008), race and gender (Elliott, 1995), concert dress and physical attractiveness

(Griffiths, 2008), and band labels (Silvey, 2009). None of these studies considered the

implications of participants potentially falling for the repeated recording illusion. Thus, in

none of these studies it is possible to determine whether the illusion occurred in the sample of

105

C.1 Introduction

participants. We considered the repeated recording illusion to be a phenomenon that merits

further investigation. Exploring this phenomenon in detail could provide relevant and unique

insights to the fields of aesthetics, music perception, cognition, and choice behaviour. There-

fore, the present study attempts to measure systematically the repeated recording illusion,

investigating individual difference factors that contribute to it as well as extrinsic factors

responsible for differences in musical judgements when the acoustic input remains the same.

In relation to the individual difference factors, we suggest that the amount of musical training

of participants may play an important role in the repeated recording illusion. A large number

of previous studies have shown that people with high levels of musical training (i.e.,

musicians) outperform non-musicians on many music-related tasks, indicating that musical

training has a positive influence on the efficiency and accuracy with which characteristics

of sounds (e.g., pitch and timbre) are encoded in memory (see Pearce, 2015 for a review).

For instance, musicians show greater sensitivity to fine variations and nuances in music (e.g.,

slurs, rests, articulation, and timbre) (Deliege, 1987) and better recognition memory for

melodies than non-musicians (Dowling & Bartlett, 1981; Dowling, 1978; Halpern, Bartlett, &

Dowling, 1995; Orsmond & Miller, 1999). We therefore hypothesized that musical training

would have an effect on the illusion. However, the tasks involved in the above research (e.g.,

to recognize a melody) are very different to the task that requires an individual to realize that

the same music recording is played in succession. Thus, it is difficult to predict the direction

in which musical training may affect the repeated recording illusion. The present

study only

attempts to assess whether musicians perform differently on this task compared to

non-

musicians.

Arguably, the paradigm used in the repeated recording illusion relies on a judgement bias

exerted by a figure of authority (i.e., participants are told by a researcher in a lab condition

that they will listen to different performances). In line with Milgram’s obedience to authority

experiment (1963), Radocy (1976) found that the bias exerted by a figure of authority

significantly influenced participants’ evaluations of musical events. We therefore considered

that

individual differences on suggestibility could be an important factor contributing to the

illusion. We hypothesized that people with higher levels of susceptibility would be more

likely to fall for the repeated recording illusion.

The present research also explored music preferences and personality as possible individual

difference factors related to the illusion. Individuals tend to have stronger preferences for

certain genres of music, becoming more familiar with the preferred style as a result of

repeated listening. Repeated exposure to a piece of music increases the liking for it and

decreases its subjective complexity (see North & Hargreaves, 2008 for a review). In relation

106

C.1 Introduction

to personality, research shows that personality traits relate to specific preferences for music

styles (see Greasley & Lamont, 2016 for a review). For instance, openness to experience

is positively linked to preference for reflective and complex styles (e.g., classical music)

(Rentfrow & Gosling, 2003). Furthermore, research on individual differences has found links

between personality and suggestibility, showing for example a positive (but low) relationship

between suggestibility and neuroticism (see Gudjonsson, 2003 for a review). Therefore, we

hypothesized that preferences for music style and personality traits would affect participants’

susceptibility to the repeated recording illusion, although we could not specify in which

direction.

Extrinsic factors that may be responsible for differences in musical judgements when the

acoustic input is identical include the effect of explicit information. Presenting music with

explicit information has been shown to be influential in the evaluation of musical perfor-

mances (Cassidy & Sims, 1991; Cavitt, 1997, 2002; Kroger & Margulis, 2016; Margulis,

2010; Margulis, Kisida, & Greene, 2015; North & Hargreaves, 2005; Silveira & Diaz, 2014;

Silvey, 2009; Vuoskoski & Eerola, 2013). In an fMRI study, Kirk, Skov, Hulme, Christensen,

and Zeki (2009) presented the same images of artworks with different contextual information,

varying in prestige (i.e., labelled as ‘gallery’ or ‘computer generated’). The findings revealed

that

when the artworks were labelled as ‘gallery’ they were rated higher in an aesthetic value scale

than when labelled as ‘computer generated’. The fMRI data showed more activity in the

medial orbitofrontal cortex under the gallery context compared to the computer one,

suggesting a neural system supporting contextual modulation of aesthetic ratings. In the

present study, we hypothesized that participants would evaluate the same recording more

positively when presented with a text suggesting high prestige of the performer than when

presented with texts of lower prestige levels.

Another important extrinsic factor responsible for differences in musical judgements when

the acoustic input is identical may be the effect of repeated exposure. In line with the domain-

general mere exposure effect (Zajonc, 1968), liking to an initially neutral stimuli increases

with repeated exposure. While the effect of mere exposure has been extensively studied using

particular pieces of music as stimuli (see North & Hargreaves, 2008 for a review), only a few

studies have examined this effect on evaluations of performances of individual pieces. In

a recent study, Kroger and Margulis (2016) presented participants with pairs of solo piano

performances and informed them that one was played by a conservatory student and the other

a world-renowned professional. After listening to each pair, participants had to select

which

they considered to have been performed by the professional. The results indicated that

participants selected the second performance as professional more frequently than the first

107

C.2 Method

performance, although this effect was modulated by the actual identity of the performer. In

relation to the repeated recording illusion, we hypothesized that participants’ ratings of the

same recording would improve with repeated exposure.

The present research had three main aims. The first was to construct an experimental

paradigm to enable the systematic measurement of the repeated recording illusion. The

second aim was to investigate possible individual difference factors that contribute to the

illusion (i.e., musical training, suggestibility, music preferences and personality). The third

aim was to investigate extrinsic factors responsible for differences in musical judgements

when the acoustic input remains the same (i.e., explicit information and repeated exposure).

In addition, in order to capture higher-order interactions between the extrinsic and individual

difference factors, an exploratory analysis of the same data aimed to identify conditions that

lead to particularly positive or negative judgements.

In constructing the experimental paradigm of the repeated recording illusion, participants

were misled to think that they had heard three different performances of an original music

piece. However, we played the exact same recording three times in succession. Each time

the recording was accompanied by a text suggesting low, medium or high prestige of the

performer. We repeated this experimental procedure with two different pieces of music, a

piece of classical music and a piece of popular music for which we assumed a high stylistic

familiarity for most participants. In order to study the repeated recording illusion without an

effect of explicit information, we examined a non-prestige group where we did not manipulate

prestige of the performer.

C.2 Method

C.2.1 Participants

A sample of seventy-two university students took part in the experiment (36 male, 36 female),

aged 19-39 (M = 24.26, SD = 3.60). Twenty-nine participants were considered as trained

musicians (M = 45.74, SD = 5.73 on the Musical Training subscale of the Goldsmiths Musical

Sophistication Index, Müllensiefen, Gingras, Musil, & Stewart, 2014; and had 6 to 8 years

of formal musical training). Forty-five participants were considered as non-musicians (M =

22.71, SD = 7.34 on the Gold-MSI; and had 1 year of formal musical training on average).

Twelve participants were randomly allocated to a non-prestige condition (6 male, 6 female),

aged 21-29 (M = 24.34, SD = 3.45). Participation was on a volunteer basis and unpaid.

108

C.2 Method

C.2.2 Design

The study employed a 3x3x2 repeated measures design. Explicit information (low vs.

medium vs. high prestige text), repeated exposure (first vs. second vs. third position), and

genre of the original music piece (popular vs. classical music) were the within-participant

factors. The three levels of the explicit information factor were fully counter-balanced with

presentation order across participants. Half of the participants started with the popular music

piece condition and the other half started with the piece of classical music. The dependent

variables consisted of a diverse range of musical judgements provided immediately after

each listening and at the end of each music condition. In order to explore the repeated

recording illusion without an effect of explicit information, we examined a non-prestige

group where we did not manipulate prestige of the performer. In addition, we measured

individual difference factors that were expected to contribute to the illusion (i.e., musical

training, suggestibility, music preferences and personality).

C.2.3 Materials

In the popular music condition participants listened to a live recording of ‘Jailhouse Rock’

by Elvis Presley recorded in NBC studios in 1968. The length of the recording was 1 minute

and 36 seconds. This piece was selected because we assumed a high stylistic familiarity

for most participants. In the classical music condition participants listened to the final part

of a live recording of ‘Bruckner Symphony No. 4 Die Romantische’ conducted by Günter

Wand and performed by the Berliner Philarmonic Orchestra in 1998. The length of the

recording was 2 minutes and 48 seconds. This piece was selected in order to replicate

empirically the experiment carried out in the German radio station WDR 3(Behne, 1987).

The original recordings were edited and normalised using ableton live computer software.

In the

popular live recording we edited the start and end points of the original recording in order

contain only the musical performance element of the recording. Similar to the German radio

experiment (Behne, 1987), the start and end points of the classical music piece were edited

to contain the final part of the performance. We then normalised the volume of the two

recordings to be fixed on the same threshold. Then each recording was duplicated three times

and written to the same compact disc, using iTunes 12.2.2. Each copy of the music recording

was saved under a different name, which included performers’ names as used in the texts

suggesting different levels of prestige. In the non-prestige condition, the names were

‘performance 1’, ‘performance 2’, and ‘performance 3’.

To manipulate the effect of explicit information we created three texts suggesting low,

medium and high prestige of the performer. The texts had the same format, organisation and

109

C.2 Method

a length of 150 words. In the popular music condition (‘Jailhouse Rock’), the three ‘different’

performers were presented as different Elvis impersonators. The prestige texts provided

information about the three impersonators, who differed on skill and success (see Appendix

A in the published paper online). In the classical music condition (‘Bruckner’s Symphony

No.4’), the three ‘different’ performers were presented as different classical conductors. The

prestige texts provided information about the conductors, who differed on skill and success

(see Appendix B in the published paper online). Günter Wand, the actual conductor of the

recording, was not among these conductors. In the non-prestige condition, three different

texts were created with the same format, organisation and length of 150 words. While in the

popular music condition the three texts provided neutral information from different parts

of Elvis Presley’s biography, in the classical music condition the texts provided neutral

information from different parts of Anton Bruckner’s biography.

In order to evaluate liking as well as more objective aspects of the performance (e.g.,

pitch accuracy and tempo appropriateness), we designed an evaluation form consisting of

ten Likert rating scales and two open-text boxes. Nine of the rating scales consisted in sliders

ranging from 0 to 100. The rating scales were provided to evaluate the following

dimensions:

(1) liking of the interpretation, (2) timing and rhythm, and (3) tone quality (from

‘dislike

strongly’ to ‘like strongly’), (4) tempo appropriateness (from ‘very inappropriate’ to ‘very

appropriate’), (5) pitch accuracy (from ‘very inaccurate’ to ‘very accurate’), (6)

emotional

quality and (7) overall quality of the performance (form ‘very bad’ to ‘very good’),

and degree

of agreement to two statements: (8) some aspects regarding the singer’s vocal technique/

orchestral technique could be improved, and (9) some aspects of the overall interpretation

could improve (from ‘strongly disagree’ to ‘strongly agree’). In addition, (10) participants

were asked to rate each recording using a 5-star rating scale, ranging from 1 star (strongly

dislike) to 5 stars (like strongly). The Likert rating scales were designed to examine

differences in musical judgements when the acoustic input is the same. After the ten Likert

rating scales, two open-text boxes were provided where participants could write down

anything to describe the performance and whether or not they enjoyed it. Answering the

open-text boxes was optional.

At the end of each music condition, participants were requested to fill out a final evaluation

form. In this final evaluation, participants were asked to rate how much they liked each

recording compared to the others, on a scale from 0 (much less than the others) to 100 (much

more than the others), where the midpoint of the scale (‘50’) was labelled as ‘as much as the

others’. Participants also had to evaluate the familiarity to the original piece of music, on a

scale from 0 (‘don’t know at all’) to 100 (‘know very well’). In all rating scales, participants

110

C.2 Method

were able to see the number attributed to their specific rating. We also provided an open-text box

where participants could write down any optional comments regarding the experience of

the

experiment. The information from the open-text boxes was used to determine whether

participants fell for the illusion or not. When the information from the open-text boxes was

not sufficient to make a clear and objective decision, the final comparative rating scales were

taken into consideration to determine whether participants fell for the illusion or not. The

open-text boxes were used in conjunction with the final comparative rating scales, designed

to address a clear limitation in this experiment: we could not ask participants explicitly

whether the recordings were the same or different as this would have biased their subsequent

evaluations and behaviour in the experiment.

In order to measure the individual difference factors, participants filled out different ques-

tionnaires corresponding to each factor. To measure participants’ musical training and active

engagement with music we used the Goldsmiths Musical Sophistication self-report ques-

tionnaire (Gold-MSI, Müllensiefen et al., 2014). To measure participants’ suggestibility, we

used

the Social Desirability Scale (SDS-17) (Stöber, 2001) and 8 items adopted from the

Susceptibility Persuasive Strategies Scale (STPS) (Kaptein, Ruyter, Markopoulos, & Aarts,

2012), which measured bias to authority, consensus and persuadability, used in a previous

study (Unal, Temizel, & Eren, 2014). To assess music preferences and stylistic familiarity,

we used the Short Test of Music Preferences revised (STOMP-R, Rentfrow & Gosling, 2003).

To measure personality, we used the Big Five Inventory (BFI) (John & Srivastava, 1999).

C.2.4 Procedure

Participants were tested individually in small cubicle rooms. They listened to the music

recordings using professional headphones (KNS 8400 Studio Headphones, KRK systems)

and at a comfortable listening level that could be adjusted by the individual participants prior

to the actual experiment. Participants were told that the main purpose of the study was to

measure people’s skills in evaluating technical and musical aspects of different musical

performances of the same original piece. After filling out the Gold-MSI questionnaire,

participants were instructed to listen to three different interpretations of the same piece of

music and to evaluate them as accurately as possible. Before listening to each recording,

participants were presented with the corresponding text suggesting different levels of prestige.

Immediately after reading the text participants listened to the recording. Immediately after

listening to each recording, participants completed the evaluation form, where they were

presented with the ten Likert rating scales and two open-text boxes. The experiment had two

parts with exactly the same procedure and experimental instructions, but using popular

111

C.3 Results

music (‘Jailhouse Rock’) and classical music (‘Bruckner’s Symphony No.4’) respectively.

Immediately after listening the three recordings of each part, participants filled the final

evaluation form consisting in the final comparative rating scales and the open-text box.

Between completing the two parts of the experiment participants were asked to fill out

the STOMP-R questionnaire. In the non-prestige condition the procedure was the same.

Participants were also instructed that they would listen to three different performances of the

same piece, but the three recordings were presented as ‘performer 1’, ‘performer 2’, and

‘performer 3’, and the texts presented with the music did not induce any kind of prestige. Two

weeks after the experiment, participants were asked via email to fill out the BFI, SDS-17, and

the

8 items measuring suggestibility. The experiment and questionnaires were implemented in

Qualtrics software (Qualtrics, Provo, UT). This research was granted ethical approval by

the

Ethics Committee of the Department of Psychology of Goldsmiths College, University of

London.

C.3 Results

C.3.1 The Repeated Recording Illusion

In order to determine whether participants fell for the repeated recording illusion or not we

used the following procedure: We first assessed the information provided in the open-text

boxes. From a total of 14 open-text boxes (7 in the popular music condition and 7 in the

classical music conditions), on average participants provided information in 12.65% of the

boxes (6.33% in the popular music condition and 6.32% in the classical music condition). By

using the information provided in the open-text boxes we were able to identify 48 participants

out of 72 (66.67%) in the popular music condition and 50 participants out of 72 (69.45%) in

the classical music condition, who provided specific information either reporting differences

between performances or reporting that the recordings were the same.

There were cases wherein the information from the open-text boxes was not sufficient to

make a clear and objective decision but suggested a direction: either that the participant was

not aware that the recordings were identical or that the participant suspected that they were

the same. In these cases, we took into consideration the scores from the final comparative

rating scales where participants had to compare how much did they like each recording in

comparison to the others, on a scale from 0 (much less than the others) to 100 (much more

than the others), where the midpoint of the scale (‘50’) was labelled as ‘as much as the

others’. We only classified the participant when the scores from the final comparative ratings

112

C.3 Results

confirmed the suggested direction from the text boxes. It is important to note that we never

took into consideration the scores form the final comparative ratings on its own.

When the information from the open-text boxes was not sufficient and/or too ambiguous

to make a clear and objective decision, we did not include the participant’s data in the

subsequent analyses. Two participants provided highly ambiguous statements in the open-

text boxes for both music conditions and the two participants were therefore excluded from

the subsequent analyses. Furthermore, one participant provided ambiguous information in

the popular music condition and a different participant in the classical music condition. Thus,

had a total of 69 participants in each music condition.

As a consequence of using the above mentioned procedure, we had a total of four possible

criteria to determine whether participants fell for the repeated recording illusion or not (see

Appendix C, in the paper published online, for a decision diagram depicting the decision

procedure and criteria; Tables in Appendix F and G (in the paper published online) from the

supplementary materials show the information used to make each individual decision per

participant in the two music conditions):

When the information provided in the open-text boxes specifically indicated any

differences between performances: In the popular music condition, 37 out of 69 par-

ticipants (53.62%) specifically reported information indicating differences between

performances, such as “more upbeat than the two others, a happier sounding perfor-

mance” or “this piece sounds more aggressive than the previous one. The tempo for

me is faster”. In the classical music condition, 42 out of 69 participants (60.87%)

specifically reported information indicating differences between performances, such as

“the mood in this piece seemed to escalate a lot more naturally than in the other pieces”

or “this interpretation sounded a bit more hesitant. Again, it was not as dramatic as the

first performance, but it was clearer than the second one”.

When the information in the open-text boxes specifically indicated that the participant

realized that the recordings were the same: In the popular music condition, 11 out 69

participants (15.94%) specifically reported information indicating that the recordings

were the same (e.g., “I reckon this is the same file repeated three time” or “this is

absolutely the same as the first two”). In the classical music condition, 8 out 69

participants (11.59%) specifically reported information indicating that the recordings

were the same (e.g., “This sounds exactly like the two others” or “I thought all 3 were

the same”).

113

C.3 Results

When the information provided in the open-text boxes was not sufficient to make a

clear and objective decision but suggested that the participant was not aware that the

recordings were identical: In these cases, in addition to the open-text boxes, we took

into consideration the scores from the final comparative rating scales. If at least one

score from the final comparative ratings differed by 10% from the midpoint of the

scale (‘50’), or any two scores differed by 10% from each other, we considered the

participant as falling for the illusion. 19 participants (27.54%) in the popular music

condition and 17 participants (24.64%) in the classical music condition were classified

using this third criterion.

When the information provided in the open-text boxes was not sufficient to make

a clear and objective decision, but suggested that the participant suspected that the

performances were the same: In these cases, in addition to the open-text boxes, we

took into consideration the scores from the final comparative rating scales. If the three

scores from the final comparative ratings did not differ more than 10% from the

midpoint of the scale (‘50’), we considered the participant as not falling for the illusion.

Two participants (2.90%) in the popular music condition and two different participants

(2.90%) in the classical music condition were classified using this fourth criterion.

Table C.1 shows the number of participants who fell for the repeated recording illusion. In

the total sample of participants, 52 out of 69 participants (75.36%) believed that they had

heard different musical performances in at least one of the two music conditions. By contrast,

participants (24.64%) recognised that the performance was the same in at least one of the two

music conditions. Only 6 out of 69 participants (8.7%) realized that the recordings were

identical in both music conditions. When looking at the music conditions separately, in the

popular music condition 56 participants (81.16%) fell for the illusion and 13 participants

(18.84%) did not. In the classical music condition, 59 participants (85.51%) fell for the

illusion and 10 participants (14.49%) did not. Additionally, in the non-prestige condition

(where the effect of explicit information was not manipulated), 9 out of 12 participants

(75%) were susceptible to the illusion. According to a X2 test, there was no significant

association between the music conditions (popular and classical piece) and the occurrence of

the repeated recording illusion, X2 (1) = .47, p = .49. According to Fisher’s Exact test, there

was no significant association between the presence of prestige (i.e., prestige-suggestion and

non-

prestige group) and the occurrence of the illusion (p = .65).

114

C.3 Results

Table C.1 Numbers of participants falling for the repeated recording illusion.

Participants were classified as NO if they identified the three recordings as identical in at least one of the two

music conditions.

Generally, participants rated the popular music piece as more familiar (M = 72.16, SD =

21.93 on 100-point rating scale) than the classical piece (M = 13.73, SD = 21.10). This

difference in familiarity was highly significant as indicated by a paired samples t-test, t (68)

= 16.43, p < .001.

C.3.2 Individual Difference Factors

The analysis of individual difference factors was conducted using a data classification

method known as the random forest (Breiman, 2001), in which the aim was to examine

whether individual differences contributed to the repeated recording illusion. Random forest

procedures differ in a number of ways from other classification methods in that they

can handle large sets of predictor variables and do not assume a linear relationship between

predictors (see Hastie, Tibshirani, Friedman, & Franklin, 2009; see Pawley & Müllensiefen,

2012 for the use of random forests in music psychology). We used the conditional random

forest based on permutation tests as implemented in the R package “party” (Hothorn,

Buehlmann, Dudoit, Molinaro, Van der Laan, 2006; Hothorn, Hornik, & Zeileis, 2006;

Strobl, Boulesteix, Kneib, Agustin, Zeileis, 2008; Strobl, Malley, & Tutz,

2009). The random

forest model was run with a size of 5000 trees. We employed a measure

of variable importance

for each predictor variable, which is designed to produce unbiased estimates of variable

importance even in situations where significant correlations between predictor variables

exist and when the dependent variable is very unequally distributed (atza, Strobl, &

Boulesteix, 2013).

As predictor variables, we used 6 demographic variables as well as musical variables that

were collected during the experimental session (age, gender, Gold-MSI Musical Training and

Active Engagement scores, STOMP preference scores for Reflective & Complex, Intense

115

C.3 Results

& Rebellious, Upbeat & Conventional, and Energetic & Rhythmic). Data for 9 additional

variables were collected via the follow-up questionnaire measuring the big five personality

traits (Extraversion, Agreeableness, Conscientiousness, Neuroticism, and Openness) as well

as suggestibility (Authority score, Consensus score, Persuadability score, and Social

Desirability score). Using these 17 predictor variables we computed two different models

with two different binary dependent variables: (a) a strict criterion model in which only those

participants who fell for the illusion in both music conditions were considered as not

falling

for the illusion and, (b) a less strict criterion model where we considered as not falling

for the

illusion those participants who fell for the illusion in at least one of the two music conditions.

A variable importance score was obtained for each predictor variable, describing

how

predictive each variable was compared to the others. We applied a “confidence interval”

criterion

in order to select the top performing variables. Only the variables whose variable importance

scores were positive and greater than the absolute value of the lowest negative variable

importance score were selected (Strobl et al., 2008; Strobl et al., 2009).

The two models (strict and less strict criterion) delivered very similar results, indicating that

there were two variable importance scores that met the above criterion (neuroticism and

openness). In both models, neuroticism was the most important variable contributing to the

repeated recording illusion, followed by openness (see Appendix D, in the paper published

online, for graphs with the 17 variable important scores in the two models). In the strict

criterion model, neuroticism was approximately 3.5 times more important than openness. In

this model, those participants falling for the illusion in the two music conditions had a higher

mean neuroticism score of 23.41 (SD = 5.17) and a higher mean openness score of 40.12

(SD = 5.14) compared to those participants who did not fall for the illusion (M = 17.43, SD

= 6.85 on the neuroticism factor; M = 35.28, SD = 7.02 on the openness factor). In the less

strict criterion model, neuroticism was approximately 3 times more important than openness.

In this model, those participants who fell for the illusion in at least one of the two music

conditions had a higher mean neuroticism score of 23.14 (SD = 5.55) and a higher mean

openness score of 40.12 (SD = 5.42) compared to those participants who did not fall for

the illusion (M = 17.43, SD = 6.85 on the neuroticism factor; M = 35.28, SD = 7.02 on the

openness factor).

C.3.3 Extrinsic Factors: The Effects of Explicit Information and Re-

peated Exposure

The subsequent analyses included the sixty participants of the main experimental group (i.e.,

where we manipulated the effect of explicit information). In the popular music condition,

116

C.3 Results

three participants were excluded from the analyses and ten fell for the illusion. Hence, in the

popular music condition we had an overall of 47 participants. In the classical music

condition, three participants were excluded from the analyses and nine fell for the illusion.

Hence, in the classical music condition we had an overall of 48 participants.

Participants’ ratings on the ten Likert rating scales were aggregated into a single scale. First,

ratings of each participant on each rating scale were transformed into z-scores across ratings

of all six recordings (three in the popular music condition and three in the classical). Then,

a principal component analysis (PCA) was conducted on the z-transformed data of the ten

rating scales. The Kaiser-Meyer-Olkin (KMO) measure verified the sampling adequacy for

the analysis, KMO = .93 (‘marvellous’ according to Hutcheson & Sofroniou, 1999). In

addition, all KMO values for individual rating scales were greater than .86, which is well

above commonly accepted limit of .5 (Field, 2013). The scree plot of the different factor

solution was very clear and indicated a solution with just one factor. Moreover, there was

only one PCA component with an eigenvalue >1 which explained 64.56% of the variance.

Thus, this 1-factor PCA solution was accepted and component scores for all participant

ratings were computed using the regression method.

Because the two music recordings used in the popular and classical music conditions differed

substantially in several aspects (i.e., musical genre, familiarity, presence of words/ vocaliza-

tions, duration of the excerpt and quality of the recording), we ran two separate models, one

with the ratings obtained in the popular music condition and one with the ratings obtained in

the classical music condition (see Appendix E, in the paper published online, for a summary

table of both models). Participants’ ratings were standardised separately for each music

condition.

To test the hypothesis regarding the effects of explicit information and repeated exposure

we used the R packages lme4 (Bates, Mächler, Bolker, & Walker, 2015) and lmerTest

(Kuznetsova, Brockhoff, & Christensen, 2016) to perform a linear mixed effects analysis

with the z-scores of the participants’ ratings as the dependent variable. In the two models,

explicit information (low, medium, and high prestige of the text) and repeated exposure (first,

second, and third position) were the fixed effect independent factors, whereas participants

were the random effect factor.

The linear mixed-effect model of the popular music condition revealed that there were signifi-

cant main effects of explicit information (p < .001) and repeated exposure (p < .001). Because

the interaction between explicit information and repeated exposure was not significant we

ran the model again only with the two main factors. The effects of explicit information and

117

C.3 Results

repeated exposure become visible in Figure C.1. The effect of explicit information shows

that when the recording was presented with a high prestige text the ratings were significantly

higher than when presented with low and medium texts. The effect of repeated exposure of

the recording shows that when the recording was heard in the second and third positions the

ratings were significantly higher than when heard in the first position.

The linear mixed-effect model of the classical music condition revealed that there was a

significant main effect of explicit information (p < .001). However, the effect of repeated

exposure and the interaction between explicit information and repeated exposure were not

significant. Because the interaction between explicit information and repeated exposure was

not significant we ran the model again only with the two main factors. The effect of explicit

information shows that when the recording was presented with a high prestige text the ratings

were significantly higher than when presented with low and medium texts (Figure C.2).

Figure C.1 Effects of explicit information and repeated exposure in the popular music

condition (error bars represent the standard error).

118

C.3 Results

Figure C.2 Effects of explicit information and repeated exposure in the classical music

condition (error bars represent the standard error).

The R2 for the classical music model was 0.16 and therefore lower than the R2 of 0.277 of

the popular music model, indicating that the extrinsic factors explained more of the variance

in the more familiar popular music condition.

C.3.4 Exploratory Analysis (Regression Model Tree)

In order to capture higher order interactions between extrinsic and individual difference

factors and identify conditions that lead to particularly low and high ratings, we computed a

regression tree model based on permutation tests as implemented in the R package “party”

(Hothorn et al., 2006; Hothorn et al., 2006; Strobl et al., 2008; Strobl, et al., 2009). Statistical

tree models differ in a number of ways from linear regression models (see Hastie et al., 2009)

in that they use a built-in variable selection mechanism and therefore can handle large sets of

predictor variables. In addition, tree models do not assume a linear relationship between

predictors and the dependent variable and they are very useful for modelling higher-order

interaction effects between predictor variables automatically. For this study we used a

particular family of tree models called conditional inference trees that combine the

119

C.3 Results

rigorous theory of permutation statistics (Hothorn et al., 2006) with the principle of recursive

partitioning (Zeileis, Hothorn, & Hornik, 2008).

For the regression tree model, the z-transformed participants’ ratings served as dependent

variable. In addition to the two extrinsic factors (explicit information and repeated exposure), we

added the factor musical genre (popular and classical music) and six individual difference

variables (1. musical training, 2. self-rated familiarity with the music piece, 3. preference for

the STOMP meta-genre reflective & complex, 4. preference for the STOMP meta- genre

Intense & Rebellious, 5. neuroticism, and 6. Openness), resulting in a total of nine

independent variables. Figure C.3 shows the structure of the regression tree. The model

makes use of only 3 of the nine independent variables and has an R2 value of 0.234. For each

node of the tree, the p-values indicating the significance of the split based on the permutation

statistics are presented as well as a description of the two subgroups of the split on the

independent variable. For the terminal nodes at the bottom of the graph, the distribution of

ratings on the standardised rating scale are depicted as box- and whiskers plots.

The tree model can be interpreted by starting at the top and following each branch down, to

arrive at a terminal node. A path to a terminal node describes the interaction of experimental

conditions that lead to a particular subset of ratings. To arrive at the subset with the highest

(i.e. most positive) average ratings, follow the first “Explicit Information” node down

the “High Prestige” branch (left-hand side) and then descend to the left at the “Repeated

Exposure” node down the “2nd and 3rd Position” branch. This branch can be interpreted as

follows: when participants listened to the music recording presented with a high prestige text

in the second and third positions, the average ratings were around 1 and, therefore, the highest

compared to the other terminal branches of the model. In contrast, the lowest ratings, which

were around -1, were given when the recording was presented with low and medium prestige

texts, in the popular music condition, and when the recording was heard for the first time.

Overall, the regression tree model confirms the effects of explicit information and repeated

exposure, but it also shows higher-level interactions between the extrinsic factors and the

two pieces of music. None of the individual difference factors were significant in the tree

model. This indicates that after participants had fallen for the illusion, individual difference

factors did not play an important role and musical judgements were mainly influenced by the

extrinsic factors.

120

C.4 Discussion

Figure C.3 Regression tree model.

C.4 Discussion

The primary aim of the present study was to construct an experimental paradigm to enable

the systematic measurement of the repeated recording illusion. Participants were misled to

think that they had heard three different performances of an original piece when in fact they

were exposed to the same repeated recording. Each time, the recording was accompanied by

a different text suggesting a low, medium or high prestige of the performer. Most participants

(75.36%) were under the impression of hearing different musical performances when in fact

they were identical. In contrast, seventeen participants (24.64%) recognised that the

performance was the same in at least one of the two music conditions. Only six participants

(8.7%) realized that the recordings were identical in both music conditions. Nearly three-

quarters of the participants provided verbal comments indicating specific differences between

the

performances (e.g., “this piece sounds more aggressive than the previous one. The tempo

for

me is faster”) or that they were the same (e.g., “I reckon this is the same file repeated three

121

C.4 Discussion

times”). Thus, it can be concluded that the majority of the participants fell for the repeated

recording illusion. This finding suggests that musical judgements are sometimes not based

on perceptual features and musical cues but are influenced by factors that do not depend on

the music itself. This is at least true when a mild deception is applied and participants believe

that they had heard different performances.

It could be argued that the repeated recording illusion occurs in part because participants are

not familiar with the original piece of music. Therefore, we examined the illusion using two

different pieces that were significantly different on familiarity, a highly familiar piece of

popular music (‘Jailhouse Rock’ by Elvis Presley) and a highly unfamiliar piece of classical

music (‘Bruckner’s Symphony No. 4’).

The repeated recording illusion occurred similarly in

the

two music conditions. However, these two recordings differed substantially in several other

aspects, including musical genre, complexity, length of the excerpt, presence of vocals and

quality of the recording. Thus, these variables are confounded in this experimental setup. Any

interpretation of differences between the two musical stimuli will have to take this into

account. Further studies should explore the repeated recording illusion with a larger range of

different performances and recordings.

It is important to note that there is a main methodological restriction to be considered in

the experimental design used here: an implicit bias of authority figure. In other words, the

fact that participants were told they would listen to ‘three different performances’ by an

investigator in a lab situation may account, at least partly, for the occurrence of the illusion.

It would be interesting for future research to investigate the repeated recording illusion using

an experimental paradigm without any implicit bias of authority. This paradigm could consist

in presenting participants with pairs of different and identical musical performances.

Participants would be instructed to rate how different are the two performances using several

rating scales. In the cases where the performances were identical, participants’ ratings would

indicate to what extent people hear differences when listening to the same repeated recording

without relying on a judgements bias excreted by a figure of authority.

The second aim of the study was to investigate possible individual difference factors that

contribute to the repeated recording illusion. The most important individual difference

factor related to the illusion was the personality trait of neuroticism, which is in line with

previous research showing a positive (but low) link between vulnerability to suggestion and

neuroticism (see Gudjonsson, 2003). This finding suggests that people who tend to be

anxious, pessimistic, shy, fearful, vulnerable and emotionally unstable

are more likely to fall for

the repeated recording illusion. Although less important, openness to experience also was

significant factor related to the occurrence of the illusion, suggesting that people who tend

122

C.4 Discussion

to be curious, imaginative, artistic, excitable and unconventional are more likely to fall for

the illusion. Importantly, none of the other individual difference factors that were expected

to contribute to the illusion were significant, including musical training, suggestibility and

preferences for music style. We consider particularly interesting that different levels of

suggestibility (including bias to authority, consensus, persuadabiliy and social desirability)

were not related with the occurrence of the illusion. Moreover, in our sample of participants,

highly trained musicians were not any more or any less susceptible to the repeated recording

illusion than participants with low levels of musical training. Thus, it remains still open the

question of which are the main individual differences contributing to the repeated recording

illusion. For instance, what would occur when using participants with a greater range of

musical training and expertise (e.g., top-level professional musicians and music critics)?

Would other individual differences (e.g., intelligence, memory, perceptual abilities) be able

to explain why some people fall for the illusion while others seem no be unaffected by it?

The third aim of the present research was to investigate extrinsic factors responsible for

differences in musical judgements when the acoustic input remains the same. As predicted,

we found that the effect of explicit information contributed significantly to differences in

musical judgements. This effect was clear in the two music conditions, where participants

rated the same music recording significantly better when presented with a high prestige text

than when presented with low and medium prestige texts. This finding is consistent with

previous research on the effects of explicit information upon aesthetic reactions to music

(e.g., Kroger & Margulis, 2016; Margulis, 2010; Margulis, Kisida, & Greene, 2015; North

& Hargreaves, 2005). Using a similar paradigm, where identical artworks were presented

with different contextual explicit information varying in prestige, Kirk et al. (2009) found

that prefrontal and orbitofrontal cortices recruited by aesthetic judgements were significantly

influenced by the explicit information presented with the same stimuli. We suggest that this

neural system could also be responsible for the modulation of aesthetic reactions to music by

explicit contextual information.

The effect of repeated exposure was only significant in the more familiar popular music

condition, but not in the more unfamiliar classical music condition. This finding supports

partly previous research on the effects of repeated exposure to music (North & Hargreaves,

2008 for a review). In one of the few studies using musical performances as stimuli, Kroger

and Margulis (2016) found that evaluations of performances were driven by a combination

of repeated exposure and the actual identity of the performer. Interestingly, in a second

experiment, Kroger and Margulis (2016) found that the effect of explicit information was

mitigated by the influence of the actual performer and repeated exposure, showing interplay

123

C.4 Discussion

between intrinsic and extrinsic factors. In the present study, the two original pieces of music

differed in a number of important aspects. For instance, the classical piece was a minute

longer than the popular piece, did not contain vocals and was highly unfamiliar to most

of the participants. Furthermore, while the popular music piece was a live recording from

1968 that had a notably worse recording quality than ordinary studio recordings, the quality

of the classical music piece (recorded live in 1998) was superior. Therefore, it may be

possible that the effect of repeated exposure did not affect participants in the classical music

condition because of the nature of the music recording. Moreover, the explicit information

presented with the recordings might have had a different impact on participants in the two

music conditions. Future studies will need to explore the strength of the effect of repeated

exposure across a larger range of different performances and recordings.

In an attempt to explore higher-order interactions between the extrinsic and individual

difference factors, we used a regression tree model in which we identified conditions that lead

to particularly low and high ratings. The highest ratings were given when the music recording

was presented with a high prestige text and heard in the second and third positions. In

contrast, the lowest ratings were found when participants listened to the popular music piece

in the first position and presented with low and medium prestige texts. Overall, the regression

tree model confirmed the effects of explicit information and repeated exposure, but it also

showed higher-level interactions between the extrinsic factors and the two pieces of music.

None of the individual difference factors used in the model (musical training, familiarity

with the original piece, music preferences, neuroticism and openness) were significant in the

regression tree model. This finding suggests that after participants had fallen for the illusion,

individual difference factors did not play an important role and musical judgements were

mainly influenced by the extrinsic factors.

The present study focussed on extrinsic factors in order to examine differences in musical

judgements when the acoustic input remains the same. Nevertheless, one could argue that

the factors of explicit information and repeated exposure might also be responsible, in part,

for the occurrence of the illusion. The results from a non-prestige group, where the effect of

explicit information was not manipulated, indicated that 75 % participants were susceptible

to the illusion. This finding suggests that the effect of explicit information is not essential for

the occurrence of the illusion. By contrast, we consider it likely that the effect of repeated

exposure contributes to the illusion. In an extensive investigation of repetition in musical

experience, Margulis (2014) provides relevant insights to this matter. She stated that, “[a]t

a minimum, a repeated element will sound different from its initial presentation by virtue of

coming later and having been heard before” (Margulis, 2014, p. 35). Although in this

124

C.4 Discussion

quote Margulis refers to repetition within individual pieces of music, we find it plausible

that the same principle should apply to the repeated recording illusion: while the musical

input remains the same, repeated exposure modifies the listening experience, giving rise to

the feeling that the performances are different.

Two relevant questions arise from the results of this study. Why are some individuals more

susceptible to the illusion than others? One way to approach this question is the study of

further individual difference factors (e.g.

intelligence, memory, perceptual abilities) that may

associated with the repeated recording illusion. The second question refers to a more

fundamental issue: did participants in this study actually perceive differences between the

repetitions of the same recording? Or, alternatively, did they believe they heard differences

because they were misled to think so? We encourage the use of neuroimaging techniques as

one possible approach to investigate whether the illusion is a perceptual phenomenon or

rather a bias in a secondary and later stage of cognitive processing and decision-making.

Taking a wider perspective, the research framework developed by Tversky and Kahneman

(Kahneman & Tversky, 1984; Tversky & Kahneman, 1974; see Kahneman, 2011 for a

review) could provide a theoretical framework by which the results of the current study

could be interpreted. Although it does not involve music and is mainly concerned with

economic decision processes, Tversky and Kahneman’s framework offers insight into how to

investigate traditional psychological biases in musical judgements by using recent research

on human judgements and decision-making. However, this framework has not yet been

applied explicitly to the study of evaluative judgement processes involving music.

The effect of explicit information may fall within a broad heuristic principle, namely, the

affect heuristic (Kahneman & Frederick, 2002; Slovic, Finucane, Peters, & MacGregor,

2002), which refers to the reliance on good or bad feelings experienced in relation to a

stimulus. Thus, if the emotions associated with a stimulus are positive, people will be more

likely to judge characteristics of the pertinent stimulus more positively, as found in the

present study when the music recording was presented with a high prestige text. Similarly,

the effect of repeated exposure is one of several mechanisms within the bias of perceptual

fluency (Kahneman, 2011), which has been widely shown to influence human judgements

and decision-making in many areas (see Reber, Schwarz, & Winkielman, 2004 for a review).

Such findings suggest that perceptual fluency gives rise to feelings of familiarity and a

positive affective response that results in an increase in preference judgements. In the present

study, this is evident only when participants listened to the more familiar popular music

recording.

125

C.5 References

Our results suggest that at least in certain situations, evaluations of music rely on judge-

ment biases and heuristics that do not depend on the stimuli themselves, which is in line with

models of decision-making and the research framework developed by Tversky and

Kahneman. However, when applying Tversky and Kahneman’s framework to the study of

evaluative and judgment processes involving music, one should consider the implications

and difficulties of using music as stimuli (e.g., familiarity, complexity, presence of vocals,

individual preferences to music, personality). This approach wherein biases in musical

judgements are linked to comparable research in behavioural economics could be used to

investigate and better understand musical judgements, preferences and choice behaviour.

This general approach, that could be termed the behavioural economics of music, would

attempt to create a solid understating of the role that behavioural economics can play in

the study of musical judgements and preferences, two fields that have been surprisingly

unconnected in the literature so far.

In summary, the findings of the present study show that most participants were under the

impression of hearing different musical performances when in fact they were identical. This

illusion occurred regardless of participants’ levels of suggestibility, musical training, and

preferences for music style. However, high levels on the personality traits of neuroticism

and openness made it significantly more likely that an individual would fall for the illusion.

While the explicit information presented with the music influenced participants’ evaluations

of music significantly, the effect of repeated exposure affected participants’ ratings only

in the more familiar popular music recording. These findings support previous research

showing that musical judgements are sometimes not based on musical cues and features but

are influenced by factors that do not depend on the music itself. Beyond the findings and

limitations of the present research, the repeated recording illusion can constitute a useful

paradigm for investigating psychological biases and individual differences in aesthetic and

musical judgements because the illusion allows for the study of their effects while the music

remains the same.

C.5 References

Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models

using lme4. Journal of Statistical Software, 67(1), 1-48.

Breiman, L. (2001). Random forests. Machine Learning, 45, 5–32.

Behne, K. -E. (1987). Urteile und Vorurteile: Die Alltagsmusiktheorien jugendlicher Hörer.

In H. Motte-Haber (Ed.), Psychologische Grundlagen des Musiklernens (pp. 221-272).

Kassel, DE: Bärenreiter.

126

C.5 References

Behne, K. -E., & Wöllner, C. (2011). Seeing or hearing the pianists? A synopsis of an early

audiovisual perception experiment and a replication. Musicae Scientiae, 15(3), 324-

342.

Cassidy, J. W., & Sims, W. L. (1991). Effects of special education labels on peers’ and adults’

evaluations of a handicapped youth choir. Journal of Research in Music Education,

39(1), 23.

Deliege, I. (1987). Grouping conditions in listening to music: An approach to Lerdahl and

Jackendoff’s grouping preference rules. Music Perception, 4(4), 325-60.

Dowling, W. J. (1978). Scale and contour: Two componenets of a theory of memory for

melodies. Psychological Review, 85(4), 341-354.

Dowling, W. J., & Bartlett, J. C. (1981). The importance of interval information in long-term

memory for melodies. Psychomusicology, 1, 30-49.

Duerksen, G. L. (1972). Some effects of expectation on evaluation of recorded musical

performance. Journal of Research in Music Education, 20(2), 268-272.

Elliott, C. A. (1995). Race and gender as factors in judgments of musical performance.

Bulletin of the Council for Research in Music Education, 127, 50-56.

Field, A. (2013). Discovering statistics using IBM SPSS statistics. London, UK: Sage.

Greasley, A., & Lamont, A. (2016). Musical Preferences. In S. Hallam, I. Cross, & M. Thaut

(Eds.), Oxford handbook of music psychology (second edition) (pp. 263-281). Oxford,

UK: Oxford University Press.

Griffiths, N. K. (2008). The effects of concert dress and physical appearance on perceptions

of female solo performers. Musicae Scientiae, 12(2), 273-290.

Gudjonsson G.H. (2003). The psychology of interrogations and confessions: A handbook.

West Sussex, UK: John Wiley & Sons.

Halpern. A. R., Bartlett, J., & Dowling, W. (1995). Aging and experience in the recognition

of musical transpositions. Psychology and Aging, 10(3), 325–342.

Hastie, T., Tibshirani, R., & Friedman, J. (2009). Hierarchical Clustering. In T. Hastie, E.

Tibshiran, & J. Friedman (Eds.), The elements of statistical learning: Data Mining,

inference and prediction (2nd ed.) (pp. 520-528). New York, NY: Springer.

Hothorn, T., Buehlmann, P., Dudoit, S., Molinaro, A, & Van Der Laan, M. (2006). Survival

ensembles. Biostatistics, 7(3), 355-373.

Hothorn, T., Hornik, K., & Zeileis, A. (2006). Unbiased recursive partitioning: A conditional

inference framework. Journal of Computational and Graphical statistics, 15(3), 651-

674.

Janitza, S., Strobl, C., & Boulesteix, A. –L. (2013). An AUC-based permutation variable

importance measure for random forests. BMC Bioinformatics, 14(1), 119.

127

C.5 References

John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement,

and theoretical perspectives. Handbook of Personality: Theory and Research, 2(510),

102-138.

Juchniewicz, J. (2008). The influence of physical movement on the perception of musical

performance. Psychology of Music, 36, 417-427

Kahneman, D. (2011). Thinking, fast and slow. New York, NY: Farrar, Straus and Giroux.

Kahneman, D., & Frederick, S. (2002). Representativeness revisited: Attribute substitution

in intuitive judgment. In T. Gilovich, D. Friffin, D. Kahneman (Eds.), Heuristics and

biases: The psychology of intuitive thought (pp. 49-81). New York, USA: Cambridge

University Press.

Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American psychologist,

39(4), 341.

Kaptein, M., De Ruyter, B., Markopoulos, P., & Aarts, E. (2012). Adaptive persuasive

systems: A study of tailored persuasive text messages to reduce snacking. ACM

Transactions on Interactive Intelligent Systems, 2(2), 1-25.

Kirk, U., Skov, M., Hulme, O., Christensen, M. S., & Zeki, S. (2009). Modulation of aesthetic

value by semantic context: An fMRI study. NeuroImage, 44(3), 1125-1132.

Kroger, C., & Margulis, E. H. (2016). “But they told me it was professional”: Extrinsic

factors in the evaluation of musical performance. Psychology of Music, 45(1), 49-64.

Margulis, E. H. (2010). When program notes don’t help: Music descriptions and enjoyment.

Psychology of Music, 38, 285-302.

Margulis, E. H. (2014). On repeat: How music plays the mind. New York, NY: Ocford

University Press.

Margulis, E. H., Kisida, B., & Greene, J. P. (2015). A knowing ear: The effect of explicit

information on children’s experience of a musical performance. Psychology of Music,

43(4), 596-605.

Milgram, S. (1963). Behavioral study of obedience. Journal of Abnormal Psychology, 67(4),

371–378.

Müllensiefen, D., Gingras, B., Musil, J., & Stewart, L. (2014). The musicality of non-

musicians: An index for assessing musical sophistication in the general population.

PloS ONE, 9(2), e89642.

North, A. C., & Hargreaves, D. J. (2005). Brief report: Labelling effects on the perceived

deleterious consequences of pop music listening. Journal of Adolescence, 28(3), 433-

440.

North, A., & Hargreaves, D. (2008). The social and applied psychology of music. New York,

NY: Oxford University Press.

128

C.5 References

Orsmond, G. I., & Miller, L. K. (1999). Cognitive, musical and environmental correlates of

early music instruction. Psychology of Music, 27, 18-37.

Pawley, A., & Müllensiefen, D. (2012). The science of singing along: A quantitative field

study on sing-along behavior in the north of England. Music Perception, 30(2), 129–

146.

Pearce, M. T. (2015). Effects on processes involved in musical appreciation. In J. P. Huston,

M. Nadal, F. Mora, L. Agnati, F. Mora, & C. J. Cela-Conde (Eds.), Art, aesthetics and

the brain (pp. 319-338). Oxford, UK: Oxford University Press.

Radocy, R. E. (1976). Effects of authority figure biases on changing judgments of musical

events. Journal of Research in Music Education, 24(3), 119-128.

Reber, R., Schwarz, N., & Winkielman, P. (2004). Processing fluency and aesthetic pleasure:

Is beauty in the perceiver’s processing experience? Personality and Social Psychology

Review, 8(4), 364-382.

Rentfrow, P. J., & Gosling, S. D. (2003). The do re mi’s of everyday life: The structure and

personality correlates of music preferences. Journal of Personality and Social

Psychology, 84(6), 1236-1256.

Silveira, J. M., & Diaz, F. M. (2014). The effect of subtitles on listeners’ perceptions of

expressivity. Psychology of Music, 42(2), 233-250.

Silvey, B. A. (2009). The effects of band labels on evaluators’ judgments of musical

performance. Applications of Research in Music Education, 28(1), 47-52.

Slovic, P., Finucane, M., Peters, E., & MacGregor, D. G. (2002). Rational actors or rational

fools: Implications of the affect heuristic for behavioral economics. Journal of Socio-

Economics, 31(4), 329-342.

Stöber, J. (2001). The Social Desirability Scale-17 (SDS-17): Convergent validity, discrimi-

nant validity, and relationship with age. European Journal of Psychological Assessment,

17(3), 222-232.

Strobl, C., Boulesteix, A. -L., Kneib, T., Augustin, T. & Zeileis, A. (2008). Conditional

variable importance for random forests. BMC Bioinformatics, 9(23), 307.

Strobl, C., Malley, J., & Tutz, G. (2009). An introduction to recursive partitioning: Rationale,

application, and characteristics of classification and regression trees, bagging, and

random forests. Psychological Methods, 14(4), 323–348.

Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases.

Science (New York, N.Y.), 185(4157), 1124-31.

Unal, P., Temizel, T. T., & Eren, P. E. (2014, May). An exploratory study on the outcomes of

influence stra-tegies in mobile application recommendations. Paper presented at Pro-

129

C.5 References

ceedings of the Second International Workshop on Behavior Change Support Systems

(BCSS2014). Padova, IT.

Vuoskoski, J. K., & Eerola, T. (2013). Extramusical information contributes to emotions

induced by music. Psychology of Music, 43(2), 262–274.

Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social

Psychology, 9 (2p2), 1-27.

Zeileis, A., Hothorn, T., & Hornik, K. (2008). Model-based recursive partitioning. Journal

of Computational and Graphical Statistics, 17(2), 492-514.

Appendix D

False memories in music listening (S4)

This is an Accepted Manuscript of an article published by Taylor & Francis

in Memory on

4th November 2018, available online: https://doi.org/10.1080/09658211.2018.1545858.

The paper is not the copy of the record and may not exactly replicate the authoritative

document published in the journal. For presentation in this thesis, the appendices of the

paper have been removed and the passages referring to each Appendix in the text modified

to indicate where to find the materials online. Moreover, there may be minor modifications

in the text to guarantee a consistent typographic style throughout the thesis, such as the

position of figures and tables. Please do not copy or cite without author’s permission.

Citation

Anglada-Tort, M., Baker, T., & Müllensiefen, D. (2019). False memories in music listening:

exploring the misinformation effect and individual difference factors in auditory memory.

Memory, 27(5), 612-627. DOI: https://doi.org/10.1080/09658211.2018.1545858

Author contribution

I conceived the idea of this project and supervised it along with Prof. Dr. Daniel

Müllensiefen (Goldsmiths, University of London). The study was conducted and developed

by Thomas Baker as part of his master thesis in the MSc in Music, Mind, and Brain, at

Goldsmiths, University of London (2017-2018). After Thomas completed his masters, I

reanalysed the data and wrote the paper for publication.

The paper is deposited under the terms of the Creative Commons Attribution-NonCommercial 4.0 Internatinal

License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and

reproduction in any medium, provided the original work is properly cited

False memories in music listening: Exploring the

misinformation effect and individual difference factors in

auditory memory

The study of false memory has had a profound impact on our understanding of how and

what we remember, as shown by the misinformation paradigm (Loftus, 2005). Though

misinformation effects have been demonstrated extensively within visual tasks, they have not

yet

been explored in the realm of non-visual auditory stimuli. Thus, the present study aimed to

investigate whether post-event information can create false memories of music listening

episodes. In addition, we explored individual difference factors potentially associated with

false memory susceptibility in music, including age, suggestibility, personality, and musical

training. In two music recognition tasks, participants (N = 151) listened to an initial music

track, which unbeknownst to them was missing an instrument. They were then presented

with post-event information which either suggested the presence of the missing instrument

or did not. The presence of misinformation resulted in significantly poorer performance on

the music recognition tasks (d = .43), suggesting the existence of false musical memories. A

random forest analysis indicated that none of the individual difference factors assessed were

significantly associated with misinformation susceptibility. These findings support previous

research on the fallibility of human memory and demonstrate, to some extent, the generality

of the misinformation effect to a non-visual auditory domain.

Keywords: : false memory, misinformation effect, auditory memory, individual differences,

music listening.

132

D.1 Introduction

To have a false memory is to recall something which did not happen (Smelser & Baltes,

2001). Motivated in part by its potentially serious implications - such as within courtroom

testimonies - false memory has become one of the most widely explored topics in psychology

in recent decades (see Brainerd & Reyna, 2005; Loftus, 2005; Neuschatz, Lampinen, Toglia,

Payne, & Cisneros, 2017; Shaw, 2016; Scoboria et al., 2017, for reviews). From this wealth

of research, it has now become clear that false memories can significantly influence our

visual memory, from deteriorating the accuracy of eyewitness’ memory (Davis & Loftus,

2007; Gabbert, Memon, & Allan, 2003; Liebman et al., 2002; Weingardt, Toland, & Loftus,

1994), to creating autobiographical memories of entire events that never occurred (Bernstein

& Loftus, 2009a; Hyman, Husband, & Billings, 1995; Hyman & Kleinknecht, 1999; Wade,

Garry, Read, & Lindsay, 2002). Nevertheless, it remains unclear to what extent memories of

non-visual auditory stimuli, such as music, may prove exceptional or congruent within the

broader realm of false memory. To what extent are false musical memories consistent with

findings from the visual domain? The present study addresses this question by adapting for

the first time the misinformation paradigm to use music instead of visual materials.

Two of the most prominent paradigms to induce false memory in the visual domain are

the misinformation paradigm (see Frenda, Nichols, & Loftus, 2011; Loftus, 2005; Pickrell,

McDonald, Bernstein, & Loftus, 2016, for reviews) and the Deese-Roediger-McDermoott

(DRM) paradigm (Roediger & McDermott, 1995). The misinformation paradigm typically

involves three stages: experiencing an event, receiving post-event misinformation, and a

memory test. Findings across studies show that after exposure to post-event misinformation,

the accuracy of individuals’ memory is altered (e.g., Loftus, Miller, & Burns, 1978; Loftus

& Hoffman, 1989). In the original form of the DRM paradigm, participants are presented

with a list of semantically related words (e.g., bed, night, rest, awake, dream, blanket, snore,

nap). These words are all related by a word known as a “lure” (e.g., sleep), which is missing

in the list. When participants are then required to remember as many words from the list as

possible, they usually recall the lure as frequently as they recall the other presented words

(Roediger & McDermott, 1995).

To the best of our knowledge, there are only two studies in the published literature that

explored false memory in music, adapting a version of the DRM paradigm using music

stimuli (Curtis & Bharucha, 2009; Vuvan, Podolak, & Schmuckler, 2014). In both studies,

participants heard a melody followed by a test tone and were asked to indicate whether they

had heard the test tone in the previously presented melody. Findings in both studies showed

that participants were more likely to falsely remember context-congruent tones than

133

D.1 Introduction

context-incongruent tones. These studies demonstrate the generality of the DRM paradigm

to a non-verbal auditory domain, suggesting that general theories for DRM false memory

(see Brainerd & Reyna, 2005, for a review) may also apply to musical memory. However,

there are no studies in the published literature that have applied the misinformation paradigm

using music materials. This is surprising given the vast interest that misinformation effects

have received in the last decades (Figure D.1) and the implications that this paradigm has

had for many disciplines, including psychology, social sciences, and law (see Brainerd &

Reyna, 2005; Loftus, 2005; Neuschatz, Lampinen, Toglia, Payne, & Cisneros, 2017; Shaw,

2016; Scoboria et al., 2017, for reviews). The current research attempted to fill this gap by

creating a musical version of the misinformation paradigm as implemented in Loftus et al.

(1978) seminar paper, “Semantic integrations of verbal information into a visual memory”.

Figure D.1 Total number of publications on the misinformation effect from 1986 to 2017.

Data retrieved from Scopus on the 4th of July 2018. We searched for all available literature containing the

keyword “misinformation effect*” in the title, abstract, or author keywords. Only peer-reviewed articles

published in English from 1986 to 2017 were included, resulting in a total of 203 publications.

D.1.1 Practical implications

There are both real-world and theoretical implications of studying the misinformation effect

within music listening. When listening to music or attending a live music event, people are

often exposed to post-event information about the musical episode, such as descriptions by

134

D.1 Introduction

peers, reports and reviews from concerts, or judgments from music auditions and competi-

tions. In these situations, exposure to post-event misinformation could have an impact on

listeners’ attitudes and behaviours towards music, influencing their preferences, taste, and

habits. Evidence for this assumption comes from research on food preferences. For instance,

suggesting to people that they had previously become ill after eating a certain food (e.g.,

boiled eggs or pickles) affects how much of that food the person consumes in the present

(Bernstein & Loftus, 2009b; Bernstein, Laney, Morris, & Loftus, 2005; Bernstein, Scoboria,

& Arnold, 2015). Interestingly, this influence in the individual’s current behaviour towards

food can be reversed if the information suggests, instead, that he or she previously “loved”

the food in question, increasing their current preference and consumption.

Under examination situations or competitions, music performances are often evaluated from

memory. The outcome of these evaluative processes can be decisive for a musician’s career,

determining, for example, whether a student is accepted in a prestigious music university

or awarded in a competition situation. The Queen Elisabeth Musical Competition is one

of the most well-known international competitions for violin and piano, considered among

the most demanding in the world. Members of the examination board include some of the

worlds’ greatest musicians, teachers, and music critics (Flôres & Ginsburgh, 1996). To win

this competition would significantly change a musicians’ future career. In the first two

stages of the competition, the members of the jury listen to a number of different

performances each day. It is only at the end of the day that the jury provide the final scores

of the witnessed candidates. These evaluations reduce the total number of musicians to

the final list of candidates. In the last stage of the competition, the jury listens to two

candidates per day. Again, it is only at the end of each day that the jury provides their final

scores. The grades are given without any discussion between judges and cannot be changed

(Flôres & Ginsburgh, 1996; see Delhasse, 1985, for further information about the working

of the competition). Thus, it is likely that the evaluative process in this and similar music

competitions is susceptible to post-event misinformation.

There are even court cases where misinformation about music events might be influential,

such as court decisions on music plagiarism. In the context of western pop music, melodic

plagiarism has been a passionately debated phenomenon (Fruehwald, 1992; Müllensiefen &

Pendzich, 2009), often because of the royalties gained from author’s rights. In these cases, the

goal of the courts and the Copyright Office is to judge whether the melodic material of two

musical pieces (an original copyrighted piece and another performance by a different artist)

is sufficiently similar or not. Although in most plagiarism cases there are music recordings

that can be listened to repeatedly, there are also situations where recordings are not available.

135

D.1 Introduction

Thus, the evidence brought to court may instead simply rely on the memory of those who

have attended concert performances from where music is alleged to have been plagiarised.

For example, former Thin Lizzy guitarist Gary Moore had to pay damages after performing a

guitar solo on a recording considered to be plagiarised from a song written by the German

band Jud’s Gallery in 1974. This was even though the Jud’s Gallery song was not available

on record at the time of Gary Moore’s recording (Graham, 2008). During court proceedings,

Gary Moore was shown to be present at concert performances where the song was played,

and it was argued that post-hoc information could have distorted his musical memory.

D.1.2 Theoretical implications

Beyond these implications, a musical version of the misinformation paradigm could provide

valuable insights into the malleable nature of long-term representations of music. For in-

stance, it could shed light on the interplay between abstract structure and surface features,

the two types of musical information that people rely on when remembering music (Halpern

& Müllensiefen, 2008; Peretz, Gaudreau, & Bonnel, 1998; Peretz & Zatorre, 2005; Schel-

lenberg, Stalinski, & Marks, 2014; Trainor, Wu, & Tsang, 2004). A person’s ability to

remember music relies, to some extent, on surfaces features of the music. This includes the

exact key (pitch level), precise tempo (speed), and timbre (the instrument on which the

melody is performed). Nevertheless, the abstract structure of the music also plays an

important role, including the relative pitch patterns (i.e., the pitch distance between tones

regardless of their absolute pitch), relative durations (i.e., the ratios between durations of

subsequent notes regardless of their absolute length), and contour (i.e., the sequence of ups

and downs in a melody regardless of interval size). When processing the abstract structure of

a given piece of music, listeners’ commonly disregard surface characteristics. This is why

people can effortlessly recognise “Happy Birthday” when played at almost any tempo, on

any instrument, and sung at any pitch.

Based on this, one could expect misinformation effects to be likely in music listening: while

a general abstract representation of the music piece may remain intact, specific perceptual

features may be easily altered after exposure to post-event misinformation. However, we

know that listeners (even when musically untrained) can remember the precise key, tempo,

and timbre of familiar and unfamiliar music (Halpern & Müllensiefen, 2008; Levitin, 1994;

Levitin & Cook, 1996; Poulin-Charronnat et al., 2004). These studies show that changes in

these surface features significantly impair listeners’ ability to recognise music. Thus, it re-

mains unclear to what degree perceived surface features will be susceptible to misinformation

effects.

136

D.1 Introduction

Another important reason to study the misinformation paradigm within music listening tasks

is to demonstrate the generality of the misinformation effect to the non-visual auditory

domain, including theoretical accounts of false memory and general principles of memory. At

a general level, the existence of misinformation effects in music could indicate that Bartlett’s

(1932) view on the reconstructive nature of memory and schema-based effects (see Alba &

Hasher, 1983, for a review) also applies to memory for music. That is, musical events are not

stored in memory like songs on a CD. Instead, musical events are reconstructed from memory

using available schemata and knowledge structures. But demonstrating misinformation

effects in music could also allow us to test and generalise more specific accounts of false

memory to the music domain, such as the source-monitoring framework (SMF; Johnson,

Hashtroudi, & Lindsay, 1993; Johnson & Raye, 2000), or fuzzy trace theory (FTT; Brainerd

& Reyna, 2002; Reyna & Brainerd, 1995). Today, the extent to which these and other

theoretical accounts of false memory apply to the music domain is largely unknown.

D.1.3 Individual differences

Finally, a musical version of the misinformation paradigm could constitute a suitable individ-

ual

difference test to study listeners’ susceptibility to false musical memories. In fact, this

paradigm has been repeatedly used to examine individual differences and memory distortion

using visual materials, showing that not everyone is equally susceptible to false memory (see

Eisen, Winograd, & Qin, 2002; Loftus, 2005; Zhu et al., 2010, for reviews). The present

study focused on four individual difference aspects that could be potentially associated with

misinformation susceptibility in music, namely, age, personality, suggestibility, and musical

training.

Firstly, there is evidence indicating that age matters. In general, young children and the

elderly are more susceptible to misinformation effects than adolescents and adults (Davis

& Loftus, 2005; see Wylie et al., 2014, for a review). Secondly, studies have identified

associations between particular personality traits and false memory. For example, introverts

are more likely to be affected by false memories than extroverts (Loftus, 2005; Porter, Birt,

Yuille, & Lehman, 2000; Ward & Loftus, 1985), although this relationship was not supported

Liebman et al. (2002), who instead found associations between neuroticism and false

memory occurrence. Even though findings on the relationship between personality variables

and false memory are not always consistent, personality traits may be useful to increase our

understanding of the processes underlying memory distortion (Frenda, Nichols, & Loftus,

2011). The third individual difference factor is suggestibility, which has also been linked to

susceptibility to false memory. For example, false memory was positively associated with

137

D.1 Introduction

measures of social desirability (Tousignant, 1983, as cited in Schooler & Loftus, 1993) and

hypnotisability (e.g., Barnier & McConkey, 1992). Though potential associations between

false memory and these three individual differences (age, personality, and suggestibility)

have been demonstrated in past research (see Eisen et al., 2002; Loftus, 2005; Zhu et al.,

2010, for reviews), the links between these individual differences and false memory in music

have not yet been explored.

Musical training is another individual difference factor that could potentially mediate listeners’

performance in a musical version of the misinformation paradigm. Traditionally, studies on

musical memory have compared two groups of participants varying in their level of musical

expertise (see Talami, Altoè, Carreti, & Grassi, 2017, for a review): musicians (i.e.,

participants with expertise playing a musical instrument, determined by the number of years

of musical training or attendance to music conservatories or music schools) and

nonmusicians (i.e., participants with little or no experience of playing a musical instrument).

Using music stimuli (e.g., tones, chords, melodies), there is evidence indicating a superiority

of musicians over nonmusicians in short-term memory tasks (Bidelman, Hutka, & Moreona,

2013; Monahan, Kendall, & Carterette, 1987; Pallesen et al., 2010; Williamson, Baddeley,

& Hitch, 2010) and working memory tasks (Pallesen et al., 2010; Schulze, Mueller, &

Koelsch, 2011). However, in the domain of long-term memory the evidence is less consistent.

Cohen, Evans, Horowitz, and Wolfe (2011) and Weiss, Vanzella, Schellenberg, and Trehub

(2015) found empirical evidence suggesting that musicians long-term memory for melodies

is superior to nonmusicians, whereas Schiavio and Timmers (2016) found no differences

between these two groups. The paradigm used in the current study can shed light on this issue by

examining whether high levels of musical training reduce susceptibility to misinformation in

music recognition tasks. In this study, musical training is measured on metrical scales and,

therefore, it overcomes the traditional musician vs. nonmusician dichotomy.

D.1.4 Aims

The main aim of the present study was to investigate whether false memories can be induced

within music listening tasks through the misinformation paradigm. As false memory has

been consistently demonstrated within a wide range of visual scenarios (e.g., Brainerd &

Reyna, 2002; Brainerd & Reyna, 2005; Liebman et al., 2002; Loftus, 2005; Neuschatz et

al., 2017; Scoboria et al., 2017), it was hypothesised that participants will demonstrate

susceptibility to false memories in music listening tasks when presented with misinformation.

The second aim of the current research was to explore potential individual difference factors

related to misinformation susceptibility in music listening. Although this second analysis was

138

D.2 Methods

exploratory in nature, we had three hypotheses: (i) older participants will be more affected

by post-event misinformation, (ii) suggestibility will be positively correlated with the effect

of misinformation, (iii) and participants with more musical training will demonstrate less

misinformation susceptibility. The assessment of the different measurements of personality

traits and their potential to predict false memory occurrence was only exploratory within this

study and no hypotheses were presented.

D.2 Methods

D.2.1 Participants

A sample of 151 participants took part in the experiment. Of those, 143 disclosed their gender

(80 female, 63 male) and age, ranging from 18 to 63 (M = 31.04, SD = 8.10). A total of 93

participants were allocated to the misinformation group and a total of 58 participants were

allocated to the control group. Participants’ mean score in the Gold-MSI musical training

factor (Müllensiefen, Gingras, Musil, & Stewart, 2014) was 25.42 (SD = 11.36), which

indicates an overall average level of musical training, corresponding to the 44th percentile

of the data norms reported in Müllensiefen et al. (2014). In the misinformation group, 88

participants disclosed their gender (49 female, 39 male) and age, ranging from ages 21-63

(M = 33, SD = 8.27). In the control group, 55 participants disclosed their gender (31 female,

24 male) and age, ranging from ages 18-49 (M = 27.87, SD = 6.77). Participation was on a

voluntary basis and was unpaid.

An a-priori power analysis using an F-test for mixed within- and between-participants designs,

with two groups (misinformation vs. control) and eight within-participant measurements (the

total number of critical pairs), indicated a total sample size of at least 114 participants was

necessary to detect a significant main effect of misinformation. Based on the estimated effect

size from the results reported in Experiment 5 by Loftus et al. (1978), we set the effect size

to .20.

D.2.2 Design

The present study used a mixed within- and between-participants design. The within-

participant variables were the type of clip (critical vs. noncritical clips), music piece (Bebop

Jazz vs. Cool Jazz), and timbral manipulation (piano vs. drums). The between-participant

variable was the presence of misinformation (misinformation vs. control group). The

experiment had two parts. In each part, participants listened to and were tested on one

139

D.2 Methods

of the two music pieces, which unbeknownst to them was missing an instrument (either

piano or drums). The second part of the experiment used exactly the same procedure as

the first, but altered the music piece and instrumental manipulation. Thus, all participants

were tested two times, using two different music pieces and instrumental manipulations. The

order of presentation of the two pieces of music and the instrumental manipulation was fully

counterbalanced across participants.

Constructing an adapted musical version of the misinformation paradigm

The experimental design used in our study was inspired by Experiments 5 from Loftus et al.

(1978), where participants were first shown a series of 20 slides depicting an auto-pedestrian

accident: “A male pedestrian is seen carrying some items in one hand and munching on

an apple held with the other. He leaves a building and strolls toward a parking lot. In the lot,

a maroon Triumph backs out of a parking space and hits the pedestrian” (pp. 28-29). Four

of the 20 slides were critical. Each version of the critical slides contained a particular object

(“a pair of skis leaning against a tree”), whereas the other version contained the same identical

slide with a changed detail (“a shovel leaning against a tree”). Participants only saw one

version of the critical slides. Following a series of filler activities for approximately 10

minutes, participants read a three-paragraph description of the event. The description

contained four critical sentences that either did or did not mention the incorrect critical

objects (e.g., “if the participant had seen skis leaning against a tree, the statement might

include a sentence that mentioned the shovel leaning against the three). After another filler

task of 10 min approximately, participants were given a two forced-choice recognition test,

where they had to indicate which of the two slides on each pair was seen before. The test had

an overall of 10 pairs of slides. Four of the 10 pairs were critical, containing a slide depicting

the event as it corresponded with the incorrect information and another slide depicting the

actual event. The authors found that the percentage of times a correct selection occurred was

significantly lower (55.3%) when the information was misleading than when it was not

(70.8%). This difference was statistically significant (Loftus et al., 1078).

The main challenge of constructing an adapted musical version of this paradigm was to use

music stimuli instead of visual. In the exposure stage, instead of seeing 20 slides depicting

an auto-pedestrian accident, participants listened to an instrumental piece of jazz music,

which unbeknownst to them was missing an instrument, either piano or drums. Instead of

manipulating 4 critical pieces of information, we only manipulated one: the presence or

absence of an instrument (piano or drums). While the visual presentation of 20 different

slides allowed for different manipulations of critical information, the auditory presentation

140

D.2 Methods

of a single piece of music did not. In the post-event information stage, participants read a

descriptive text with a single critical piece of information, which could either suggest the

presence of the missing instrument or did not. Finally, we tested participants using a two

forced-choice recognition task analogous to Loftus et al. (1978). On this recognition task,

participants had to decide which clip of the pair would correspond to the original recording

they had heard. Four of the 10 pairs of clips were critical, containing a clip from the original

track missing the target instrument (correct response) and a clip from the original track

containing the target instrument (incorrect response). The order of presentation of the 10

trials as well as the order of presentation of the two clips within each trial was randomised

for each participant.

D.2.3 Materials

Music stimuli

Two core tracks were used from the MedleyDB database (Bittner et al., 2014) by the artist

“Music Delta”, namely the track “Bebop Jazz”, and the track “Cool Jazz”. MedleyDB is a

dataset of annotated, royalty-free multitrack recordings, which were curated primarily to

support music research (Bittner et al., 2014). Both tracks were instrumental jazz pieces,

similar in style and using the same instrumentation. The original complete tracks both

featured a drum set, a double bass, a piano, and a brass section. The original version of

“Bebop Jazz” was 1 minute 43 seconds in length, and the original version of “Cool Jazz”

was 1 minute 42 second in length. Alternative mixes were created using the Ableton Live

software featuring the core tracks either without drums or without piano (all other instruments

and attributes remained the same). Each track was then shortened to 60 seconds, featuring

50 seconds of the original track from the start, and then a gradual 10-second fade out. In

addition, we used a third track from the MedleyDB database called “Fusion Jazz”. This track

was also an instrumental jazz track featuring drum set and saxophone, though it also featured

electric bass, electric piano and synthesizer (rather than only acoustic instruments), creating

a clear distinction with the original tracks.

To create the testing stimuli, the original and alternative mixes were used along with the third

track to create three types of pairs of clips: (i) critical pairs (i.e., a clip from the original track

either missing piano or drums and a clip from the same original track including the suggested

target instrument), (ii) noncritical-brass pairs (i.e., a clip from the original track either missing

piano or drums and a clip from the same original track missing the target instrument as well

as an additional instrument, namely, the brass section), and (iii) noncritical-easy pairs (i.e.,

a clip from the original track missing piano or drums and a clip from a different hitherto

141

D.2 Methods

unheard but stylistically similar track). The critical clips were used to measure the effect of

post-event information on memory, whereas the noncritical clips were used to measure

memory performance in the absence of misinformation. For each track and instrument

condition, there was a total of 10 pairs of clips: four critical and six noncritical pairs (three

noncritical-brass and three noncritical-easy pairs).

Post-event information

Descriptive texts consisting of a single paragraph of information on the original track were

created. Except for the instrumentation referenced - which was incorrect only for the

misinformation group - this text was wholly accurate and was identical across conditions. To

increase credibility, the paragraph also included a web citation, pointing to the Music Delta

website. To ensure that participants read the post-event information, a sentence from the

descriptive text was copied at the bottom of the survey page, with a blank space instead of

one of the words for the participant to fill in and show they had read the text (see Appendix

A, in the paper published online, for the descriptive texts of all conditions, including the fill-

in-the-blank sentence).

Individual difference factors

A variety of self-reporting questionnaires were used to measure individual difference factors.

The Goldsmiths Musical Sophistication Index (Gold-MSI; Müllensiefen et al., 2014), a mea-

surement instrument used to assess musical skills in the general population. The dimensions

of musical sophistication measured by the Gold-MSI are (1) Active Engagement, Perceptual

Abilities, Musical Training, Emotion, Singing Abilities, and General Musical Sophistication.

Eight items from the Susceptibility to Persuade Strategies Scale (STPS; Kaptein, De Ruyter,

Markopoulos, & Aarts, 2012; Eren, Unal & Temizel, 2014) that provide a measurement of

general suggestibility in relation to two of the six persuasion principles identified by Cialdini

(2001), namely, bias to authority and consensus. The Social Desirability Scale-17 (SDS-17;

Stöber, 1999, 2001), which captures the tendency for individuals to self-describe with socially

desirable attributes, as per Paulhus’ impression management construct (1986).The Big Five

Inventory (BFI; John & Srivastava, 1999), which measures an individual on the Big Five

Dimensions of personality defined by Goldberg (1993), namely, Extraversion, Introversion,

Agreeableness, Conscientiousness, Neuroticism, and Openness. And the Revised Short Test

of Music Preference (STOMP-R; Rentfrow & Gosling, 2003), measuring music preferences

across 23 genres, which represent four higher-order dimensions of music preference. These

four dimensions of music preference are named Intense and Rebellious, Upbeat and Conven-

142

D.2 Methods

tional, Energetic and Rhythmic, and Reflective and Complex (including preferences for jazz).

These questionnaires were provided to fill the time between the exposure and the recognition

test

stages.

D.2.4 Procedure

The core procedure for the experiment (shown in Figure D.2) involved participants listening

to an initial track (“Cool Jazz” or “Bebop Jazz”), which unbeknownst to them was missing

an instrument (either piano or drums). At this stage, participants were provided with the

following instructions: “you will listen to a piece of music; you will later be tested on

your memory for this piece of music, so please concentrate on the piece closely”. After

listening to the piece of music, participants had to complete a filler task, consisting of filling

several questionnaires for approximately 5-minutes. Following the filler task, participants

had to read a descriptive text of the original track which either suggested the presence of the

missing instrument in the original track (misinformation group) or did not (control group).

Following another set of filler questionnaires for approximately 5-minutes, participants took

the music recognition task, a 2-alternative-forced-choice recognition test of 10 pairs of

clips, where participants had to choose the clip that was presented previously. In this

recognition test, four pairs of clips were critical and six were noncritical. The order of the

pairs, as well as the position of the two clips on the screen (left or right), were randomized

for each participant. The entire procedure was repeated for a second time in part 2, where

the initial track (“Cool Jazz” or “Bebop Jazz”) and the instrumental manipulation (piano

or drums) was counterbalanced. Participants were sent the experiment to complete online

via a single link, with an explanation that this was a test on memory and music, along with

instructions providing an outline of the test structure. The experiment was constructed on

the online survey software Qualtrics (Qualtrics, Provo, UT). The experiment was granted

ethical clearance by the Ethics Committee of the Department of Psychology, University of

Goldsmiths, London, on 5 May 2017.

143

D.2 Methods

Figure D.2 Diagram of the misinformation paradigm procedure used in the experiment.

The entire procedure was repeated once after a brief break, using a new track as well as a different instrumental

manipulation (drums or piano).

D.2.5 Statistical analysis

Three participants were excluded from the subsequent analysis because they scored 0%

in either the noncritical-easy clips or the noncritical-brass clips in the two parts of the

experiment, indicating that they did not understand the experimental task. The average time

to complete the experiment was 105.36 minutes (SD = 342.88). Five participants took more

than two standard deviations from the mean to complete the survey and were excluded,

resulting on a total average time of 54 minutes (SD = 85.79). Thus, the subsequent analysis

included an overall of 143 participants, 88 in the misinformation group and 55 in the control

group.

Matching procedure

Because individual differences in age have been associated with the susceptibility to false

memories (see Wylie et al., 2014, for a review) and individual’s musical training plays an

important role in the ability to remember music (see Talami et al., 2017, for a review), we

carried out a logistic regression analysis to investigate whether the two groups (misinforma-

tion vs. control) differed on age and musical training. The binary dependent variable was the

group, and the two predictors were age and musical training. Table 1 shows the output of

144

D.2 Methods

the model, which indicates that the two groups differed significantly on age (misinformation

group, M =32.96, SD = 8.39; control group, M =27.71, SD = 6.33; p = .002), but not in

musical training (misinformation group, M =23.66, SD = 11.00; control group, M =28.40,

SD = 11.53; p = .06).

In order to correct for age differences, we used a nonparametric multivariate procedure

to match the two samples on age and musical training, as implemented in the R package

MatchIT (Stuart, King, Imai, & HO, 2011), which matches participants in two groups on

the basis of several covariates at once. We used the nearest neighbour matching method and

matched with the replacement method to enable one-to-many matching to accommodate the

different sample sizes. The optimal solution to match participants in the two groups to

correct for age differences excluded a total of 23 participants. After this procedure, we

repeated the logistic regression analysis with the matched dataset. The two groups did not

differ significantly in age (misinformation group, M =29.97, SD = 4.74; control group, M

=27.71, SD = 6.33; p = .09), nor in musical training (misinformation group, M =23.68, SD =

11.68; control group, M = 28.92, SD = 11.50; p = .05). Table D.1 shows the output of the

logistic regression model with the matched dataset. Thus, the final sample size for the main

analysis regarding the effect of misinformation comprised a total of 120 participants (65 in

the misinformation group and 55 in the control group).

Table D.1 Summary of the logistic regression analyses before and after the matching

procedure.

The misinformation effect

The main analyses planned initially to measure misinformation effects in music listening

followed the analysis strategy used in Experiment 5 by Loftus et al. (1978). This analysis

aimed to compare our results, using music stimuli, with those reported in Loftus et al. (1978),

using visual stimuli. This analysis only considered the critical pairs (i.e., a clip from the

original track either missing piano or drums and a clip from the same original track including

145

D.2 Methods

the suggested target instrument in the misinformation text). The dependent variable was the

total number of correct and incorrect responses, collapsed across all critical pairs in the

two parts of the experiment. The independent variable was the experimental group

(misinformation group vs. control group). An independent t-test was employed to test

for significant differences between groups. We reported effect sizes in terms of Cohen’s d

because this measure can be used across several model types and is intuitively understood by

most researchers. Moreover, we calculated the odds ratio to compare the effect sizes in our

study and the ones reported in Loftus et al., (1978).

To expand this analysis using more advanced statistical techniques and also accounting for

the timbral manipulation and different types of music clips, we performed an exploratory

analysis using mixed-effects logistic regression as implemented in the R packages lme4

(Bates, Mächler, Bolker, & Walker, 2015) and car (Fox et al., 2017). Mixed-effects logistic

regression models have several advantages compared to ordinary logistic regression models.

They can handle missing values and binomial or non-normal distributions, do not assume

independence among observations, and can work with correlated observations. Mixed-effects

logistic regression can also model random variability by assuming random intercepts for

different relevant factors, such as participants’ memory abilities, providing unbiased estimates

the coefficients of the predictor variables (Baayen, Davidson, & Bates, 2008; Pinheiro &

Bates, 2000).

This analysis was exploratory and different combinations of predictor and outcome variables

were considered. In this paper, we report what we consider the most comprehensive model,

including fixed effect factors for misinformation (presence vs. absence), type of clip (critical

vs. noncritical), timbral manipulation (piano vs. drums), and the interaction between

misinformation and type of clip. The binary response on each pair of clips (correct or

incorrect) was the dependent variable. In addition, we specified a random intercept for

participants because the individual ability of the participants to perform on the recognition

task contributed to the variance of the responses, inflating the overall variance of the data.

The interaction term was used to study the effect of misinformation in the two types of pairs

of clips, whereas the timbral manipulation factor was included to test whether memory

performance was affected by the instrument under manipulation. The mixed-effects analysis

was conducted using effects coding as opposed to the default treatment coding and Type-III

Wald chi-square tests. The mixed-effect logistic regression model met the assumptions of

linearity and normality. Collinearity was not an issue because the predictor variables were

part of the orthogonal experimental design and, therefore, the association between these

146

D.2 Methods

variables was 0. In addition, there were no correlations between residuals and the error

variance of the residuals did not change across the range of fitted values.

Individual differences

A person-wise dependent variable was created, using the ratio of the individual’s overall

performance on the critical clips, divided by their performance on the noncritical clips. Ratio

values of >1 suggested low susceptibility to false memories, whereas values of <1 indicated

high susceptibility. For this analysis, we only used participants in the misinformation group,

as they were the only exposed to incorrect information. We used the dataset from before the

matching procedure was carried out, comprising a total of 88 participants. The subsequent

analysis only includes the individual difference factors for which we had clear previous

hypothesis based on the published literature (i.e., age, musical training, suggestibility, and

personality; see Appendix B, in the paper published online, for a correlation matrix of the

person-wise dependent variable and all 17 individual difference variables measured during

the filler task phases of this study).

To examine which individual difference factors were associated with misinformation sus-

ceptibility, we used a data analysis method known as random forest (Breiman, 2001), based

on permutation tests, as implemented in the R package party (Hothorn, Bühlmann, Dudoit,

Molinaro, & Van Der Laan, 2006; Hothorn, Hornik, & Zeileis, 2006; Strobl, Boulesteix,

Kneib, Augustin, & Zeileis, 2008; Strobl, Malley, & Tutz, 2009). Compared to other classifi-

cation and regression methods, random forests have several advantages, as they can handle

complex interactions, large sets of predictor variables (even if they are highly correlated),

and do not assume a linear relationship between predictors and dependent variable (see

Hastie, Tibshirani, & Friedman, 2009). Moreover, random forest models use an in-built out-

of-the-bag cross-validation mechanism that protects against alpha error inflations and

overfitting.

The person-wise memory score was the dependent variable and the individual difference

variables the predictors, including age, musical training, suggestibility (SPSS and SDS-17),

and personality traits (i.e., Extraversion, Agreeableness, Conscientiousness, Neuroticism,

and Openness). The model was run with a size of 10,000 trees. The number of randomly

preselected predictor variables to be chosen in each split was 4. The R2 was calculated using

the R package caret (Kuhn, 2008), which uses cross-validation and prevents model

overfitting.

147

D.3 Results

A measure of variable importance for each predictor was used to produce unbiased estimates

even when there are significant correlations between predictor variables and/ or when the

dependent variable is very unequally distributed (Janitza, Strobl, & Boulesteix, 2013). The

variable importance score for each predictor variable describes how predictive the variable in

question was relative to the other predictors. This includes their influence as main and in

interaction effects. To select those variables that were associated with misinformation effects,

used a confidence interval criterion (Strobl et al., 2008; Strobl et al., 2009). This criterion

indicates that only those variables whose importance score is positive and greater than the

absolute value of the lowest negative variable importance score should be selected.

D.3 Results

The percentage of times a correct response occurred was 64.23% (SD = 22.95) when the

descriptive text contained incorrect information and 74.09% (SD = 23.18) when it did not.

This difference was statistically significant, as indicated by an independent t-test, t(118) =

2.33, p = .02. The effect size was small to medium, d = .43. Based on the number of correct

and incorrect selections in the presence and absence of misinformation, the odds ratio was

also calculated, OR = 1.59, log odds = .46, 95% CI [.87, 2.91].

To study whether the matching procedure affected the outcome of the main analysis regarding

the misinformation effect, we repeated the same analysis using the dataset before the matching

procedure was carried out (N = 143). Overall, the results were very similar. The percentage of

times a correct response occurred was significantly lower when the descriptive text contained

incorrect information (M = 63.09, SD = 22.05) than when it did not (M = 74.09, SD = 23.18),

t(141) = 2.63, p = .009; d = .49; OR = 1.61, log odds = .48, 95% CI [.88, 2.96].

The results of the exploratory analysis using mixed-effects logistic regression can be seen in

Figure D.3, which shows the proportion of correct selections in the presence and absence

of misinformation in the two types of clips (critical vs. noncritical) and the two timbral

manipulations (piano vs. drums). The mixed-effects logistic regression analysis revealed a

main effect of misinformation, X

2(1) = 5.96, p = .01, type of clip X

2(1) = 191.25, p <. 001,

instrumentX 2(1) = 22.55, p <. 001, but the interaction between misinformation and type of

clip was nonsignificant, X

(1) = .00, p = .99. The classification accuracy of the model 1 was

.84.

To test whether the order of the experiment (part 1 vs. part 2) had an effect on participants’

performance on the music recognition task, we ran a mixed-effects logistic regression

analysis with order, misinformation, and the order-misinformation interaction as fixed factors.

148

D.3 Results

Participants ID was used as a random effect factor. The model indicated that the order of the

experiment, clip X 2(1) = .22, p = .63, and the interaction term, clip X 2(1) = .1.32, p = .25,

were nonsignificant, whereas the main effect of misinformation was significant, clip X

(1) =

6.88, p =. 008.

Because timbral manipulation was significant we performed mixed-effect logistic regression

models to study the effects of misinformation in the two timbral manipulation. In the two

models, misinformation and the interaction between misinformation and type of clip were

the fixed factors, whereas participants the random effect factor. The mixed-effects logistic

regression model with piano indicated a main effect of misinformation, X

2(1) = 5.87, p =

.01, and type of clip X 2(1) = 76.64, p <. 001, but no significant interaction between these

two factors, X

(1) = .00, p = .97. The classification accuracy of model 2 was .86. The model with

drums revealed a main effect type of clip X

(1) = 110.50, p <. 001, but misinformation,

X 2(1)

= 2.01, p = .16, and the interaction between these two factors, X 2(1) = .01, p = .92, were

nonsignificant. The classification accuracy of model 2 was .88.

The random forests analysis revealed that none of the 9 individual difference factors were

significantly associated with misinformation susceptibility. None of the positive variable

importance scores were greater than the absolute value of the lowest negative variable

importance score, indicating the person-wise dependent variable was not associated with any

of the predictor variables. The overall R2 of the random forest model was .19.

149

D.4 Discussion

Figure D.3 Proportion of correct selections in the presence and absence of misinformation

in the two types of clips and the two timbral manipulations.

D.4 Discussion

The main aim of the present study was to investigate whether false memory can be induced

within music listening tasks through the misinformation paradigm. The presence of post-event

misinformation significantly deteriorated participants’ performance in a music recognition

task. Participants were more likely to select the wrong music clip (i.e., a clip containing an

instrument that was never actually experienced) when the descriptive text included incorrect

information suggesting the presence of the target instrument (36%) than when it did not (26%).

This finding supports the malleable nature of long-term memory for music and suggests the

existence of false musical memories. Regarding our initial question, to what extent are false

musical memories consistent with findings from the visual domain?, the results reported in

this study are congruent with research on the misinformation effect in visual recognition

tasks (see Frenda et al., 2011; Loftus, 2005; Pickrell, et al., 2016, for reviews) and more

broadly with the existence of false memories (Brainerd & Reyna, 2002; Brainerd & Reyna,

2005; Neuschatz et al., 2017; Shaw, 2016; Scoboria et al., 2017). Therefore, the findings

reported here are also subject to the misinformation debate and related issues, such as whether

150

D.4 Discussion

impairment by post-event information exists or not (Loftus, Schooler, & Wagenaar, 1985;

McCloskey & Zaragoza, 1985). To the best of our knowledge, this is the first published study

demonstrating, at least partly, the generality of the misinformation effect to a non-visual

auditory domain.

The present study attempted to create a musical version of the misinformation paradigm

based on Experiment 5 in Loftus et al. (1978). Nevertheless, there are important differences

to consider between these two studies. While we used a design in which the presence of

misinformation was manipulated between-participants, Experiment 5 in Loftus’ et al. (1978)

used a repeated-measures design. Moreover, we used shorter filler tasks (approximately 5

minutes each) compared to Loftus et al. (1978; approximately 10 minutes each); and allowed

participants to listen to the clips multiple times. Despite these differences, the misinformation

effect reported in Loftus et al. (1978) and the one observed in the current research were

fairly similar in size, as indicated by the odds ratio (1.59 in the present study and 2.12 in

Loftus’ Experiment 5). Using visual stimuli, participants in Loftus’ study (1978) selected the

right clip 53% of the times in the presence of misinformation and 71% in the absence. In the

present study, using musical materials, participants selected the right clip 64% of the times

when the text included misinformation and 74% when it did not.

The initial analysis used to measure misinformation effects in music listening was based on

Loftus’ et al. (1978) and, therefore, only considered the proportion of wrongly selected clips

on the critical trials. Accordingly, to establish a misinformation effect, the proportion of

wrongly selected critical clips should be significantly higher on the misinformation group

than in the control group. Using this criterion, the misinformation effect in our dataset was

clear. Nevertheless, we carried out a second exploratory analysis using a more comprehensive

statistical model (i.e., mixed-effects logistic regression) and accounting also for type of music

clips (critical vs. noncritical) and type of misinformation (piano vs. drums). Note that the

critical trials were designed explicitly to measure misinformation effects (i.e., including the

target instrument in one of the two clips), whereas the noncritical clips were designed to

measure general musical memory ability.

Thus, a second criterion could be used to determine whether the misinformation effect was

established or not. That is, the increase in wrongly selected clips from noncritical to critical

trials should be significantly higher only for those participants in the misinformation group.

According to this, we were not able to establish a statistically significant misinformation effect.

In other words, the interaction between type of clip and misinformation was nonsignificant.

However, there was a clear tendency supporting the interaction term between misinformation

and

type of clip (Figure D.3). Moreover, it is likely that the presence of misinformation

151

D.4 Discussion

increased the overall difficulty of the test only in the misinformation group, as they had to read

and process counterfactual information. Thus, the interference produced by the post-event

misinformation could have resulted in an overall poorer performance in both critical and

noncritical pairs compared to the control group. This second analysis was only exploratory

and future research could improve our design by planning this type of analysis in advance

and using a measurement of general long-term memory for music in which misinformation is

not

confounded, such as Gordon (1989) or Harrison, Collins, and Müllensiefen (2017). We also

encourage future researchers to use within-participants designs to avoid the comparison of

independent groups as well as use more comprehensive analysis strategies that take the

performance on critical and noncritical clips into account within the same model.

In addition, the mixed-effects logistic regression analysis also revealed a significant main

effect of instrumental manipulation. Participants performed significantly better when the

misinformation paradigm involved a piano manipulation than when it involved drums. To ex-

plore this further we conducted two separate analyses for each instrument (piano and drums).

Results indicated that misinformation regarding the presence of piano had a significant effect

on participants’ performance in the recognition task, whereas misinformation suggesting the

presence of drums did not. A potential explanation for this could be that drums are

perceived

as a louder and more prominent instrument than piano, occupying a larger range of

the frequency

spectrum in a recorded track. Because drums are perceptually a more obvious element to be

missing or introduced into a musical piece, this timbral manipulation may be

less susceptible

to misinformation effects compared to a less prominent timbral manipulation,

such as piano.

However, this distinction on the two types of timbral manipulation did not affect the overall

misinformation effect in the general model (including both piano and drums). This might be

due to a clear trend supporting the misinformation effect in the drums manipulation as well

(Figure D.3).

The present research focused on musical timbre and instrumentation, which has been iden-

tified as a particularly important surface feature influencing the ability to remember music

(Halpern & Müllensiefen, 2008; Poulin-Charronnat et al., 2004; Schellenberg & Habashi,

2015; Trainor et al., 2004). For example, research shows that participants’ long-term memory

for

music deteriorates significantly when the instrumentation used at the recognitions test phase

does not match the one used at the learning phase (Halpern & Müllensiefen, 2008; Poulin-

Charronnat et al., 2004). In addition, there is evidence that information about the key and

tempo is forgotten at a faster rate than information about the timbre (Schellenberg &

Habashi,

2015), with infants even retaining memory of timbre after 1 week of daily exposure

to music

(Trainor et al., 2004). Results from our study indicate that post-event information

152

D.4 Discussion

about instrumentation significantly interfered with the actual surface information encoded at

the exposure phase.

When creating an adapted musical misinformation paradigm, there were practical reasons to

manipulate timbre instead of other surface features. In order to manipulate misinformation in

a given piece of music, a musical feature was required that could be addressed specifically in

the

post-event descriptive text and could be later scored unambiguously as correct/ incorrect in

the recognition test. This is most easily done by using a categorical feature such as the

presence of an instrument which most people, even without musical training, are able to

identify. Extracting other types of information from a musical excerpt (e.g., key, mode,

or tempo) requires special musical skills or training and would have limited our choice of

participants and, therefore, the wider applicability of this study. Moreover, manipulating the

presence or absence of an instrument in a piece of music is analogous to the missing objects

manipulated in Loftus et al. (1978).

Nevertheless, there is a wide range of other musical aspects that could be manipulated, if

these can be made congruent with the practical constraints of designing a task that does not

require special musical trainings and skills. For example, the intonation accuracy of the

vocalist, appropriateness of tempo, expressivity, emotionality, overall quality, and aesthetical

aspects of the performance. Moreover, when creating false memories of music listening

episodes, the perceived familiarity of the tracks and the jazz genre used may have played

a role. Jazz music remains a relatively specialist and less popular music genre in the West,

with jazz album sales accounting for only 2.1% of the total albums sold in the USA in 2014

(Nielsen, 2015). Thus, replications of this study should consider the use of a wider range of

types of misinformation as well as music genres and styles in order to bolster the effects

observed, and allow for a more nuanced understanding of what specific factors contribute to

the fallibility of musical memory. Looking beyond misinformation there are clearly more

experimental paradigms that could be employed to study false memory in music. For instance,

Curtis and Bharucha (2009) and et al. (2014) studied false memories in music using an

adapted version of Roediger and McDermott’s paradigm (1995). The authors found that

participants falsely remembered more notes in situations where the target note was congruent

with the context (more expected) than when it was incongruent (less expected).

The existence of misinformation effects in music listening situations sheds light into the

nature of long-term memory for music. Our results suggest that memory for music is, to some

extent, malleable and susceptible to post-event misinformation about surface features, such

as timbre. Participants failed to accurately monitor the source of information, misattributing

information from the musical source to the verbal one. This finding is in line with the source-

153

D.4 Discussion

monitoring framework (SMF; Johnson, Hashtroudi, & Lindsay, 1993; Johnson & Raye,

2000) and fuzzy trace theory (FTT; Brainerd & Reyna, 2002, 2004; Reyna & Brainerd, 1995).

According to FTT, there are two parallel types of memory, namely, verbatim and gist. While

verbatim memory represents the surface details of physical stimuli, gist memory represents

the main meaning or theme. These two types of memories are encoded separately and can be

retrieved independently (Brainerd & Reyna, 2002).Thus, false musical memory may have

occurred because verbatim (surface features) declined faster than gist (abstract structure) and

integrated with schematic-gist information (Brainerd & Reyna, 2002). Although providing a

comprehensive theoretical account of the effect of misinformation in music is beyond the

scope of this study, we consider that efforts in this direction are essential to demonstrate the

generality of theoretical accounts of false memory and general principles of memory to the

non-visual auditory domain.

Finally, the present study aimed to explore potential individual difference factors related to

the susceptibility to false memories in music listening. Based on previous literature (see Eisen

al., 2002; Loftus, 2005; Talami et al., 2017; Zhu et al., 2010, for reviews), we explored the

role of age, musical training, suggestibility, and personality traits. Contrary to our

hypotheses, we found no evidence to support that any of these individual difference factors

were significantly associated with misinformation susceptibility in music. It is important to

mention, however, that this analysis was purely exploratory and one would need to devise

and carefully calibrate a proper individual difference test of misinformation susceptibility in

music in order to confirm and generalise these results. Moreover, we did not investigate a

sample of professional musicians and this finding may change when testing individuals with

greater musical expertise. However, this result is in line with Anglada-Tort and Müllensiefen

(2017), who also showed that musical expertise did not have a protective effect against a

musical memory illusion. It also supports Schiavio and Timmers (2016), who did not find an

advantage of musicians’ long-term memory over nonmusicians.

Overall, our findings support, at least partly, previous research that post-event misinformation

has a significant effect on the reliability of memory, suggesting that false memory can be

induced in music listening tasks. When people listen to music or experience music in a

live performance, they are normally exposed to related information at some point after

the event. In our experimental setting, the presence of post-event misinformation about

instrumentation impaired listeners’ ability to remember music. Participants used verbal

information that was never actually experienced to reconstruct a memory of a piece of music,

demonstrating the generality of the misinformation effect to the non-visual auditory domain.

Furthermore, a random forest analysis indicated that the misinformation effect occurred

154

D.5 References

regardless of participants’ levels of musical training, suggestibility, age, and personality

traits. The existence of misinformation effects in music listening situations has implications

for any area in which musical memory is involved, including aesthetics, music education,

performance evaluation, preferences for music, marketing, and advertising. We conclude

that memory for non-visual auditory stimuli can be fallible and the extent to which humans

can memorise and remember music reliably should be, at least, questioned and further

investigated.

D.5 References

Alba, J. W., & Hasher, L. (1983). Is memory schematic? Psychological Bulletin, 93, 203-231.

Anglada-Tort, M., & Müllensiefen, D. (2017). The repeated recording illusion: The effects

of extrinsic and individual difference factors on musical judgements. Music Perception:

An Interdisciplinary Journal, 35(1), 92-115.

Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models

using lme4. Journal of Statistical Software, 67(1), 1-48.

Baayen, R. H., Davidson, D. J., & Bates, D. M. (2008). Mixed-effects modeling with crossed

random effects for subjects and items. Journal of memory and language, 59(4), 390-412.

Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology.

Cambridge: Cambridge University Press.

Barnier, A. J., & McConkey, K. M. (1992). Reports of real and false memories: The

relevance of hypnosis, hypnotizability, and context of memory test. Journal of Abnormal

Psychology, 101(3), 521.

Bernstein, D., & Loftus, E. F. (2009a). How to tell if a particular memory is true or false.

Perspectives on Psychological Science, 4, 370–374.

Bernstein, D. M., & Loftus, E. F. (2009b). The consequences of false memories for food

preferences and choices. Perspectives on Psychological Science, 4(2), 135-139.

Bernstein, D. M., Laney, C., Morris, E. K., & Loftus, E. F. (2005). False memories about

food can lead to food avoidance. Social Cognition, 23(1), 11-34.

Bernstein, D., Scoboria, A., & Arnold, R. (2015). The consequences of suggesting false

childhood food events. Acta Psychologica, 156, 1–7.

Brainerd, C. J., & Reyna, V. F. (2002). Fuzzy-trace theory and false memory. Current

Directions in Psychological Science, 11(5), 164-169.

Brainerd, C. J., & Reyna, V. F. (2005). The science of false memory. Oxford: Oxford

University Press.

155

D.5 References

Bidelman, G. M., Hutka, S., & Moreno, S. (2013). Tone language speakers and musicians

share enhanced perceptual and cognitive abilities for musical pitch: evidence for

bidirectionality between the domains of language and music. PloS one, 8(4), e60676.

Bittner, R., Salamon, J., Tierney, M., Mauch, M., Cannam, C. & Bello, J.P. (2014, October).

MedleyDB: A Multitrack Dataset for Annotation-Intensive MIR Research (pp.155-160).

Paper presented at the 15th International Society for Music Information Retrieval

Conference, Taipei, Taiwan.

Breiman, L. (2001). Random forests. Machine Learning, 45, 5-32.

Cialdini, R.B. (2001). Harnessing the Science of Persuasion. Harvard Business Review,

79(9), 72-81.

Cohen, M. A., Evans, K. K., Horowitz, T. S., & Wolfe, J. M. (2011). Auditory and visual

memory in musicians and nonmusicians. Psychonomic bulletin & review, 18(3), 586-

591.

Costa, P. T., & McCrae, R. R. (1992). Normal personality assessment in clinical practice:

The NEO Personality Inventory. Psychological assessment, 4(1), 5-13.

Curtis, M. E., & Bharucha, J. J. (2009). Memory and musical expectation for tones in cultural

context. Music Perception: An Interdisciplinary Journal, 26(4), 365-375.

Davis, D., & Loftus, E. F. (2007). Internal and external sources of misinformation in adult

witness memory. In M. P. Toglia, J. D. Read, D. F. Ross, & R. C. L. Lindsay (Eds.), The

handbook of eyewitness psychology, Vol. 1. Memory for events (pp. 195-237). Mahwah,

NJ:

Lawrence Erlbaum Associates.

Delhasse, P. (1985). Le Concours Reine Elisabeth des Originesa Aujord’hui. Bruxelles:

Vander.

Eisen, M. L., Winograd, E., & Qin, J. (2002). Individual differences in adults’ suggestibility

and memory performance. In M. L. Eisen, J. A. Quas, & G. S. Goodman (Eds.), Memory

and suggestibility in the forensic interview. New York, NY: Routledge.

Eren, P. E., Temizel, T. T., & Unal, P. (2014, May). An Exploratory Study on the Outcomes

of Influence Stra-tegies in Mobile Application Recommendations. In Proceedings of

the Second International Workshop on Behavior Change Support Systems (BCSS2014),

Padova, Italy.

Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical

power analysis program for the social, behavioral, and biomedical sciences. Behavior

Research Methods, 39, 175-191.

Flôres Jr, R. G., & Ginsburgh, V. A. (1996). The Queen Elisabeth musical competition: how

fair

is the final ranking?. The Statistician, 97-104.

156

D.5 References

Frenda, S. J., Nichols, R. M., & Loftus, E. F. (2011). Current issues and advances in

misinformation research. Current Directions in Psychological Science, 20(1), 20-23.

Fruehwald, E. S. (1992). Copyright infringement of musical compositions: A systematic

approach. Akron Law Review, 26(1), 15–44.

Gabbert, F., Memon, A., & Allan, K. (2003). Memory conformity: Can eyewitnesses

influence each other’s memories for an event? Applied Cognitive Psychology, 17(5),

533-543.

Graham, D. (2008, December 04). Ex-Thin Lizzy guitarist loses German plagiarism case.

Retrieved from https://www.reuters.com/article/us-moore/ex-thin-lizzy-guitarist-loses-

german-plagiarism-case-idUSTRE4B28HQ20081203

Goldberg, L. R. (1993). The structure of phenotypic personality traits. American Psychologist,

48(1), 26.

Gordon, E. E. (1989). Advanced measures of music audiation. Chicago, IL: GIA Publicaitons.

Halpern, A. R., & Müllensiefen, D. (2008). Effects of timbre and tempo change on memory

for music. The Quarterly Journal of Experimental Psychology, 61(9), 1371-1384.

Harrison, P. M., Collins, T., & Müllensiefen, D. (2017). Applying modern psychometric

techniques to melodic discrimination testing: item response theory, computerised

adaptive testing, and automatic item generation. Scientific reports, 7(1), 3618.

Hastie, T., Tibshirani, R., & Friedman, J. (2009). Hierarchical clustering. In T. Hastie, E.

Tibshirani, & J. Friedman (Eds.), The elements of statistical learning: Data mining,

inference and prediction (2nd ed., pp. 520-528). New York: Springer.

Hothorn, T., Bühlmann, P., Dudoit, S., Molinaro, A., & Van Der Laan, M. (2006). Survival

ensembles. Biostatistics, 7, 355-373.

Hothorn, T., Hornik, K., & Zeileis, A. (2006). Unbiased recursive partitioning: A conditional

inference framework. Journal of Computational and Graphical Statistics, 15, 651-674.

Hyman, I. E., Husband, F., & Billings, J. (1995). False memories of childhood experiences.

Applied Cognitive Psychology, 9, 181–197.

Hyman, I. E. Jr., & Kleinknecht, E. (1999). False childhood memories: Research, theory,

and applications. In L. M. Williams & V. L. Banyard (Eds.), Trauma and memory (pp.

175–

188). Thousand Oaks, CA: Sage.

Janitza, S., Strobl, C., & Boulesteix, A.–L. (2013). An AUC- based permutation variable

importance measure for random forests. BMC Bioinformatics, 14, 119.

John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and

theoretical perspectives. Handbook of Personality: Theory and Research, 2, 102-138.

Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological

bulletin, 114(1), 3-28.

157

D.5 References

Johnson, M. K., & Raye, C. L. (2000). Cognitive and brain mechanisms of false memories

and beliefs. In D. L. Schacter & E. Scarry (Eds.), Memory, brain and beliefs (pp. 35-86).

Cambridge, MA, USA: Harvard University Press.

Kaptein, M., De Ruyter, B., Markopoulos, P., & Aarts, E. (2012). Adaptive persuasive

systems: a study of tailored persuasive text messages to reduce snacking. ACM Trans-

actions on Interactive Intelligent Systems (TiiS), 2(2), 10.

Kuhn, M. (2008). Caret package. Journal of statistical software, 28(5), 1-26. Retrieved from

http://www.download.nextag.com/cran/web/packages/caret/caret.pdf

Levitin, D. J. (1994). Absolute memory for musical pitch: Evidence from the production of

learned melodies. Perception & Psychophysics, 56(4), 414-423.

Levitin, D. J., & Cook, P. R. (1996). Memory for musical tempo: Additional evidence that

auditory memory is absolute. Perception & Psychophysics, 58(6), 927-935.

Liebman, J. I., McKinley-Pace, M. J., Leonard, A. M., Sheesley, L. A., Gallant, C. L.,

Renkey, M. E., & Lehman, E. B. (2002). Cognitive and psychosocial correlates of

adults’ eyewitness accuracy and suggestibility. Personality and Individual Differences,

33(1), 49-66.

Loftus, E. F. (2005). Planting misinformation in the human mind: A 30-year investigation of

the

malleability of memory. Learning & Memory, 12(4), 361–366.

Loftus, E. F., & Hoffman, H. G. (1989). Misinformation and memory: The creation of new

memories. Journal of Experimental Psychology: General, 118(1), 100-104.

Loftus, E. F., Miller, D. G., & Burns, H. J. (1978). Semantic integration of verbal information

into a visual memory. Journal of Experimental Psychology: Human Learning and

Memory, 4(1), 19-31.

Loftus, E. F., Schooler, J. W., & Wagenaar, W. A. (1985). The fate of memory: Comment

on McCloskey and Zaragoza. Journal of Experimental Psychology: General, 114(3),

375-380.

McCloskey, M., & Zaragoza, M. (1985). Misleading postevent information and memory for

events: Arguments and evidence against memory impairment hypotheses. Journal of

Experimental Psychology: General, 114(1), 1-16.

Monahan, C. B., Kendall, R. A., & Carterette, E. C. (1987). The effect of melodic and

temporal contour on recognition memory for pitch change. Perception & Psychophysics,

41(6), 576-600.

Müllensiefen, D., Gingras, B., Musil, J., & Stewart, L. (2014). The musicality of non-

musicians: An index for assessing musical sophistication in the general population.

PloS ONE, 9(2), e89642.

158

D.5 References

Müllensiefen, D., & Pendzich, M. (2009). Court decisions on music plagiarism and the

predictive value of similarity algorithms. Musicae Scientiae, 13, 257–295.

Neuschatz, J. S., Lampinen, J. M., Toglia, M. P., Payne, D. G., & Cisneros, E. P. (2017). False

memory research: History, theory, and applied implications. In M. P. Toglia, J. D., Read,

D. F. Ross, & R. C. Lindsay (Eds.), The Handbook of Eyewitness Psychology: Volume

I: Memory for Events (pp. 239-260). Mahwah, NJ: Lawrence Erlbaum Associates.

Paulhus D.L. (1986) Self-Deception and Impression Management in Test Responses. In:

Angleitner A., Wiggins J.S. (eds) Personality Assessment via Questionnaires. Springer,

Berlin, Heidelberg.

Pickrell, J. E., McDonald, D., Bernstein, D. M., & Loftus, E. F. (2016). Misinformation

effect. In R. F. Pohl (Ed.) Cognitive Illusions: Intriguing Phenomena in Judgement,

Thinking and Memory (pp. 406-423). London: Psychology Press.

Porter, S., Birt, A. R., Yuille, J. C., & Lehman, D. R. (2000). Negotiating false memories:

Interviewer and rememberer characteristics relate to memory distortion. Psychological

Science, 11(6), 507-510.

Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words

not

presented in lists. Journal of Experimental Psychology: Learning, Memory, and

Cognition, 21(4), 803-814.

Stuart, E. A., King, G., Imai, K., & Ho, D. E. (2011). MatchIt: nonparametric preprocessing

for parametric causal inference. Journal of Statistical Software, 42(8).

Pallesen, K. J., Brattico, E., Bailey, C. J., Korvenoja, A., Koivisto, J., Gjedde, A., & Carlson,

S. (2010). Cognitive control in auditory working memory is enhanced in musicians.

PloS one, 5(6), e11120.

Peretz, I., Gaudreau, D., & Bonnel, A. M. (1998). Exposure effects on music preference and

recognition. Memory & Cognition, 26(5), 884-902.

Peretz, I., & Zatorre, R. J. (2005). Brain organization for music processing. Annual Review

of Psychology, 56, 89-114.

Poulin-Charronnat, B., Bigand, E., Lalitte, P., Madurell, F., Vieillard, S., & McAdams, S.

(2004). Effects of a change in instrumentation on the recognition of musical materials.

Music Perception: An Interdisciplinary Journal, 22(2), 239-263.

Reyna, V. F., & Brainerd, C. J. (1995). Fuzzy-trace theory: An interim synthesis. Learning

and individual Differences, 7(1), 1-75.

Schellenberg, E. G., & Habashi, P. (2015). Remembering the melody and timbre, forgetting

the key and tempo. Memory & Cognition, 43(7), 1021-1031.

159

D.5 References

Schellenberg, E. G., Stalinski, S. M., & Marks, B. M. (2014). Memory for surface features of

unfamiliar melodies: Independent effects of changes in pitch and tempo. Psychological

research, 78(1), 84-95.

Schiavio, A., & Timmers, R. (2016). Motor and audiovisual learning consolidate auditory

memory of tonally ambiguous melodies. Music Perception: An Interdisciplinary

Journal, 34(1), 21-32.

Schooler, J. W., & Loftus, E. F. (1993). Multiple mechanisms mediate individual differences

in eyewitness accuracy and suggestibility. Mechanisms of everyday cognition, 177-203.

Schulze, K., Mueller, K., & Koelsch, S. (2011). Neural correlates of strategy use during

auditory working memory in musicians and non-musicians. European Journal of

Neuroscience, 33(1), 189-196.

Scoboria, A., Wade, K. A., Lindsay, D. S., Azad, T., Strange, D., Ost, J., & Hyman, I. E.

(2017). A mega-analysis of memory reports from eight peer-reviewed false memory

implantation studies. Memory, 25(2), 146–163.

Shaw, J. (2016). The Memory Illusion: Remembering, Forgetting, and the Science of False

Memory. New York, NY: Random House.

Smelser, N. J., & Baltes, P. B. (Eds.). (2001). International encyclopedia of the social &

behavioral sciences (Vol. 11). Amsterdam: Elsevier.

Stöber, J. (2001). The Social Desirability Scale-17 (SDS-17): Convergent validity, discrimi-

nant validity, and relationship with age. European Journal of Psychological Assessment,

17, 222-232.

Strobl, C., Boulesteix, A.-L., Kneib, T., Augustin, T., & Zeileis, A. (2008). Conditional

variable importance for random forests. BMC Bioinformatics, 9, 307.

Strobl, C., Malley, J., & Tutz, G. (2009). An introduction to recursive partitioning: Rationale,

application, and characteristics of classification and regression trees, bagging, and

random forests. Psychological Methods, 14, 323–348.

Talamini, F., Altoè, G., Carretti, B., & Grassi, M. (2017). Musicians have better memory

than nonmusicians: A meta-analysis. PloS ONE, 12(10), e0186773.

Trainor, L. J., Wu, L., & Tsang, C. D. (2004). Long-term memory for music: Infants

remember tempo and timbre. Developmental Science, 7(3), 289-296.

Vuvan, D. T., Podolak, O. M., & Schmuckler, M. A. (2014). Memory for musical tones: the

impact of tonality and the creation of false memories. Frontiers in Psychology, 5, 582.

Wade, K. A., Garry, M., Read, J. D., & Lindsay, D. S. (2002). A picture is worth a thousand

lies: Using false photographs to create false childhood memories. Psychonomic Bulletin

& Review, 9, 597–603.

160

D.5 References

Ward, R. A., & Loftus, E. F. (1985). Eyewitness performance in different psychological

types. The Journal of General Psychology, 112(2), 191-200.

Weingardt, K. R., Toland, H. K., & Loftus, E. F. (1994). Reports of suggested memories:

Do people truly believe them? In D. F. Ross, J. D. Read, & M. P. Toglia (Eds.), Adult

eyewitness testimony: Current trends and developments (pp. 3-26). Cambridge:

Cambridge University Press.

Weiss, M. W., Vanzella, P., Schellenberg, E. G., & Trehub, S. E. (2015). Pianists exhibit

enhanced memory for vocal melodies but not piano melodies. The Quarterly Journal of

Experimental Psychology, 68(5), 866-877.

Williamson, V. J., Baddeley, A. D., & Hitch, G. J. (2010). Musicians’ and nonmusicians’

short-term memory for verbal and musical sequences: Comparing phonological similar-

ity

and pitch proximity. Memory & Cognition, 38(2), 163-175.

Wylie, L. E., Patihis, L., McCuller, L. L., Davis, D., Brank, E., Loftus, E. F., & Bornstein,

B. (2014). Misinformation effect in older versus younger adults: A meta-analysis and

review. In M. P. Toglia, D. F. Ross, J. Pozzulo, & E. Pica (Eds.), The Elderly Eyewitness

in Court (pp. 38.66). London: Psychology Press.

Zhu, B., Chen, C., Loftus, E. F., Lin, C., He, Q., Chen, C., & Dong, Q. (2010). Individual

differences in false memory from misinformation: Cognitive factors. Memory, 18(5),

543-555.

Ziegler, M., Danay, E., Heene, M., Asendorpf, J., & Bühner, M. (2012). Openness, fluid

intelligence, and crystallized intelligence: Toward an integrative model. Journal of

Research in Personality, 46(2), 173-183.

Appendix E

Names and titles matter: Linguistic

fluency and the affect heuristic (S5)

This is an Accepted Manuscript of an article published by APA in Psychology of Aesthetics,

Creativity, and the Arts. ©American Psychological Association, 2019. This paper is not the

copy of record and may not exactly replicate the authoritative document published in the

APA journal. Please do not copy or cite without author's permission. The final article is

available, upon publication, at: https://doi.org/10.1037/aca0000172. For presentation in this

thesis, the appendices of the paper have been removed and the passages referring to each

Appendix in the text modified to indicate where to find the materials online. Moreover, there

may be minor modifications in the text to guarantee a consistent typographic style

throughout the thesis, such as the position of figures and tables.

Citation

Anglada-Tort, M., Steffens, J., & Müllensiefen, D. (2019). Names and titles matter: The

impact of linguistic fluency and the affect heuristic on aesthetic and value judge- ments

of music. Psychology of Aesthetics, Creativity, and the Arts, 13(3), 277–292. DOI:

https://doi.org/10.1037/aca0000172

Author contribution

I conceived the idea of this project, conducted the experiments, analysed the data, and write

the paper. Prof. Dr. Daniel Müllensiefen (Goldsmiths, University of London) and Prof. Dr.

Jochen Steffens (Technische Universität Berlin) supervised the work at all stages.

Names and Titles Matter: The Impact of Linguistic

Fluency and the Affect Heuristic on Aesthetic and Value

Judgements of Music

It has been shown that titles influence peoples’ evaluation of visual art. However, the question

of whether titles and artist names affect listeners when evaluating music has not yet been

investigated. By using two well-known cognitive heuristics, we investigated whether names

presented with music pieces influenced aesthetic and value judgements of music. Experiment

1 (N = 48) focused on linguistic fluency. The same music excerpts were presented with easy-

to-pronounce (fluent) and difficult-to-pronounce (disfluent) names. Experiment 2 (N = 100)

studied the affect heuristic. The same music excerpts were presented with positive (e.g., Kiss),

negative (e.g., Suicide), and neutral (e.g., Window) titles. In both studies, aesthetic and value

judgements of music were significantly influenced by the linguistic manipulation of the

names. Participants in Experiment 1 evaluated the same music more positively when

presented with fluent names compared to disfluent names. In Experiment 2, presenting the

music with negative titles resulted in the lowest judgements. Moreover, music excerpts

presented with neutral and negative titles were remembered significantly more often than

positive titles. Finally, a comparison of the music presented with and without titles indicated

that music excerpts were more liked in the presence of titles than in their absence. The present

research shows different ways in which aesthetic and value judgements can be influenced

by the names presented with music. Results suggest that like any other human judgement,

evaluations of music also rely on heuristic principles that do not necessarily depend on the

aesthetic stimuli themselves.

Keywords: : music evaluation, artist name, title, fluency, affect heuristric.

163

E.1 Introduction

The idea is straightforward, as argued by Danto (1981). Imagine an art exhibition where four

identical plain red paintings are placed next to each other. The only difference between them

is that they are presented with different titles. One painting is called “The Israelites Crossing

the Red Sea”, another “Kierkegaard’s mood”. There is also a painting titled “Red Square”

and another named “Nirvana”. Visitors to this exhibition would perceive and appreciate

these identical paintings in different ways, influenced by the titles and resulting in different

aesthetic judgements. Danto concluded (1981): “A title is more than a name: frequently it

is a direction for interpretation or reading, which may not always be helpful” (p. 3). The

influence of titles on art appreciation and evaluation has been largely studied in the world of

visual arts, but to the best of our knowledge, there are no studies in the published literature

that examined the extent to which titles presented with music impact aesthetic and value

judgements. Thus, the present study endeavours to make its contribution by investigating the

effects of titles and artist names on the evaluation of music.