Document [original]

Article

Visualizing Music Psychology:

A Bibliometric Analysis of Psychology

of Music,Music Perception, and Musicae

Scientiae from 1973 to 2017

Manuel Anglada-Tort

and Katie Rose M. Sanfilippo

Abstract

Music psychology has grown drastically since being established in the middle of the 19th century. However, until now, 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 collaborations), consistent with

the general growth observed in science. The retrieved documents received a total of 33,771 citations (M¼16.17,

SD ¼26.93), 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 discussed. This type of analysis, not without its limitations, can help understand music psychology and

identify future directions within the field.

Keywords

Bibliometrics, music psychology, psychology of music, scientometrics, visualization technique

Introduction

The beginnings of what we now regard as music psychol-

ogy started in the middle of the 19th century as a branch of

both psychology and musicology (Thaut, 2011). 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 per-

ceived, created, responded to, and incorporated into every-

day life (Tan, Pfordresher, & Harr´e, 2017; Thompson,

2009). 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 edu-

cation, music performance, composition and improvisa-

tion, the use of music in everyday life, and music therapy

and wellbeing (Hallam, Cross, & Thaut, 2011). But the

psychology of music can also contribute to broader fields,

such as social psychology, behavioral science, aesthetics,

computer science, medicine and health, consumer psychol-

ogy, 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). Psy-

chology of music conferences have been held in the UK

Technische Universita

¨t Berlin, Berlin, Germany

Goldsmiths College, University of London, London, UK

Corresponding author:

Manuel Anglada-Tort, Department of Audio Communication, Technische

Universita

¨t Berlin, Berlin, Germany.

Email: [email protected]

Music & Science

Volume 2: 1–18

ªThe Author(s) 2019

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/2059204318811786

journals.sagepub.com/home/mns

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).

since 1972 under the umbrella of SEMPRE (formerly

SRPMME). Since then, 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 scien-

tific understanding of all psychological aspects of music

and music education” (journals.sagepub.com/home/pom).

Music Perception, established in 1983, was developed with

a primary focus on cognitive-psychological research with

broader and multidisciplinary draw, including work from

“psychology, psychophysics, neuroscience, 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, the-

oretical and critical articles directed at increasing under-

standing 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 journals, including other APA

journals and journals from related disciplines, such as

musicology, music theory, music therapy, music education,

aesthetics, marketing, and neuroscience. This includes, for

example, the Journal of Research in Music Education,

International Journal of Music Education,Journal of

Music Therapy,Empirical Musicology Review,andPsy-

chomusicology: Music, Mind, and Brain.

The current research focuses on the three most

prominent scientific journals in music psychology, namely

Psychology of Music,Music Perception,andMusicae

Scientiae. We used two criteria to select these journals:

content and impact. Regarding content, the focus was on

journals covering specifically the psychology of music.

Impact was determined by the SJR ranking provided in

SCImago (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 fourth, Musicae

Scientiae sixth, and Music Perception seventh. The first

(IEE Signal Processing Magazine), second (Journal of

Research in Music Education), third (Music Education

Research), and fifth (International Journal of Music Edu-

cation) journals did not meet the first criterion of content,

focusing on other topics rather than on 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 systematically on

what research has been published and what gaps can still be

filled. Bibliometrics and scientometrics allow for the mea-

surement and analysis of published scientific literature,

giving objective and measurable data to help us understand

the discipline’s trajectory thus far. By using computational,

mathematical and statistical techniques, bibliometrics anal-

yses the quantity and quality of published scientific litera-

ture, including citation analysis, authorship and country

productivity and collaborations, impact of publications,

and research trends (e.g., Bla´zquez-Ruiz, Guerrero-Bote,

& Moya-Anego´n, 2016; Blaz

ˇun, Kokol, & Vosˇner, 2015;

Chen, Arsenault, Gingras, & Larivie`re, 2015; De Bellis,

2009; Laengle et al., 2017; Mryglod, Holovatch, Kenna,

& Berche, 2016; Naukkarinen & Bragge, 2016; Sweileh,

2017; Sweileh et al., 2016, 2017; Sweileh, Al-Jabi,

Sawalha, & Zyoud, 2016).

Bibliometric analyses have rarely been applied to music

psychology literature. 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’s

(2011) study, the authors looked at publications within one

journal (Music Perception), covering a total of 578 articles.

The retrieved literature was coded to look at the most fre-

quent topics, populations, stimuli, materials, outcome mea-

sures, and music styles, indicating their trends 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 Percep-

tion. 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 Per-

ception. The authors used strict inclusion criteria to select

articles related to topics relevant to affective aspects of

music, resulting in a total of 286 articles.

A recent paper by Sloboda and Ginsborg (2018) inves-

tigated the country and discipline spread of the contributors

and members of Musicae Scientiae as well as compared the

topics of early meetings and publications with current

meetings. Although a strict bibliometric approach was not

used, the paper is the only publication to specifically exam-

ine the international spread of music psychology research-

ers. While the studies by Triovolas & Levitin (2011) and

Sloboda and Ginsburg (2018) only focused on one scien-

tific journal, the study by Diaz and Silveria (2014) had a

very narrow topical focus (i.e., music and affect research).

Thus, these studies cannot give insight into trends through-

out music psychology as a whole. Moreover, the studies

used very few bibliometric indicators. For instance, they

did not provide information about the growth of publica-

tions, more elaborated citation analysis, or author

2Music & Science

productivity and collaborations. Other important limita-

tions in these studies are the relatively small datasets

and the use of human coders to analyze the content of

the articles.

The present study aims to produce a large-scale compu-

tational 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 identify-

ing 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 bib-

liometrix (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 tech-

niques for generating and visualizing maps based on net-

work data. VOSviewer software has been used in a large

body of published literature (www.vosviewer.com/publica

tions), 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,andMusicae

Scientiae, while 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 num-

ber of citations over time), (3) authorship analysis (i.e.,

productivity, dominance, collaboration index, visualiza-

tions of authorship collaboration, and Lotka’s law coeffi-

cient for scientific productivity), (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.

Methods

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 (Fala-

gas, Pitsouni, Malietzis, & Pappas, 2008) and offers many

relevant features that facilitate bibliometric analysis (e.g.,

author, country, and affiliation contributions, citation anal-

ysis, and the “source type” function).

We searched all available literature, by “source title”, in

Psychology of Music,Music Perception,andMusicae

Scientiae. Using the Scopus “source type” function, we

limited the search to empirical and review articles only,

excluding book chapters, conference papers, and editorial

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, lan-

guage of original document, correspondence address, and

abbreviated source title), abstracts, as well as keywords. All

data was retrieved on April 20, 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 spel-

lings. This might generate inaccuracy and inconsistencies in

the computational analysis of authorship. There is not a gen-

eral solution to this problem. Researchers can decide between

two imperfect approaches: (a) to analyze the data without any

previous processing, which would count the same authors as

different authors when their names were spelled using differ-

ent initials; or (b) to remove the initials of the second (and

third) names from all authors, which would count two differ-

ent authors as the same when they had the same surname and

first initial. We chose the latter approach because it was con-

sidered to have a smaller negative impact on the analysis.

Moreover, we did not find any case in which two different

authors shared the same surname and first initial names in our

dataset. Thus, we removed the second and third initial in all

authors, including only the first surname and first initial.

In addition, it is important to note that our data has a gap

in the literature retrieved from Music Perception between

2002 and 2004, as Scopus does not contain any documents

from this source during these three years. Although Music

Perception published articles during this period of time, we

do not have a clear reason to explain this gap.

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

coefficient for scientific productivity (1926).

Visualization and bibliometric maps were created using

VOSviewer (Van Eck & Waltman, 2010), which uses a

Anglada-Tort and Sanfilippo 3

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 place of the items in the map and their colors

are used to cluster the items. The color of an item is deter-

mined 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. Further-

more, we used the R package rworldmap (South, 2011) to

generate visualizations of the geographical distribution of

countries’ productivity.

Results

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 1

shows the total number and type of articles retrieved per

journal and in total. The majority of documents were

research articles (1,987; 95.12%), whereas review articles

only represented a minimal portion (102; 4.88%). Psy-

chology 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 jour-

nals (20.76, 23.31, and 19.48, respectively). Table 2

shows the top 20 contributions made by author, keywords,

and countries. See Appendix A 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.

Growth in number of publications

The mean number of publications from 1973 to 2017 was

46.42 (SD ¼35.56). The total percentage of relative growth

was 11%. The highest productivity was observed in 2016

Table 2. Top 20 contributions of authors, keywords, and countries.

Author TP Keywords TP Country* TP

Repp, B. H. 27 Music 183 U.S. 484

Kopiez, R. 22 Emotion 87 UK 330

Schubert, E. 22 Performance 45 Canada 130

Huron, D. 21 Music perf. 39 Australia 118

Sloboda, J. A. 21 Rhythm 37 Germany 102

Davidson, J. W. 20 Music training 33 Finland 62

MacDonald, R. A. R 20 Perception 33 France 56

North, A. C. 19 Music therapy 29 Netherlands 47

Clarke, E. F. 17 Creativity 28 Japan 41

Eerola, T. 17 Singing 28 Israel 33

Toiviainen, P. 17 Motivation 25 Sweden 31

Trehub, S. E. 17 Memory 24 Belgium 29

Hargreaves, D. J. 16 Music listening 24 Austria 22

Thompson, W. F. 16 Music perception 24 Italy 21

Welch, G. F. 16 Personality 24 Spain 15

Krumhansl, C. L. 15 Practice 24 Georgia 13

Mu¨llensiefen, D. 14 Preference 24 Poland 9

Williamon, A. 14 Arousal 23 Switzerland 9

Cross, I. 13 Expertise 23 Norway 8

Hallam, S. 13 Communication 22 Greece 7

Note. TP: total publications. *Country of corresponding author.

Table 1. Number and type of articles retrieved.

Type of document PoM (1973–2017) MP (1983–2017) MS (1997–2017) Total %

Research articles 872 710 405 1,987 95.12

Review articles 62 36 4 102 4.88

Total 934 746 409 2,089 100

% 44.71 35.71 19.58 100 –

Note. PoM: Psychology of Music; MP: Music Perception; MS: Musicae Scientiae.

4Music & Science

with a total of 135 publications (6.46%)andthelowest

productivity was observed in 1975 with a total of nine

publications (.43%). Figure 1 shows the total number of

publications in the three journals over time. The total num-

ber of publications increased significantly over time, as

indicated by a simple linear regression, F(1,43) ¼141.1,

p< .001, with an R

of .766.

Table 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 over one year. The AGR is calculated using

the following equation: AGR ¼[(TP ending value - TP

beginning value)/TP beginning value] *100, where TP is

total number of publications. The RGR indicates the

growth rate relative to the total number of publications per

year. The RGR was calculated based on the following

equation: RGR¼[log

–log

]/(T2 - T1), where

log

is the log of the final number of publications after

a specific period of interval; log

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 shows the annual number of publications,

AGR, and RGR in the three journals separately. In Psy-

chology of Music, the average number of publications

from 1973 to 2017 was 20.76 (SD ¼16.48), with a total

relative growth rate of 9%.InMusic Perception the

mean number of publications from 1983 to 2017 was

23.31 (SD ¼8.31), with a total relative growth rate of

15%.InMusicae Scientiae, the average number of

publications from 1997 to 2017 was 19.48 (SD ¼

10.38), with a total relative growth rate of 18%.

Citation analysis

Table 4 shows the summary of the citation analysis of

all three journals combined. Retrieved documents

received a total of 33,771 citations, a mean of 16.17

(SD ¼26.93) citations per document, and median

(Q1—Q3) of 7 (2—20). While the highest number of

total citations was in 2007, with 1,978 (M¼23.3, SD

¼30.1) citations, the lowest was in 1975, with 25 cita-

tions (M¼2.8, SD ¼2.9). Figure 2 shows the average

total number of citations 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) ¼.21, p¼.65, R

¼.005. However,

the relationship between the average citations and year

followed an inverted-U shape, as indicated by a statis-

tically significant quadratic regression, F(2,42) ¼52.65,

p< .001, R

¼.715.

Appendix D 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 ¼26.12) 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 ¼33.25) citations per document, and median

(Q1–Q3) of 14 (5–29). In Musicae Scientiae,the

Figure 1. Total number of publications per journal over time. Note: PoM: Psychology of Music; MP: Music Perception; MS: Musicae

Scientiae.

Anglada-Tort and Sanfilippo 5

documents received a total of 3,358 citations, a mean of

10.17 (SD ¼15.04) citations per document, and median

(Q1–Q3) of 5 (2–12).

The top 10 cited articles and authors in the retrieved

literature are shown in Table 5a and Table 5b respectively.

The publication that received the highest amount of cita-

tions 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.

Authorship analysis: Productivity, dominance,

collaboration, and Lotka’s law

A total of 2,632 authors were covered in the retrieved lit-

erature, with a mean of 1.26 authors per article and a mean

of .79 articles per author. The mean number of co-authors

per article was 2.08. Table 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

Table 3. Annual number of publications, AGR, and RGR.

Year Frequency (%) AGR Cumulative total Log

W RGR

1973 17 (.81) – 17 2.83 –

1974 12 (.57) –.29 29 3.37 .53

1975 9 (.43) –.25 38 3.64 .27

1976 10 (.48) .11 48 3.87 .23

1977 11 (.53) .10 59 4.08 .21

1978 12 (.57) .09 71 4.26 .19

1979 13 (.62) .08 84 4.43 .17

1980 11 (.53) –.15 95 4.55 .12

1981 14 (.67) .27 109 4.69 .14

1982 11 (.53) –.21 120 4.79 .10

1983 19 (.91) .73 139 4.93 .15

1984 38 (1.82) 1.00 177 5.18 .24

1985 36 (1.72) –.05 213 5.36 .19

1986 25 (1.2) –.31 238 5.47 .11

1987 31 (1.48) .24 269 5.59 .12

1988 31 (1.48) .00 300 5.70 .11

1989 30 (1.44) –.03 330 5.80 .10

1990 35 (1.68) .17 365 5.90 .10

1991 38 (1.82) .09 403 6.00 .10

1992 34 (1.63) –.11 437 6.08 .08

1993 28 (1.34) –.18 465 6.14 .06

1994 40 (1.91) .43 505 6.22 .08

1995 34 (1.63) –.15 539 6.29 .07

1996 26 (1.24) –.24 565 6.34 .05

1997 28 (1.34) .08 593 6.39 .05

1998 43 (2.06) .54 636 6.46 .07

1999 46 (2.2) .07 682 6.53 .07

2000 42 (2.01) –.09 724 6.58 .06

2001 32 (1.53) –.24 756 6.63 .04

2002 22 (1.05) –.31 778 6.66 .03

2003 38 (1.82) .73 816 6.70 .05

2004 32 (1.53) –.16 848 6.74 .04

2005 57 (2.73) .78 905 6.81 .07

2006 59 (2.82) .04 964 6.87 .06

2007 85 (4.07) .44 1049 6.96 .08

2008 72 (3.45) –.15 1121 7.02 .07

2009 88 (4.21) .22 1209 7.10 .08

2010 96 (4.6) .09 1305 7.17 .08

2011 90 (4.31) –.06 1395 7.24 .07

2012 93 (4.45) .03 1488 7.31 .06

2013 97 (4.64) .04 1585 7.37 .06

2014 113 (5.41) .16 1698 7.44 .07

2015 133 (6.37) .18 1831 7.51 .08

2016 135 (6.46) .02 1966 7.58 .07

2017 123 (5.89) –.09 2089 7.64 .06

Note. AGR: annual growth rate and RGR: relative growth rate.

Table 4. Summary of the citation analysis.

Year Frequency (%) TC Mean (SD) Median (Q1–Q3)

1973 17 (.81) 86 5.06 (9.17) 1 (0–5)

1974 12 (.57) 118 9.83 (15.44) 2.5 (0.75–10.5)

1975 9 (.43) 25 2.78 (2.86) 2 (0–6)

1976 10 (.48) 60 6 (9.55) 2.5 (0.25–5.75)

1977 11 (.53) 33 3 (2.83) 2 (1–5)

1978 12 (.57) 75 6.25 (8.11) 1.5 (0.75–11.75)

1979 13 (.62) 100 7.69 (12.19) 1 (0–8)

1980 11 (.53) 96 8.73 (10.17) 4 (0–14)

1981 14 (.67) 178 12.71 (13.77) 9 (2–19)

1982 11 (.53) 101 9.18 (10.01) 4 (1.5–15.5)

1983 19 (.91) 272 14.32 (18.3) 9 (1.5–16.5)

1984 38 (1.82) 901 23.71 (27.95) 13 (2.25–33.25)

1985 36 (1.72) 1165 32.36 (65.77) 10.5 (2.75–26.5)

1986 25 (1.2) 507 20.28 (18.12) 18 (8–30)

1987 31 (1.48) 718 23.16 (40.68) 8 (4–19.5)

1988 31 (1.48) 663 21.39 (24.21) 13 (5–29.5)

1989 30 (1.44) 922 22.49 (28.7) 13 (1–31)

1990 35 (1.68) 1045 29.86 (28.41) 16 (11–38.5)

1991 38 (1.82) 1293 34.03 (57.73) 16.5 (9.25–31)

1992 34 (1.63) 628 18.47 (17.58) 16 (4.25–26.75)

1993 28 (1.34) 778 27.79 (38.07) 16 (6–37)

1994 40 (1.91) 875 21.88 (36.17) 14 (6.5–28.25)

1995 34 (1.63) 816 24 (43.25) 14 (3–25.75)

1996 26 (1.24) 942 36.23 (43.43) 21.5 (9.5–50.75)

1997 28 (1.34) 452 16.14 (18.31) 8.5 (3–23.5)

1998 43 (2.06) 808 18.79 (25.39) 11 (2–24.5)

1999 46 (2.2) 1047 22.76 (42.45) 10 (4–22.25)

2000 42 (2.01) 886 21.1 (21.75) 13.5 (7–24.75)

2001 32 (1.53) 996 31.13 (32.66) 24.5 (12.25–38.75)

2002 22 (1.05) 430 19.55 (16.69) 17 (7.25–24.75)

2003 38 (1.82) 812 21.37 (27.93) 7.5 (3–33.25)

2004 32 (1.53) 492 15.38 (30.82) 4 (0.75–14)

2005 57 (2.73) 1308 22.95 (23.87) 17 (7–28)

2006 59 (2.82) 1586 27.34 (29.31) 19 (7.25–27)

2007 85 (4.07) 1978 23.27 (30.1) 12 (5–25)

2008 72 (3.45) 1587 22.04 (21.81) 14 (7.75–27.5)

2009 88 (4.21) 1594 18.11 (22.85) 10.5 (4–23)

2010 96 (4.6) 1443 15.03 (19.11) 8 (3–20)

2011 90 (4.31) 1524 16.93 (21.48) 12 (4.25–17.75)

2012 93 (4.45) 1068 11.48 (10.75) 9 (4–17)

2013 97 (4.64) 888 9.15 (10.45) 6 (3–11)

2014 113 (5.41) 526 4.65 (4.19) 4 (2–4)

2015 133 (6.37) 384 2.89 (3.48) 2 (2–4)

2016 135 (6.46) 173 1.28 (1.62) 1 (0–2)

2017 123 (5.89) 44 .37 (.81) 0 (0–1)

Note. TC: total citations.

6Music & Science

Figure 2. Average total citations per year over time.

Table 5. (a) Top 10 cited articles in the retrieved literature. (b) Top 10 cited authors in the retrieved literature.

Article title Authors Journal (year) TC TC per year

(a) Top 10 cited articles in the retrieved literature

Perception of temporal patterns Povel, D. J., Essens, P. MP (1985) 364 11.03

Music structure and emotional response: Some empirical findings Sloboda, J. A. PoM (1991) 333 12.33

A cross-cultural investigation of the perception of emotion in

music: Psychophysical and cultural cues

Balkwill, L. L., Thompson, W. F. MP (1999) 277 14.58

The emotional sources of “chills” induced by music Panksepp, J. MP (1995) 257 11.17

A perceptual model of pulse salience and metrical accent in

musical rhythms

Parncutt, R. MP (1994) 236 9.83

Emotional expression in music performance: Between the

performer’s intention and the listener’s experience

Gabrielsson, A., Juslin, P. N. PoM (1996) 204 9.27

The role of music in adolescents’ mood regulation Saarikallio, S., Erkkila

¨, J. PoM (2007) 203 18.45

Visual perception of performance manner in the movements of

solo musicians

Davidson, J. W. PoM (1993) 197 7.88

Music cognition and perceptual facilitation: A connectionist

framework

Bharucha, J. J. MP (1987) 197 6.35

A model of expressive timing in tonal music Todd, N. MP (1985) 193 5.85

(b) Top 10 cited authors in the retrieved literature

Authors TC TP h-index

Sloboda, J. A. 1,070 21 14

Thompson, W. F. 736 16 14

Krumhansl, C. L. 725 15 13

Davidson, J. W. 699 20 14

Eerola, T. 608 17 12

Hargreaves, D. J. 522 16 12

Repp, B. H. 515 27 15

North, A. C. 460 19 11

Schubert, E. 459 22 11

Trehub, S. E. 437 17 10

Note. PoM: Psychology of Music; MP: Music Perception; MS: Musicae Scientiae. TC: total citations; TP: total publications.

Anglada-Tort and Sanfilippo 7

last period of 10 years (2008–2017), F(1,43) ¼221.19,

p< .001, R

¼.837. The collaboration index (CI) for

multi-authored papers (CI ¼number 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) ¼78.91, p<.001, R

¼.653.

Figure 3 shows the number of single-authored and

multi-authored publications over time. While a total of

828 documents (39.67%) were single-authored publica-

tions, a total of 1,262 publications (60.41%) were multi-

authored. Both the number of single-authored papers,

F(1,43) ¼27.92, p< .001, R

¼.394, and multi-authored

papers, F(1,43) ¼123.97, p< .001, R

¼.742, increased

Table 6. Average authors per document, author productivity, and collaboration index.

Year Frequency (%) TA

Average TA

per document

Number of

single- authored

publications (%)

Number of

multi-authored

publications (%)

Average TA in

multi-authored

publications CI

1973 17 (.81) 17 1.00 17 (100) 0 (0) 0 .00

1974 12 (.57) 13 1.08 11 (91.67) 1 (8.33) 2 2.00

1975 9 (.43) 10 1.11 8 (88.89) 1 (11.11) 2 2.00

1976 10 (.48) 11 1.10 9 (90) 1 (10) 2 2.00

1977 11 (.53) 15 1.36 7 (63.64) 4 (36.36) 8 2.00

1978 12 (.57) 19 1.58 6 (50) 6 (50) 13 2.17

1979 13 (.62) 14 1.08 12 (92.31) 1 (7.69) 2 2.00

1980 11 (.53) 12 1.09 10 (90.91) 1 (9.09) 2 2.00

1981 14 (.67) 18 1.29 11 (78.57) 3 (21.43) 7 2.33

1982 11 (.53) 15 1.36 7 (63.64) 4 (36.36) 8 2.00

1983 19 (.91) 24 1.26 14 (73.68) 5 (26.32) 10 2.00

1984 38 (1.82) 52 1.37 27 (71.05) 11 (28.95) 25 2.27

1985 36 (1.72) 49 1.36 26 (72.22) 10 (27.78) 23 2.30

1986 25 (1.2) 37 1.48 17 (68) 8 (32) 20 2.50

1987 31 (1.48) 48 1.55 21 (67.74) 10 (32.26) 27 2.70

1988 31 (1.48) 55 1.77 16 (51.61) 15 (48.39) 39 2.60

1989 30 (1.44) 50 1.67 17 (56.67) 13 (43.33) 33 2.54

1990 35 (1.68) 62 1.77 18 (51.43) 17 (48.57) 44 2.59

1991 38 (1.82) 72 1.89 18 (47.37) 20 (52.63) 54 2.70

1992 34 (1.63) 59 1.74 16 (47.06) 18 (52.94) 43 2.39

1993 28 (1.34) 54 1.93 15 (53.57) 13 (46.43) 39 3.00

1994 40 (1.91) 61 1.53 25 (62.5) 15 (37.5) 36 2.40

1995 34 (1.63) 47 1.38 24 (70.59) 10 (29.41) 23 2.30

1996 26 (1.24) 48 1.85 15 (57.69) 11 (42.31) 33 3.00

1997 28 (1.34) 42 1.50 17 (60.71) 11 (39.29) 25 2.27

1998 43 (2.06) 89 2.07 17 (39.53) 26 (60.47) 72 2.77

1999 46 (2.2) 76 1.65 27 (58.7) 19 (41.3) 49 2.58

2000 42 (2.01) 68 1.62 20 (47.62) 22 (52.38) 48 2.18

2001 32 (1.53) 57 1.78 15 (46.88) 17 (53.13) 42 2.47

2002 22 (1.05) 40 1.82 11 (50) 11 (50) 29 2.64

2003 38 (1.82) 76 2.00 13 (34.21) 25 (65.79) 63 2.52

2004 32 (1.53) 62 1.94 17 (53.13) 15 (46.88) 45 3.00

2005 57 (2.73) 110 1.93 22 (38.6) 35 (61.4) 88 2.51

2006 59 (2.82) 118 2.00 22 (37.29) 37 (62.71) 96 2.59

2007 85 (4.07) 165 1.94 34 (40) 51 (60) 131 2.57

2008 72 (3.45) 191 2.65 15 (20.83) 57 (79.17) 176 3.09

2009 88 (4.21) 216 2.45 24 (27.27) 64 (72.73) 192 3.00

2010 96 (4.6) 198 2.06 40 (41.67) 56 (58.33) 158 2.82

2011 90 (4.31) 208 2.31 22 (24.44) 68 (75.56) 186 2.74

2012 93 (4.45) 253 2.72 18 (19.35) 75 (80.65) 235 3.13

2013 97 (4.64) 242 2.49 23 (23.71) 75 (77.32) 219 2.92

2014 113 (5.41) 275 2.43 25 (22.12) 88 (77.88) 250 2.84

2015 133 (6.37) 331 2.49 31 (23.31) 102 (76.69) 300 2.94

2016 135 (6.46) 357 2.64 24 (17.78) 111 (82.22) 333 3.00

2017 123 (5.89) 319 2.59 24 (19.51) 99 (80.49) 295 2.98

Note. Percentages in brackets. TA: total number of authors and CI: collaboration index.

8Music & Science

significantly over time, although this increase had a larger

magnitude in publications with multiple authors.

Figure 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 five co-authorship publications and a

minimum of 100 total citations are visualized, resulting in a

total of 49 authors.

Figure 3. Number of single-authored and multi-authored publications over time.

Figure 4. Network visualization map of author collaborations. Note: 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.

Anglada-Tort and Sanfilippo 9

Table 7 shows the authors with a minimum dominance

factor of > .1. The dominance factor was proposed by

Kumar and 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.

Figure 5 depicts Lotka’s law coefficient for scientific

productivity (Lotka, 1926), indicating the theoretical dis-

tribution (red) and the estimated distribution based on the

retrieved literature (blue). Lotka’s law describes the fre-

quency 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, p¼.22.

Country analysis: Productivity, collaborations,

and geographical distribution

The number of countries contributing to the retrieved liter-

ature was 49. Table 8 displays the countries with a minimum

production of five publications, including their frequency,

total number of citations, and the number of single-country

publications as well as multiple-country publications. The

U.S. 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. Nevertheless, this

analysis did not take the population of each country into

account. Thus, we ran a second analysis considering the

average population for each country. The average population

from 1973 to 2017 per country was calculated using the

World Bank population data (https://data.worldbank.org).

The total number of publications was divided by the average

population for each country (rounded to the nearest million)

to find the total number of publications per million people.

Once population was accounted for, Finland and Australia

had the highest total publications, with 15 and 8.1 publica-

tions per million, respectively.

Figure 6 shows two geographical distributions of publi-

cations, a version without correcting for country population

(map on the top) and a version correcting for country pop-

ulation (map on the bottom). The maps were created using

the R package rworldmap. The map on the top is color-coded

using six categories (1 ¼0–100, 2 ¼101–200, 3 ¼201–300,

4¼301–400, 5 ¼401–500, and 6 ¼501–600 publications),

whereas the map on the bottom used eight categories (1 ¼0–

2, 2 ¼3–4, 3 ¼5–6, 4 ¼7–8, 5 ¼9–10, 6 ¼11–12, 7 ¼13–

14, and 8 ¼15–16 publications per million). In the two

maps, countries colored in dark blue indicate the highest

number of publications, and light yellow colored countries

the lowest. Countries with no color indicate that there was no

retrieved data from these areas.

Figure 7 depicts a network visualization map of

international collaborations. The relatedness of countries

is determined based on their number of co-authored publi-

cations. 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 visua-

lized, clustering in four groups.

Conceptual language

Figure 8 shows an overlay visualization map of author key-

words occurrences (i.e., keywords listed by the authors on

each publication). Only keywords that occurred a minimum

Table 7. Authors with a minimum dominance factor of > .1.

Authors

Dominance

factor

Multi-authored

publications

First author

publications

Eerola, T. .47 17 8

Thompson, W. F. .37 16 6

North, A. C. .31 19 6

Repp, B. H. .30 27 8

Clarke, E. F. .29 17 5

Mu¨llensiefen, D. .28 14 4

Kopiez, R. .27 22 6

Krumhansl, C. L. .27 15 4

Macdonald, R. A. R. .25 20 5

Hargreaves, D. J. .25 16 4

Welch, G. F. .25 16 4

Sloboda, J. A. .24 21 5

Hallam, S. .23 13 3

Davidson, J. W. .20 20 4

Huron, D. .14 21 3

Williamon, A. .14 14 3

Toiviainen, P. .12 17 2

Figure 5. Lotka’s law coefficient for scientific productivity (the-

oretical and estimated distributions).

10 Music & Science

of 10 times were included, resulting in a total of 75 key-

words. Note, however, that Scopus only provides author

keywords data from 2005 onwards. Thus, the overlay map

only displays keywords from 2005 to 2017. 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 rep-

resents 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.

Discussion

This study aimed to analyze, through visualization and

bibliometric techniques, all published literature from Psy-

chology of Music,Music Perception,andMusicae Scien-

tiae. Using all available literature in Scopus, a total of

2,089 publications constituted the retrieved literature,

covering a time span of 44 years (1973–2017). Overall,

there is a clear increase in music psychology research (i.e.,

number of publications, authors, and collaborations), with

a total growth rate of 11%. The retrieved documents

received a total of 33,771 citations (M¼16.17, SD ¼

26.93), with a median (Q1—Q3)of7(2—20).Atotalof

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. 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. How-

ever, the magnitude of this increase was higher in the

publications with multiple authors. The collaboration

index (CI) for multi-authored papers (i.e., CI ¼number

of authors in multi-authored publications/number of

multi-authored papers) also increased significantly over

time, from 2.00 in 1974 (the first year with a multi-

authored paper) to 2.98 in 2017. Moreover, the retrieved

literature covered a total of 49 countries. The U.S. and the

UK were the most productive countries, defined as having

the highest number of publications (U.S. ¼23%and UK¼

16%) and citations (U.S. ¼26%and UK¼18%), but when

corrected for population Finland and Australia had the

highest total number of publications per million (Finland

¼15 per million and Australia ¼8.1 per million). Finally,

the keywords “music” and “emotion” had the highest

number of co-occurrences as well as connections with

other keywords.

The results of this study present objective and measur-

able patterns and trajectories seen across the development

Table 8. Countries with a minimum productivity of five publications (country of corresponding author).

Country TP (%, N ¼2,089) SCP (%, N ¼TP country) SMP (%, N ¼TP) TC (%, N ¼33,771) Average TC per publication

U.S. 482 (23.07) 447 (92.74) 35 (7.26) 8,669 (25.67) 17.99

UK 330 (15.8) 279 (84.55) 51 (15.45) 5,954 (17.63) 18.04

Canada 130 (6.22) 113 (86.92) 17 (13.08) 2,775 (8.22) 21.35

Australia 118 (5.65) 99 (83.9) 19 (16.1) 1,622 (4.8) 13.75

Germany 102 (4.88) 87 (85.29) 15 (14.71) 1,439 (4.26) 14.11

Finland 62 (2.97) 47 (75.81) 15 (24.19) 1,250 (3.7) 20.16

France 56 (2.68) 46 (82.14) 10 (17.86) 704 (2.08) 12.57

Netherlands 47 (2.25) 33 (70.21) 14 (29.79) 1,103 (3.27) 23.47

Japan 41 (1.96) 36 (87.8) 5 (12.2) 517 (1.53) 12.61

Israel 33 (1.58) 31 (93.94) 2 (6.06) 452 (1.34) 13.7

Sweden 31 (1.48) 29 (93.55) 2 (6.45) 775 (2.29) 25

Belgium 29 (1.39) 25 (86.21) 4 (13.79) 485 (1.44) 16.72

Austria 22 (1.05) 13 (59.09) 9 (40.91) 329 (.97) 14.95

Italy 21 (1.01) 17 (80.95) 4 (19.05) 127 (.38) 6.05

Spain 15 (.72) 13 (86.67) 2 (13.33) 69 (.2) 4.6

Georgia 13 (.62) 13 (100) 0 (0) 73 (.22) 5.62

Poland 9 (.43) 9 (100) 0 (0) 115 (.34) 12.78

Switzerland 9 (.43) 5 (55.56) 4 (44.44) 84 (.25) 9.33

Norway 8 (.38) 7 (87.5) 1 (12.5) 65 (.19) 8.12

Greece 7 (.34) 4 (57.14) 3 (42.86) 57 (.17) 8.14

New Zealand 7 (.34) 6 (85.71) 1 (14.29) 67 (.2) 9.57

Portugal 7 (.34) 4 (57.14) 3 (42.86) 34 (.1) 4.86

Turkey 7 (.34) 5 (71.43) 2 (28.57) 54 (.16) 7.71

Hong Kong 6 (.29) 5 (83.33) 1 (16.67) 92 (.27) 15.33

Ireland 6 (.29) 6 (100) 0 (0) 72 (.21) 12

South Africa 6 (.29) 6 (100) 0 (0) 21 (.06) 3.5

Wales 6 (.29) 6 (100) 0 (0) 11 (.03) 1.83

Estonia 5 (.24) 5 (100) 0 (0) 29 (.09) 5.59

Note. TP: total publications, TC: total citations, SCP: single-country publication, MCP: multiple-country publication.

Anglada-Tort and Sanfilippo 11

of music psychology research included within these three

journals. We hope these spark discussion and questions as

to why these patterns might exist, what gaps they leave, and

how they fit within a wider context. Compared to the above

summary of the results, below we discuss these questions

for each of our main findings.

Figure 6. Geographical distribution of publications without correcting for country population (top) and with the correction (bottom).

Note: 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 7. Network visualization map of international collaborations. Note: The width of the line shows the strength of the collabora-

tion. 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.

12 Music & Science

Comparing the three journals

Psychology of Music was the first journal to begin publish-

ing, 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 number of retrieved articles (44%),

followed by Music Perception (36%), and Musicae Scien-

tiae (20%). However, the average number of publications

per year in the three journals is very similar (20.76, 23.31,

and 19.48, respectively). Figure 1 suggests that one poten-

tial main driver in the general increase in publications over

time was Psychology of Music, which moved to a greater

number of publications per year. However, statistical tests

should be used to examine to what extent this difference is

meaningful. Interestingly, Musicae Scientiae has the high-

est relative growth rate of 18%, whereas 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 docu-

ment (M¼24.38, SD ¼33.25), followed by Psychology of

Music (M¼16.98, SD ¼26.12) and Musicae Scientiae

(M¼10.17, SD ¼15.04). Nevertheless, this pattern

changes if we look at the average citations in the three most

recent years (from 2015 to 2017), as calculated by SCIma-

go’s SJR ranking. In this case, Psychology of Music

remains in first place, but Musicae Scientiae moves for-

ward to the second position and Music Perception to the

last. These results could inspire future research to investi-

gate reasons for such differences. One example could be to

examine how funding, publication costs, access, and

editorial teams might influence or predict productivity and

citation outcomes.

Growth of publications

Our results show that from 1973 to 2017 there was an overall

growth in the number of publications across all three jour-

nals. This may not be surprising as research article publica-

tionshaveseenanoverall3%growth every year across all

disciplines and there is some indication that this growth has

accelerated even more in recent years (Ware & Mabe, 2015).

This growth may also be due to an increase in the number of

researchers overall (Ware & Mabe, 2015) and an increase in

the number of journals publishing music psychology

research. 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 have been 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 & 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 All-Party

Parliamentary Group on Arts, Health and Wellbeing

(APPGAHW) in 2014, which aims to improve awareness

Figure 8. Network visualization map of keyword occurrences. Note: 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.

Anglada-Tort and Sanfilippo 13

of the benefits that the arts can bring to health and well-

being. 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 publica-

tions 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 an academic audience.

Citation analysis

The retrieved documents received a total of 33,771 citations,

with a mean of 16.17 (SD ¼26.93) 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). How-

ever, compared to music research publications, which have

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 identi-

fied a significant inverted-U-shaped relationship between

year of publication and average number of citations, with

its highest peak in 2007, which received 1,978 citations. It is

likely that the decrease in the average total citations

observed in the last decade is due to the following two

factors: an increase in the total number of publications emer-

ging each year and in particular in this last decade; and a

natural gap between the 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 cita-

tion 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 cita-

tions 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 10 most-

cited articles (Table 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 psychol-

ogy within these three journals. The author with the highest

number of citations was John 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.

However, note that these results only cover articles pub-

lished within three music-psychology-specific journals. For

instance, we are not capturing articles published in neu-

roscience 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 April 20, 2018. The content of this database is

frequently updated, therefore, the numbers reported here will

likely change over time. Moreover, there are significant dif-

ferences 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 docu-

ments. 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 publi-

cation 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 jour-

nal of publication. For instance, Patience, Patience, Blais,

and Bertrand (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.

Authorship analysis

As noted, the magnitude of the increase in publications was

higher in the publications with multiple authors and the

collaboration index (CI) for multi-authored papers

increased significantly over time.

This growth in the total number of authors and colla-

boration are not just a significant trend in music psychology

but are observed in general scientific literature. The Econ-

omist (2016) found that in 34 million research papers pub-

lished 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 for this growth. One reason could be

the fact that research is becoming more multi- and inter-

disciplinary in general, which is particularly true in the case

of music psychology. There is also a growing need for

research teams to have different types of expertise and

represent a variety of specialist perspectives. This is evi-

dent within the field of neuroscience, where the need for

inter- and multidisciplinary research has been discussed

(e.g., Quagilo et al., 2017, Waldman, 2013). Another rea-

son may be due to authors wanting to “pad their publication

lists” and the increasing institutional pressure 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 under-

stand how this might affect the impact or rigor of published

scientific research.

14 Music & Science

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 5) includes individu-

als 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. This visualization

helps to track how collaborations across different domains

and areas may be carried or created by certain dominant

individuals within the field.

Finally, when comparing our data set to Lotka’s theore-

tical distribution (Lotka, 1926), we found no significant

differences between the observed and the theoretical dis-

tributions. Although expected, this is a clear 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.

Country analysis

When looking specifically at the international collabora-

tions and distributions of publications, we found that out

of the total 49 countries contributing to the retrieved liter-

ature, the U.S. and the UK were the most productive coun-

tries, defined as having the highest number of publications

(U.S. ¼23%and UK ¼16%) and citations (U.S. ¼26%

and UK ¼18%). However, when country population is

taken into account, it is Finland (15 per million persons)

and Australia (8.1 per million persons) that have the highest

publication productivity. By comparing the two maps side

by side (Figure 6), the difference can be seen in countries

such as the U.S., which, after accounting for population,

seems less productive, and countries like Australia and Fin-

land, become the most productive. The large differences in

these two scenarios (when not considering country popula-

tion and when considering it) bring to light the potential of

misusing and misinterpreting bibliometric indicators.

The collaboration network map shows this predomi-

nance of the UK and the U.S. as well, but also shows how

more countries collaborate with the UK, creating more

international collaborations than with the U.S. This may

have to do with the UK being within the wider EU and thus

fostering more collaboration between countries. This pro-

minence of research coming from the U.S. 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, U.S., China, Japan, Germany,

Italy, Canada and France (Kisjes, 2013). However, in our

study the top eight most productive countries were the U.S.,

UK, Australia, Canada, Germany, Finland, France, and the

Netherlands. The productivity of these countries may be

related to certain funding opportunities, 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 impor-

tant to think about which nation’s voices are being heard

and which are the loudest within music psychology

research. There is a limitation in knowledge if only a few

nations are represented. Working towards creating opportu-

nities in other countries for music psychology research and

providing places for people to train could help disperse the

distribution beyond Europe and the U.S.

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 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; Gabriels-

son & Lindstrom, 2001; Juslin & Laukka, 2003; Va¨stfja¨ll,

2001). 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 in the early

2000s, whereas keywords such as “self-regulation”, “flow”,

and “emotion regulation” appear more 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 (www.vosviewer.com/

publications).

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

Anglada-Tort and Sanfilippo 15

retrieve the literature (e.g., the journals of Psychomusicol-

ogy: Music, Mind, and Brain and Music & Science are not

indexed in Scopus). This is an important limitation for two

reasons. Firstly, these journals are all predominantly English

language journals, creating automatically a bias in the types

of publications included and researchers represented. Sec-

ondly, high-quality research on music psychology is pub-

lished in a wide range of journals from a wide variety of

disciplines, including experimental psychology, social psy-

chology, 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.

Furthermore, it is likely that authors working in psychol-

ogy departments or medical and neuroscience contexts

often prefer to publish in mainstream generic journals

(e.g., Journal of Experimental Psychology,Journal of

Cognitive Neuroscience) rather than specialist music psy-

chology journals. These generic journals have a higher

impact compared to the specialist journals covered in the

current study. Therefore, when researchers aim to dissemi-

nate their work in the most influential and prestigious jour-

nals, they might be encouraged to choose generic over

specialist journals. Future research could broaden the scope

of the present study by conducting a bibliometric analysis

covering music psychology literature in other journals,

including generic journals in general psychology, beha-

vioral sciences, medicine, and neuroscience.

The second main limitation relies on the use of 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 exists to this problem, we

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 initial. We hope that the limitations of the current

study are justified by the benefits of using large-scale com-

putational bibliometric analysis.

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 bibliometric techniques to visualize and understand-

ing 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 U.S. should be pursued, allowing for different types of

questions, methods and potential findings, steering our

field away from WEIRD (Westernized, educated, indus-

trialized, rich, and democratic) populations (Henrich,

Heine, & Norenzayan, 2010). Future studies should be

done to investigate potential predictors of music psychol-

ogyresearchcitations.Understanding how the system

around music psychology research, its funding schemes,

organizations and institutions, and the influence of certain

individuals and countries impact the dissemination and aca-

demic impact of music psychology research could shed

light on how the system is working and potential ways to

improve it. Finally, future research should continue inves-

tigating the wider impact of music psychology research on

the general public and policies. The need for efficient mea-

surements of scientific collaboration and research impact is

becoming more important. Using similar large-scale com-

putational analysis allows for these questions to be more

objectively addressed.

Music psychology is still a relatively young field. Tak-

ing the time to systematically 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 excit-

ing directions. More research, using similar methods,

should be done giving insight into the past, present and

future of music psychology research.

Acknowledgement

The authors would like to thank Professor Lauren Stewart, Peter

Harrison, Tabitha Trahan, Haia Ironside, and Kai Mueller for

constructive criticism of the manuscript.

Author contribution

MAT conceived of the idea and the analysis strategy for the study.

All other aspects of the research were done collaboratively by

MAT and KRMS.

Declaration of conflicting interests

The author declared no potential conflicts of interest with respect

to the research, authorship, and/or publication of this article.

Funding

The author disclosed receipt of the following financial support for

the research, authorship, and/or publication of this article: This work

was supported by a PhD studentship from the “Studienstiftung

des Deutschen Volkes” (Bonn, Germany), awarded to Manuel

Anglada-Tort.

ORCID iD

Manuel Anglada-Tort https://orcid.org/0000-0003-3421-9361

16 Music & Science

Peer review

Graham Welch, University College London, Institute of Education.

John Sloboda, Guildhall School of Music and Drama.

One anonymous reviewer.

Supplemental material

Supplemental material for this article is available online.

References

Aria, M., & Cuccurullo, C. (2017). bibliometrix: An R-tool for

comprehensive science mapping analysis. Journal of Infor-

metrics,11, 959–975.

Bellis, N. De. (2009). Bibliometrics and citation analysis: From

the science citation index to cybermetrics.Lanham,MD:

Scarecrow Press.

Bla´zquez-Ruiz, J., Guerrero-Bote, V. P., & Moya-Anego´n, F.

(2016). New scientometric-based knowledge map of food sci-

ence research (2003 to 2014). Comprehensive Reviews in Food

Science and Food Safety,15, 1040–1055. doi:10.1111/1541-

4337.12223

Blaz

ˇun, H., Kokol, P., & Vosˇner, J. (2015). Research literature

production on nursing competences from 1981 till 2012: A

bibliometric snapshot. Nurse Education Today,35, 673–679.

doi:10.1016/j.nedt.2015.01.002

Chen, S., Arsenault, C., Gingras, Y., & Larivie`re, V. (2015).

Exploring the interdisciplinary evolution of a discipline: The

case of Biochemistry and Molecular Biology. Scientometrics,

102, 1307–1323. doi:10.1007/s11192-014-1457-6

Diaz, F., & Silveira, J. S. (2014). Music and affective phenomena:

A 20-year content and bibliometric analysis of research in

three eminent journals. Journal of Research in Music Educa-

tion,62, 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, 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 Jour-

nal,22, 338–342.

Gabrielsson, A., & Lindstro¨m,E.(2001).Theinfluenceofmusi-

cal structure on emotional expression. In P. N. Juslin & J. A.

Sloboda (Eds.), Music and emotion: Theory and research

(pp. 223–248). New York, NY: Oxford University Press.

Greenberg, D. M. (2017, December 21). Musical preferences and

the brain. Retrieved from https://www.psychologytoday.com/

us/blog/the-power-music/201712/musical-preferences-and-

the-brain

Hallam, S., Cross, I., & Thaut, M. (2011). Oxford handbook of

music psychology (2nd ed.). Oxford, UK: Oxford University

Press.

Hancock, C. B., & Price, H. E. (2016). First citation speed for

articles in Psychology of Music. Psychology of Music,44(6),

1454–1470.

Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest

people in the world? Behavioral and Brain Sciences,33,

61–83.

Juslin, P. N., & Laukka, P. (2003). Communication of emotions in

vocal expression and music performance: Different channels,

same code? Psychological Bulletin,129, 770.

Kisjes, I. (2013, December 10). Report: How do the large

research nations compare? Retrieved from https://www.else

vier.com/connect/report-how-do-the-large-research-nations-

compare

Kumar, S., & Kumar, S. (2008). Collaboration in research pro-

ductivity in oil seed research institutes of India. In: Proceed-

ings of Fourth International Conference on Webometrics,

Informetrics and Scientometrics.

Laengle, S., Merigo´, J. 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,

803–816. doi:10.1016/j.ejor.2017.04.027

Levitin, D. J. (2006). This is your brain on music: The science of a

human obsession. New York, NY: Penguin.

Lotka, A. J. (1926). The frequency distribution of scientific pro-

ductivity. Journal of the Washington academy of sciences,

16(12), 317–323.

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, 2105–2125.

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, 1151–1166. doi:10.1007/s11192-

015-1820-2

Naukkarinen, O., & Bragge, J. (2016). Aesthetics in the age of

digital humanities. Journal of Aesthetics & Culture,8, 30072.

doi:10.3402/jac.v8.30072

Patience, G. S., Patience, C. A., Blais, B., & Bertrand, F. (2017).

Citation analysis of scientific categories. Heliyon,3, e00300.

Povel, D.-J., & Essens, P. (1985). Perception of temporal patterns.

Music Perception,2, 411–440. doi:10.2307/40285311

Quaglio, G., Corbetta, M., Karapiperis, T., Amunts, K., Kor-

oshetz, W., Yamamori, T., & Draghia-Akli, R. (2017). Under-

standing the brain through large, multidisciplinary research

initiatives. The Lancet Neurology,16, 183–184.

Sacks, O. (2007). Musicophilia: Tales of music and the brain.

Toronto, Canada: Vintage Canada.

Sloboda, J. A., & Ginsborg, J. (2018). 25 years of ESCOM:

Achievements and challenges. Musicae Scientiae,22,

147–160.

South, A. (2011). rworldmap: A New R package for Mapping

Global Data. R Journal,3.

Sweileh, W. M. (2017). Global research trends of World Health

Organization’s top eight emerging pathogens. Globalization

and Health,13, 9. doi:10.1186/s12992-017-0233-9

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, 1–12. doi:10.1186/s12911-017-0476-7

Anglada-Tort and Sanfilippo 17

Sweileh, W. M., Al-Jabi, S. W., Sawalha, A. F., & Zyoud, S. H.

(2016). Bibliometric profile of the global scientific research on

autism spectrum disorders. SpringerPlus,5. doi:10.1186/

s40064-016-3165-6

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, 25. doi:10.1186/s40413-016-0116-1

Tan, S. L., Pfordresher, P., & Harr´e, R. (2017). Psychology of

music: From sound to significance. London, UK: Routledge.

Thaut, M. (2011). History and research. In S. Hallam, I. Cross, &

M. Thaut (Eds.), Oxford handbook of music psychology (2nd

ed.). Oxford, UK: Oxford University Press.

The Economist. (2016, November 25). Scientific papers get more

authors. Retrieved from https://www.economist.com/graphic-

detail/2016/11/25/scientific-papers-get-more-authors

Thompson, W. F. (2009). Music, thought & feeling: Understanding

the psychology of music. New York, NY: 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, 23–36. doi:10.1525/mp.

2011.29.1.23

van Eck, N. J., & Waltman, L. (2010). Software survey: VOS-

viewer, a computer program for bibliometric mapping. Scien-

tometrics,84, 523–538. doi:10.1007/s11192-009-0146-3

Va¨stfja¨ll, D. (2001). Emotion induction through music: A review

of the musical mood induction procedure. Musicae Scientiae,

5, 173–211.

Waldman, D. A. (2013). Interdisciplinary research is the key.

Frontiers in Human Neuroscience,7, 562.

Waltman, L., Van Eck, N. J., & Noyons, E. C. (2010). A unified

approach to mapping and clustering of bibliometric networks.

Journal of Informetrics,4, 629–635.

Ware, M., & Mabe, M. (2015). The STM report: An overview of

scientific and scholarly journal publishing. Retrieved from

http://digitalcommons.unl.edu/scholcom/9

Zatorre, R. J., & Salimpoor, V. N. (2013). Why music makes our

brain sing. New York Times,9. Retrieved from https://www.

nzcf.org.nz/wp-content/uploads/Why-music-makes-the-brain-

sing.pdf

18 Music & Science