Scandinavian Journal of Psychology, 2013, 54, 432–441 DOI: 10.1111/sjop.

12065

Personality and Social Psychology

Musical taste, employment, education, and global region
ADRIAN C. NORTH1 and JANE W. DAVIDSON2
1
School of Psychology and Speech Pathology, Curtin University, Perth, Australia
2
Callaway Centre, University of Western Australia, Crawley, Australia

North, A. C. & Davidson, J. W. (2013). Musical taste, employment, education, and global region. Scandinavian Journal of Psychology 54, 432–441.

Sociologists have argued that musical taste should vary between social groups, but have not considered whether the effect extends beyond taste into
uses of music and also emotional reactions to music. Moreover, previous research has ignored the culture in which participants are located. The present
research employed a large sample from five post-industrial global regions and showed that musical taste differed between regions but not according to
education and employment; and that there were three-way interactions between education, employment, and region in the uses to which participants put
music and also their typical emotional reactions. In addition to providing partial support for existing sociological theory, the findings highlight the
potential of culture as a variable in future quantitative research on taste.
Key words: Music, education, employment, culture, taste.
Adrian C. North, School of Psychology and Speech Pathology, Curtin University, Perth WA 6845, Australia. Tel: +61 (0)8 9266 9558;
e-mail: [email protected]
INTRODUCTION attendance from predominantly professional and managerial,
rather than blue collar, workers; and Pegg’s (1984) survey of
Sociological research has investigated the ‘real world’ status of British concert attendance. Other studies carried out in the 1970s
music as a means of discriminating social groups (see e.g., and early 1980s (e.g., Dixon, 1981; Skipper, 1975) and North
DeNora, 2000; Frith, 1978; 1981). Specifically, research on and Hargreaves (2007) reached similar conclusions. Findings
“taste publics” and “taste cultures” has had some success in cate- and theories such as these are perhaps best considered in the
gorizing fans of particular musical styles in terms of socio- context of Marxist approaches to sociology and social psychol-
demographic variables. These “taste publics,” “express values ogy which propose that “high art” forms of culture are associ-
and standards of taste and aesthetics” (Fox & Wince, 1975, ated with higher socioeconomic status (e.g., Adorno, 1941;
p.199), such that they comprise groups of people who subscribe Bourdieu, 1971, 1984; Franc
es, 1967; Frith, 1990), and indicates
to a particular “taste culture.” The present study aimed to con- that musical taste ought to be predictable from various indicators
tribute to this research area by investigating whether education of a given individual’s social status.
and employment were related to musical taste and reasons for Given the apparent promise of research along these lines it is
listening to music among a large sample drawn from several surprising that so little has been carried out subsequent to the
countries; but also whether the nature of these relationships var- initial studies in the 1970s and 1980s (see Hutchison and
ies between differing countries. Wotring, 1993). Similarly, since the research was carried out
predominantly in North America it is possible that taste publics
do not exist in other cultures within even the western world:
Social status and musical taste increasing globalization in recent years may mean that tastes
Gans (1974) identified five major American taste cultures, which have instead become massified (see e.g., Klein, 2000). Moreover,
were associated with corresponding taste publics. These ranged effects of the nature outlined by research on taste publics might
along a social class continuum, namely “high culture”,“upper- not be limited to solely musical taste. One clear implication of
middle culture”,“lower-middle culture”,“low culture”,and “quasi- the arguments of Bourdieu and colleagues is that people of dif-
folk low culture” .Gans therefore made an explicit link between fering socioeconomic status use music for differing purposes.
different taste cultures and social structure, and in particular While the sociological theorizing outlined here addresses the
stressed the importance of the educational attitudes and values of attempts of one social group to gain superiority over another, it
different socioeconomic groups. Similarly, Fox and Wince seems reasonable to suspect that a by-product of this might be
(1975) collected demographic data on 767 sociology undergradu- that other aspects of music listening also reflect differing goals
ates, identifying five taste cultures that each had membership and priorities. For instance, if more powerful social groups
associated with factors such as age, sex, employment, education, reserve more intellectually-demanding music for themselves then
income, and religious background. Fox and Wince regarded we might also find that their reasons for listening to music are
these results as support for Gans’ concepts of taste cultures and more cerebral than those of less powerful groups who are more
taste publics, since they represented a diverse pattern of musical likely to (have little choice but to) be listening to music for the
tastes in which social class and religious preferences exerted the purpose of entertainment. As such, we might also expect to find
strongest effects. that differing social groups have differing reasons for listening
Similar findings are reported in DiMaggio and Useem’s to music on even a mundane, everyday level. There also exists
(1978) survey of concert attendance in the USA, showing the possibility that these variations in musical taste and purposes

© 2013 The Scandinavian Psychological Associations

Scand J Psychol 54 (2013) Musical taste, class and culture 433

of listening to music might feed through to differing types of to cultures being identified as vertical individualist, horizontal
common emotional reaction to music among people from differ- individualist, vertical collectivist, or horizontal collectivist. As
ing socioeconomic backgrounds: if differing social groups are Shavitt et al. (2008, p.1106) explain, “In vertical individualistic
listening to differing music and listening to it for differing rea- societies (VI; e.g., U.S., Great Britain, France), people tend to be
sons, it seems reasonable to suspect that the emotions aroused concerned with improving their individual status and distinguish-
by that music differ also between the groups in both their ing themselves from others via competition. In contrast, in hori-
strength and quality. In short, the present research predicts that zontal individualist societies (HI; e.g., Sweden, Norway,
social class should correlate with musical taste, such that higher Australia), where people prefer to view themselves as equal to
social class is associated with liking for “high art” musical others in status, the focus [of the individualism] is on expressing
styles, but also that differences between social classes might one’s uniqueness, capability, and self-reliance. In vertical collec-
extend beyond merely musical taste into reasons for listening to tivist societies (VC; e.g., Japan, Korea, India), people focus on
music and perhaps also common emotional reactions. fulfilling obligations to others, and on enhancing the status of
their ingroups in competition with outgroups, even when that
entails sacrificing their own personal goals. In horizontal collec-
Cultural factors tivist societies (HC; exemplified historically by the Israeli
It seems intuitive that any approach concerning social status kibbutz), the focus [of the collectivism] is on sociability, benevo-
must also account for the culture in which a given music listener lence, and interdependence with others in an egalitarian context.”
exists. Music is of course a cultural product in the most general Shavitt et al. (2008) describe three other dimensions that seem
terms, and the existence of (at least some) national differences in to differentiate cultures, namely power distance, uncertainty
musical taste is apparent to any international traveler who avoidance, and masculinity/femininity. Power distance is the
switches on a radio. However, the extent, nature, and possible extent to which people who occupy lower status positions in a
cause of such differences has not been the subject of previous social hierarchy accept the uneven distribution of power.
quantitative research. Moreover, the arguments of Bourdieu and Cultures with high power distance (for which Shavitt et al. offer
others further strengthen the imperative to address the role of Malaysia as an example) contain people of low status who are
culture and social structures in musical taste. more likely to accept autocratic organizational structures;
While there have been numerous attempts to define culture, whereas cultures with low power distance (for which Shavitt
perhaps the most useful in the present context are those dimen- et al. offer Australasia and Scandinavia as examples) contain
sional approaches that have been used in consumer psychology people of low status who are more likely to demand democratic
in a quantitative attempt to relate marketing and consumption to organizational structures and a right to critique those in power.
the specific features of the culture in which an individual is Uncertainty avoidance is the extent to which members of a cul-
located. Cultures obviously differ from one another along an ture are prepared to tolerate uncertainty or ambiguity. Shavitt
almost infinite number of dimensions, although the argument of et al. argue that the finding that Latin American cultures have
research in this context has been that much of the variety can be relatively high uncertainty avoidance explains why these favor
subsumed under a relatively small number of dimensions that rules and structures that provide guidance. Masculinity/feminin-
are able to capture a sufficient portion of the variance in order to ity refers to the extent to which a culture values stereotypically
have practical use. Arguably the most-commonly-used dimension masculine roles (e.g., assertiveness, striving) or stereotypically
along which differing cultures have been contrasted with one feminine roles (e.g., caring, emphasis on interpersonal relation-
another concerns individualism versus collectivism. As Shavitt, ships): Shavitt et al. note that Japan and Scandinavia respec-
Lee & Johnson’s (2008, p.1104) lengthy review notes, “In indi- tively have tended to score highly on each of these masculine
vidualistic cultures, people tend to prefer independent relation- and feminine roles respectively.
ships to others and to subordinate the goals of their ingroups to These differences in cultural values are of course only tenden-
their own personal goals. In collectivist cultures, in contrast, cies across entire populations rather than absolutes that will apply
individuals tend to prefer interdependent relationships with oth- to all individuals within a population all of the time. More gener-
ers and to subordinate their personal goals to those of their ally, it is impossible to make any inferences with confidence con-
ingroups.” In other words, the crucial distinction is in whether cerning how these cultural values might feed through to musical
the individual sees him/herself as unavoidably bound up with taste; and nor is it clear (as will be discussed in more detail later)
others via a range of interdependent social relationships and how country of residence equates with culture in predicting musi-
responsibilities, or as wholly independent of other people and cal taste. Nonetheless, arguments and findings such as these indi-
their social responsibilities and obligations towards them. Exam- cate several possible mechanisms by which people in differing
ples of more individualistic cultures that are typically cited cultures have differing musical tastes and use music for differing
include North America, Scandinavia, the United Kingdom, and purposes. For instance, the cultural dimensions outlined here sug-
the Netherlands. gest that musical taste and the uses of music in (especially verti-
Some research has considered the individualistic versus collec- cal) individualistic cultures may reflect a given person’s attempt
tivist dimension in terms of a second dimension along which cul- to obtain social status and achieve personal goals, such as mood
tures may differ. This second dimension reflects the extent to management (rather than collective, social goals such as securing
which a culture is organized horizontally, in which the equality friendship). Cultures of low power distance might have been less
of people is emphasized, versus vertically, in which an individ- prone to the processes outlined by sociologists in which a ruling
ual’s position in social hierarchies is emphasized. This then leads elite is able to retain high art for themselves such that, in these

© 2013 The Scandinavian Psychological Associations

434 A. C. North and J. W. Davidson Scand J Psychol 54 (2013)

places, higher social class is not related to liking for classical and Canada; 10,159, 35.5%), Scandinavia (Norway, Sweden, Denmark,
music, jazz and other high art musical genres. Uncertainty avoid- Finland, Iceland; 10,760, 37.6%), UK and Ireland (4,164, 14.5%),
France and Germany (2,040, 7.1%), and Australia and New Zealand
ance might promote conservative, mainstream musical tastes in
(1,503, 5.3%). The remaining 7,892 participants could not be classified
which individuals avoid risk by favoring established, well-known into similarly well-represented and coherent cultural groups, and so
music; whereas countries low for uncertainty avoidance might be were excluded from further analyses. This left a final sample of 28,626
likely to enjoy novel music which is innovative and avant garde. participants with a mean age of 28.82 years (SD = 10.68). Age was
Masculine cultures might be attracted toward assertive, raucous sorted into three groups for analysis, namely 24 years and younger
(12,117, 42.3%), 25–34 years (9,572, 33.4%), and 35 + years (6,937,
musical styles which partly serve a ‘display’ function, whereas
24.2%). There were 17,106 males (59.8%) and 11,520 females (40.2%).
feminine cultures might place greater value upon music with Participants stated their total personal annual income before tax by
stronger empathic qualities or which emphasize interpersonal selecting from nine options, namely £0–4,999 (7,734, 27.0%), £5,000–
relations. 7,499 (2,016, 7.0%), £7500–9999 (1,458, 5.1%), £10,000–14,999
Proposals such as these are clearly speculative and represent (2,640, 9.2%), £15,000–19,999 (3,385, 11.8%), £20,000–29,999 (5,202,
18.2%), £30,000–49,999 (4,066, 14.2%), £50,000–79,999 (1,432, 5.0%),
arguably too great an intellectual leap from the arguments con-
£80,000–99,999 (294, 1.0%), and £100,000 + (399, 1.4%). The values
cerning gross cultural differences proposed by Shavitt et al. were also presented in euro, US dollar, and Australian dollar equiva-
(2008) and others. It does seem plausible at this stage of data lents, based on the exchange rate on the day that the questionnaire was
collection, however, that differing cultures should have not only published. Participants’ stated the highest level of education they had
differing musical tastes, but that there should also be cross- completed by selecting from 1 = “Primary school” (485, 1.7%),
2 = “Secondary school (GCSE or equivalent)” (947, 3.3%), 3 = “Sec-
cultural differences in motives for listening to music and the
ondary school (A level or equivalent)” (9,888, 34.5%), 4 = “First
emotional effects of that music. As such, a large number of par- degree (e.g., BA, BSc)” (12,065, 42.1%), or 5 = “Higher degree (e.g.,
ticipants from several cultures were asked to provide data con- Masters, PhD)” (5,241, 18.3%). The questionnaire also stated equivalent
cerning their education and employment and also their cultural levels of education in the systems of those non-UK countries also
background, and to also rate their degree of liking for a large included in the research. Participants’ stated their employment by select-
ing from 10 options namely, 1 = “Higher professional (e.g., company
number of musical styles, their degree of agreement with a num-
directors, police ranking at inspector and above, bank managers, senior
ber of statements concerning their potential uses of music, and civil servants, staff military officers)” (1,600, 5.6%), 2 = “Middle pro-
their typical emotional reaction to the music they listen to typi- fessional (e.g., doctors, barristers, solicitors, clergy, librarians, teachers)”
cally. In the light of the evidence and argument presented here (5,920, 20.7%), 3 = “Lower professional (nurses and midwives, journal-
we might well expect the following; ists, actors, prison officers, other police, and soldiers)” (3,398, 11.9%),
4 = “Office staff or similar sedentary (e.g., clerks, secretaries, driving
Hypothesis 1. Musical taste will differ according to social class instructors, computer operator)” (4,526, 15.8%), 5 = “Small employer
and also culture. (e.g., publicans, farmers, play group leader, window cleaner, painter and
decorator)” (496, 1.7%), 6 = “Lower supervisory or craftsman (e.g.
Hypothesis 2. Uses of music will differ according to social class
printers, plumbers, butchers, bus inspectors, TV engineers, train driv-
and also culture. ers)” (667, 2.3%), 7 = “Semi-routine occupations (e.g., shop assistant,
Hypothesis 3. Common emotional reactions to music will differ traffic warden, cook, bus drivers, hairdressers, postal workers)” (897,
according to social class and culture. 3.1%), 8 = “Routine occupations (e.g., waiters, road sweepers, cleaners,
Hypothesis 4. More speculatively, differences between cultures couriers, building labourers, refuse = collectors)” (579, 2.0%),
9 = “Long-term unemployed (i.e., longer than six months)” (533,
in musical taste and uses of music might reflect those dimen-
1.9%), and 10 = “Student” (10,010, 35.0%). Participants who were
sions along which cultures differ such that (a) the notion of short-term unemployed, retired, or stay-at-home mothers were asked to
individualism versus collectivism suggests that some cultures answer on the basis of the last job they held. Respondents’ IP addresses
may be more prone to using music for individualistic purposes were used to verify that they did not complete the questionnaire twice.
of display and self-enhancement versus communitarian goals
such as enhancing relationships and facilitating communal goals;
(b) the notion of power distance suggests that some countries Questionnaire
will give rise to stronger relationships between social class and The questionnaire was publicized by the host university’s web site and a
musical taste than will others; (c) the notion of uncertainty press release targeted at newspapers and radio stations. The press release
noted that an attempt was being made, “to carry out the largest-ever aca-
avoidance suggests that some countries will have more conserva-
demic study of musical taste,” that the intention was to collect data from
tive tastes and uses of music, whereas others will be more avant a large number of respondents, and that as many people as possible were
grade and adventurous; and that (d) some countries may be in needed to visit the project web site in order to complete an online ques-
some way more masculine or feminine in their musical tastes tionnaire presented there. Data were discarded if participants failed to
and uses of music. state their country of origin, or failed to select the “unknown” option for
a fictional musical style included on the list, pancat, which was included
to identify participants producing unreliable data.
METHOD Participants’ age, sex, and income were used as control variables. Age
and sex were used as controls given the substantial body of research
identifying relationships between these and musical taste. Income was
Participants included as a control variable to allow for differential salary levels in the
Data were collected via an internet questionnaire (in English). Partici- countries from which data was obtained. Note that a previous paper
pants stated their nationality by selecting from a drop-down list. Partici- (North, 2010) also used age, sex, and income data from the present data-
pants were asked to state their nationality, and of the 36,518 set, in that case as control variables when considering the relationship
participants, 78.4% came from five clearly identifiable global regions between personality and musical taste, and these data have the same role
based on economic and cultural factors, namely North America (USA in the present analyses.

© 2013 The Scandinavian Psychological Associations

Scand J Psychol 54 (2013) Musical taste, class and culture 435

In addition to providing basic demographic information, participants between the fourth and fifth factors seemed to represent a ‘tip-
rated their liking for 104 musical styles, presented alphabetically, from 0 ping point’. The loadings on each if the four factors were dis-
(“Dislike a lot”) to 10 (“Like a lot’). Participants were also told that “if
cussed at length by the authors (with further more detailed
you have never heard of a musical style, or you don’t recognize it, then
please select ‘unknown’,” instead of giving a rating. In an attempt to discussions of specific elements involving several colleagues),
investigate the widest possible range of musical styles (given the diver- and the uses of music were subsequently labeled on the basis of
sity of the sample), the 104 styles were from all global regions, selected consensus as 1 = mood management, 2 = pleasing others,
in consultation with an ethnomusicologist to help ensure breadth of cov- 3 = creating an impression on others, and 4 = pleasure.
erage, and checked with the relevant cultural groups (via e.g., cultural
To test hypothesis 1, that musical taste will differ according to
centres, internet chat rooms, and student societies) to ensure they were
well-known. social class and also culture, a 10 (employment) 9 5 (education)
Participants were next presented with: (i) the full list of musical styles 9 5 (region) Multivariate analysis of covariance (MANCOVA)
and asked to nominate one as their “favorite”; (ii) 12 statements from was carried out on factor scores associated with each of the nine
North, Hargreaves, and O’Neill (2000; see Table 2), and asked to rate meta-styles. Income, sex, and age group were included as control
the extent to which each was a reason why they listened to their favour-
variables. Significant results arising from the multivariate and
ite style of music (0 = “Definitely not a reason,” 10 = “Definitely a rea-
son”); and (iii) nine emotional terms from North and Hargreaves (1997; univariate analyses are shown in Tables 3 and 4 respectively.
see Table 5b) on which they rated how, “Your favorite style of music Table 3 shows that, excluding the control variables, only the
usually makes you feel” (0 = “Not at all true for me,” 10 = “Really true main effect of region was statistically significant. Table 4 shows
for me”): in addition to the rating of how ‘arousing’ the music was (see that the only meta-styles to give rise to variations in factor
Berlyne, 1971; Russell, 1978), the remaining eight terms represented two
scores between the regions studied were dance and rock, and the
from each of the four ‘quadrants’ of the circumplex approach to
emotion. means concerning these are shown in Table 5. Note that a docu-
ment detailing the means for each MANCOVA result shown in
the present manuscript is available from the first author’s web
RESULTS page.
Of the 104 listed, 51 musical styles were not recognized by at To test hypothesis 2, that uses of music will differ according
least 60% the participants, and so data concerning these were to social class and also culture, a 10 (employment) 9 5 (educa-
deleted. The styles in question were afrobeat, ballero, bebop, tion) 9 5 (region) MANCOVA was carried out on ratings of the
bhajan, Bhangra, bollywood / filmi, British blues, calypso, four sets of factor scores arising from reasons for listening to
carnatic, contemporary African, corridos, cumbia, dandy, decima, music. Income, sex, and age group were included as control
dixieland, doo-wop, dub, early music, gamelan, garba, ghazals, variables. Significant results arising from the multivariate and
grime, highlife, hindustani, Indian folk, Indian rock, indigenous univariate analyses are shown in Tables 6 and 7 respectively.
Latino, juju, jump blues, khayal, kirtanams, klezmer, lavani, Table 6 shows that the three-way interaction between employ-
lounge, merseybeat, mod, qawwali, Rajasthani, renaissance, skif- ment, education and region was significant, and Table 7 shows
fle, soca, soukous, surf, tarana, thurmi, vallenato, varnas, zorba, that of the four sets of factor scores, those concerning listening
zouk, and zulujive. to music in order to please others gave rise to a significant
A principal components analysis was carried out on ratings three-way interaction. A detailed list of mean ratings is available
of liking for the remaining musical styles. Varimax rotation from the author.
yielded nine components with eigenvalues greater than one To test hypothesis 3, that common emotional reactions to
which accounted for 69.83% of the variance in the data (see music will differ according to social class and culture, a 10
Table 1). Inspection of the scree plot indicated similarly that (employment) 9 5 (education) 9 5 (region) MANCOVA was
selection of nine factors represented a “tipping point” beyond carried out on ratings of the nine adjectival responses that partic-
which further factors were of negligible importance and, as ipants were asked to use in describing their typical reactions to
importantly, difficult to interpret; whereas the nine factors with their favorite music. Income, sex, and age group were included
the largest eigenvalues were very coherent in terms of the as control variables. Significant results arising from the multivar-
musical styles they contained: in addition to the apparent iate and univariate analyses are shown in Tables 8 and Table 9
coherence of those styles contained within factors, only one respectively. Table 8 shows that the three-way interaction
musical style, show tunes, loaded onto more than one of the between employment, education, and region was significant, and
nine factors with the largest eigenvalues. The loadings on each Table 9 shows that ratings of unsettling and boring gave rise to
of the nine factors were discussed at length by the authors p values of exactly 0.05 concerning the three-way interaction,
(with further more detailed discussions of specific elements and that lower p values were found for ratings of disconcerting,
involving several colleagues), and the nine meta-styles were unstimulating, and arousing. A detailed list of mean ratings is
subsequently labeled on the basis of consensus as 1 = jazz, available from the author.
2 = ritualized, 3 = classical, 4 = mainstream mass media,
5 = folk, 6 = rebellious, 7 = MOBO, 8 = dance, and 9 = rock.
A principal components analysis was carried out on ratings of DISCUSSION
the reasons for listening to music. Varimax rotation yielded four The first hypothesis was that musical taste should vary according
components with eigenvalues greater than one which accounted to social class and also culture. Principal components analysis of
for 61.98% of the variance in the data (see Table 2). Inspection ratings of the musical styles identified nine meta-styles. Analysis
of the scree plot indicated similarly that further factors were of the resulting factor scores showed that although there were
subject to steadily declining eigenvalues that that the different clear effects of the control variables (namely, age, sex, and

© 2013 The Scandinavian Psychological Associations

436 A. C. North and J. W. Davidson Scand J Psychol 54 (2013)

Table 1. Liking for musical styles (principal components analysis)

Component

1 2 3 4 5 6 7 8 9

Cool Jazz 0.818

Jazz 0.811
Free Jazz 0.740
Smooth Jazz 0.717
Funk 0.611
Blues 0.582
Rhythm and Blues 0.515
Soul 0.512
Rumba 0.685
Samba 0.666
Polka 0.624
Latino 0.551
Marching/Military 0.548
Bauls 0.540
Show Tunes 0.527 0.603
Religious/Praise 0.521
Classical 0.787
20th Century 0.738
Baroque 0.711
Opera 0.641
Choral 0.608
Soundtracks/Theme songs 0.718
Musicals 0.638
Easy Listening 0.631
Film Scores 0.620
Romantic 0.580
Pop 0.556
American Folk 0.752
Folk Rock 0.698
Country/Western 0.676
Bluegrass 0.653
Folk Music 0.653
Indie 0.795
Punk 0.687
Alternative Rock 0.647
Ska 0.606
Garage 0.602
Rap 0.878
Hip-Hop 0.864
R&B 0.726
Reggae 0.544
Dance/Electronica 0.873
Techno 0.849
House 0.763
Disco 0.593
Rock/Heavy Metal 0.786
Rock and Roll 0.747
Glam Rock 0.505
Percentage of variance 11.11 9.88 9.22 8.12 7.82 6.88 6.45 6.18 4.17
Eigenvalue 6.00 5.34 4.98 4.39 4.22 3.72 3.48 3.34 2.25

Note: Big band, celtic, experimental rock, gospel, psychedelic rock, and swing did not load onto any of the components.

income) that have been documented extensively by previous data suggests that musical tastes are common across differing
research, there was otherwise very little evidence of variation in socioeconomic classes. The absence of any interactions between
factor scores between the meta-styles as a consequence of vary- either of the social class variables and region implies that this
ing education and employment. This runs contrary to the argu- absence of relationship between musical taste and social class
ments of several sociologists, and data collected during the was common across the differing cultures that were included in
1970s and 1980s concerning taste publics, which all suggested the research. More simply, no region was less or more likely
that musical taste ought to be fragmented such that differing than any other to give rise to musical tastes stratified by employ-
social classes have differing musical tastes. Rather, the present ment and education.

© 2013 The Scandinavian Psychological Associations

Scand J Psychol 54 (2013) Musical taste, class and culture 437

Table 2. Reasons for listening to music (principal components analysis) cultures in question, rather than because region operates as some
form of proxy variable for socioeconomic status factors of the
Component
type addressed by Adorno, Bourdieu and others. The main effect
1 2 3 4 of region manifested itself in effects on specifically the dance
and rock meta-styles. Effects of region on dance and rock factor
Difficulties 0.793 scores are interesting as they represent arguably the most rebel-
Tension 0.717 lious, counter-culture meta-styles, and so future research might
Loneliness 0.694
investigate aspects of culture that could relate to this directly.
Boredom 0.683
Emotions 0.516 0.447 For instance, we noted earlier that, in horizontal individualistic
Teachers 0.884 societies, the aim of individualism is to emphasize self-reliance
Parents 0.878 and uniqueness, and it is arguable that both of these themes can
Friends 0.505 0.547 be found prominently in dance and rock music. Indeed, although
Trendy 0.794
we would not wish to make too much of this, it is particularly
Image 0.779
Creative 0.767 interesting that Sweden, Norway, and Australia are cited by
Enjoy 0.659 Shavitt et al. (2008) as scoring highly for horizontal individual-
Percentage of variance 20.22 15.41 14.35 12.00 ism, and Scandinavia and Australia and New Zealand were
Eigenvalue 2.43 1.85 1.72 1.44 shown in Table 5 to have among the highest liking for the dance
and rock meta-styles. Similarly, Shavitt et al. argue that vertical
Notes: ‘Difficulties’ = ‘To help me get through difficult times’,
‘Tension’ = ‘To relieve tension and stress’, ‘Loneliness’ = ‘To reduce collectivist societies (e.g., Japan, Korea, and India) focus on ful-
loneliness’, ‘Boredom’ = ‘To relieve boredom’, Emotions’ = ‘To express filing obligations to others: future research could investigate
my feelings and emotions’, ‘Teachers’ = ‘To please my teachers’, whether dance and rock music might be less popular in such cul-
‘Parents’ = ‘To please my parents’, ‘Friends’ = ‘To please my friends’, tures, and whether individuals in these cultures who do like
‘Trendy’ = ‘To be trendy and fashionable’, ‘Image’ = ‘To create an
dance and rock music are perhaps more likely to be ostracized
image for myself’, ‘Creative = ‘To be creative and use my imagine’,
Enjoy’ = ‘To enjoy the music’. than those who like other musical styles.
While these later points are clearly hypotheses for future
Table 3. Liking for music (multivariate analyses) research rather than conclusions that can be drawn from the
present data, the data in Tables 3–5 appear to also provide some
Effect F DF p Partial eta-squared support for those issues concerning musical taste raised in the
more speculative Hypothesis 4. The absence of an interaction
Income 2.00 9, 1350 0.036 0.013
Age group 10.93 9, 1350 <0.001 0.068 between region and both employment and education means that
Sex 13.63 9, 1350 <0.001 0.083 there is no support for Hypothesis 4b which stated that some
Region 1.69 36, 5412 0.006 0.011 cultures may give rise to stronger relationships than do others
between musical taste and social status. However, there was
more support for Hypotheses 4c and 4d, which argued respec-
There was, however, a main effect of region on factor scores tively that some regions would give rise to more conservative
for the meta-styles. The absence of any interaction between musical tastes than would others and that some would have more
region and either education or employment implies that the masculine or feminine musical tastes than would others. Table 5
effects of region are due to qualitative differences between the shows differences between cultures in dance and rock music

Table 4. Liking for music (univariate analyses)

Effect Dependent Variable DF F p Partial eta-squared

Income Folk 1, 1358 5.78 0.016 0.004

Age group Jazz 1, 1358 36.99 <0.001 0.027
Ritualized 1, 1358 13.35 <0.001 0.010
Folk 1, 1358 30.97 <0.001 0.022
Sex Jazz 1, 1358 6.11 0.014 0.004
Ritualized 1, 1358 21.74 <0.001 0.016
Classical 1, 1358 10.37 0.001 0.008
Mainstream 1, 1358 73.15 <0.001 0.051
MOBO 1, 1358 4.54 0.033 0.003
Employment Mainstream 9, 1358 2.54 0.007 0.017
Folk 9, 1358 2.38 0.011 0.016
Region Dance 4, 1358 2.78 0.026 0.008
Rock 4, 1358 3.00 0.018 0.009
Employment 9 education Ritualized 32, 1358 1.45 0.050 0.033
Folk 32, 1358 1.69 0.010 0.038
Employment 9 education 9 region Ritualized 47, 1358 1.60 0.007 0.052

© 2013 The Scandinavian Psychological Associations

438 A. C. North and J. W. Davidson Scand J Psychol 54 (2013)

Table 5. Selected musical taste mean factor scores (and SD) by region tion between education, employment, and region on ratings of
several uses of music (see Table 6) indicates that social class did
Region Dance Rock
appear to influence the uses to which people put music, but that
North America 0.0923022 0.0145889 these differed from one region to another. Moreover, all the
(0.9797) (0.9860) effects tested, except the main effect of education, were also sta-
Scandinavia 0.0193753 0.1115015 tistically significant. Those effects involving education and
(0.9760) (0.9982) employment support the inference from the existing research on
UK and Ireland 0.1458689 0.1831126
social class and the arts that differing social groups should use
(1.01870) (1.0027)
France and Germany 0.0920010 0.2381376 music for differing reasons. Given that Table 3 showed no effect
(1.1525) (1.0526) of education or employment on factor scores derived from rat-
Australia and NZ 0.1040652 0.0380928 ings of liking for the musical styles, the data in Table 6 show
(0.9709) (0.8345) that sociological arguments concerning social class and taste
might more accurately describe the differing practical uses to
which different social classes use music, rather than different
factor scores, which arguably indicates that some cultures have aesthetic values per se. Moreover, the finding that education and
tastes which are indeed more masculine than do others. Simi- employment interacted with region indicates that the precise nat-
larly, Table 3 shows a multivariate main effect of region on ure of the way in which differing social groups use music
scores for the nine meta-styles which, put simply, indicates that depends upon the nature of the society in question: the specific
some regions gave lower liking ratings than others and therefore relationship between uses of music and social class is apparently
had less eclectic musical tastes. region-specific
The second hypothesis stated that uses of music should differ The data in Table 7 indicate that the three-way interaction
according to social class and also culture. The three-way interac- between education, employment, and region did not occur for

Table 6. Reasons for listening to music (multivariate analyses)

Effect F DF p Partial eta-squared

Age Group 123.14 4, 28353 <0.001 0.017

Sex 316.24 4, 28353 <0.001 0.043
Income 23.81 4, 28353 <0.001 0.003
Employment 2.15 36, 113424 <0.001 0.001
Region 4.88 16, 113424 <0.001 0.001
Employment 9 education 1.44 144, 113424 <0.001 0.002
Employment 9 region 1.46 144, 113424 <0.001 0.002
Education 9 region 1.57 64, 113424 0.002 0.001
Employment 9 education 9 region 1.30 492, 113424 <0.001 0.006

Table 7. Reasons for listening to music (univariate analyses)

Effect Dependent Variable DF F p Partial eta-squared

Age Group Mood management 1, 28356 206.50 <0.001 0.007

Pleasing others 1, 28356 15.93 <0.001 0.001
Creating an impression 1, 28356 260.63 <0.001 0.009
Pleasure 1, 28356 11.46 0.001 <.001
Sex Mood management 1, 28356 1179.13 <0.001 0.040
Pleasing others 1, 28356 26.28 <0.001 0.001
Creating an impression 1, 28356 49.17 <0.001 0.002
Pleasure 1, 28356 4.69 0.030 <.001
Income Mood management 1, 28356 62.05 <0.001 0.002
Creating an impression 1, 28356 4.62 0.032 0.000
Pleasure 1, 28356 23.57 <0.001 <.001
Employment Mood management 9, 28356 3.03 0.001 0.001
Pleasure 9, 28356 2.53 0.007 0.001
Education Creating an impression 4, 28356 3.33 0.010 <.001
Region Mood management 4, 28356 8.79 <0.001 0.001
Pleasing others 4, 28356 4.82 0.001 0.001
Creating an impression 4, 28356 5.00 <0.001 0.001
Employment 9 education Pleasing others 36, 28356 2.33 <0.001 0.003
Employment 9 region Pleasing others 36, 28356 2.42 <0.001 0.003
Education 9 region Pleasing others 16, 28356 2.74 <0.001 0.002
Employment 9 education 9 region Pleasing others 123, 28356 2.10 <0.001 0.009

© 2013 The Scandinavian Psychological Associations

Scand J Psychol 54 (2013) Musical taste, class and culture 439

Table 8. Emotions elicited by music (multivariate analyses)

Effect F DF p Partial eta-squared

Income 3.85 9, 28386 <0.001 0.001

Sex 94.73 9, 28386 <0.001 0.029
Age Group 23.84 9, 28386 <0.001 0.008
Employment 1.28 81, 255546 0.046 <.001
Education 1.89 36, 113556 0.001 0.001
Region 5.09 36, 113556 <0.001 0.002
Employment 9 education 1.18 324, 255546 0.016 0.001
Employment 9 region 1.24 324, 255546 0.002 0.002
Education 9 region 1.53 144, 255546 <0.001 0.001
Employment 9 education 9 region 1.18 1107, 255546 <0.001 0.005

factor scores on all the four uses of music identified. Rather, the predominant emotional responses to music considered in the cur-
three-way interaction was focused on using music in order to rent research.
please others, such that although differing social classes are less
or more likely to use music in order to please others, the spe-
cific nature of the social classes which are more likely to do this Limitations and implications
varies across regions. Moreover, although there were several There are several clear limitations on the generality of the pres-
main effects and interactions concerning the three remaining ent findings. First, the advent of digital technology means that
uses of music (see Table 2), the data in Table 7 perhaps suggest music has become increasingly portable and subsequently con-
that those involving region are most widespread (giving rise to a textualized. It seems likely therefore that reactions to music and
main effect for three of the four sets of factor scores, whereas the uses to which music is put might vary from place to place:
education and employment gave rise to main effects on respec- the data presented in Tables 6 and 7 might overview the uses of
tively only one and two of the sets of factor scores). The music among certain cultural groups, but moment-by-moment
absence of a main effect of region on ratings of listening to uses of music might reveal a good deal more subtlety than the
music for pleasure indicates that music is enjoyed equally across present data imply. Second, the present data conflates nationhood
cultures. However, the main effect of culture on listening to with culture. Future research would benefit from considering eth-
music for mood management, to please others, or to create an nicity, which cuts across legal and geographical boundaries, and
impression with other people appears to support various ele- instead focuses on social bonds which may link people in coun-
ments of Hypothesis 4 and imply that it may be worthwhile car- tries in opposing hemispheres. It would be interesting to deter-
rying out future cross-cultural research on the uses of music. mine whether the social nature of ethnicity means that it exerts a
For instance, we might speculate that factors such as individual- stronger influence on people’s uses of music than do the legal
ism / collectivism (and the importance of display and self- and geographical natures of a country.
enhancement implied by the former versus the importance of This in turn leads to the final and most important limitation of
maintaining relationships implied by the latter) may explain the the present research. It had been hoped originally to recruit par-
main effects of region on using music in order to please others ticipants from a range of cultures although this did not transpire
and to create an impression. Similarly, regional variations in – perhaps as a combination of the use of a press release in only
masculinity/femininity may explain the effects of region on English and the reliance on the internet for data collection – so
using music for mood management or to create an impression that there was a preponderance of participants from western Eur-
with others. ope and North America. Future research that samples a wider
The third hypothesis stated that common emotional reactions range of cultures would be able to test some much more detailed
to music will differ according to employment, education, and hypotheses concerning the relationship between culture and
region. In support of this, the three-way interaction between edu- musical taste (and related issues). As noted earlier, there is a
cation, employment and region gave rise to a significant multi- possibility for instance that people in vertical individualistic cul-
variate interaction (see Table 8), and there were significant tures might be more prone to using music for the purpose of
univariate three-way interactions concerning ratings of discon- gaining social status; people in horizontal individualist cultures
certing, boring, unstimulating, and arousing. With regard to the might use music to promote their uniqueness; people in vertical
other emotional descriptors, ratings of listening to music because collectivist cultures might use music to promote the social domi-
it was relaxing, peaceful, and celebratory were subject to an edu- nance of a particular(ly fashionable) ingroup; and that people in
cation 9 region interaction. Ratings of listening to music horizontal collectivist cultures might use music to promote that
because it was relaxing and peaceful were also subject to an cultures ethical code. Similarly, we would expect an interaction
employment 9 education interaction. Ratings of listening to between the degree of power distance within a culture and its
music because it was exciting were subject to a main effect of tendency to exhibit a relationship between social class and musi-
region. As such, it would appear that relationships involving cal taste, such that low power distance cultures give rise to less
education, employment, and region are reflected in all of the differentiation in musical taste among differing social groups;

© 2013 The Scandinavian Psychological Associations

440 A. C. North and J. W. Davidson Scand J Psychol 54 (2013)

Table 9. Emotions elicited by music (univariate analyses)

Effect Dependent Variable DF F p Partial eta-squared

Income Unsettling 1, 28394 31.18 <0.001 0.001

Disconcerting 1, 28394 15.81 <0.001 0.001
Sex Relaxing 1, 28394 164.97 <0.001 0.006
Peaceful 1, 28394 69.45 <0.001 0.002
Exciting 1, 28394 76.10 <0.001 0.003
Celebratory 1, 28394 55.54 <0.001 0.002
Unsettling 1, 28394 309.90 <00.001 0.011
Disconcerting 1, 28394 300.78 <.001 0.010
Boring 1, 28394 59.77 <0.001 0.002
Unstimulating 1, 28394 6.32 0.012 <.001
Arousal 1, 28394 159.55 <0.001 0.006
Age Group Relaxing 1, 28394 31.32 <0.001 0.001
Peaceful 1, 28394 38.33 <0.001 0.001
Exciting 1, 28394 101.79 <0.001 0.004
Unsettling 1, 28394 14.96 <0.001 0.001
Disconcerting 1, 28394 50.18 <0.001 0.002
Arousal 1, 28394 14.04 <0.001 <.001
Employment Celebratory 9, 28394 2.80 0.003 0.001
Education Relaxing 4, 28394 4.76 0.001 0.001
Peaceful 4, 28394 3.75 0.005 0.001
Unsettling 4, 28394 4.96 0.001 0.001
Disconcerting 4, 28394 4.46 0.001 0.001
Boring 4, 28394 2.62 0.033 <.001
Region Relaxing 4, 28394 2.93 0.020 <.001
Peaceful 4, 28394 7.78 <0.001 0.001
Exciting 4, 28394 5.91 <0.001 0.001
Celebratory 4, 28394 17.05 <0.001 0.002
Unsettling 4, 28394 2.75 0.027 <.001
Disconcerting 4, 28394 7.54 <0.001 0.001
Unstimulating 4, 28394 3.48 0.008 <.001
Employment 9 education Relaxing 36, 28394 1.54 0.020 0.002
Peaceful 36, 28394 1.60 0.012 0.002
Unsettling 36, 28394 1.55 0.019 0.002
Disconcerting 36, 28394 1.44 0.042 0.002
Employment 9 region Unsettling 36, 28394 1.51 0.025 0.002
Disconcerting 36, 28394 1.57 0.017 0.002
Boring 36, 28394 1.82 0.002 0.002
Education 9 region Relaxing 16, 28394 2.24 0.003 0.001
Peaceful 16, 28394 2.06 0.008 0.001
Celebratory 16, 28394 2.20 0.004 0.001
Unsettling 16, 28394 2.40 0.001 0.001
Disconcerting 16, 28394 2.60 <0.001 0.001
Employment 9 education 9 region Unsettling 123, 28394 1.22 0.050 0.005
Disconcerting 123, 28394 1.25 0.033 0.005
Boring 123, 28394 1.22 0.050 0.005
Unstimulating 123, 28394 1.23 0.045 0.005
Arousal 123, 28394 1.24 0.036 0.005

that cultures high in uncertainty avoidance might have conserva- outlined here in countries and cultures other than the post-indus-
tive, mainstream musical tastes; and that masculine cultures trial exemplars studied in the present research.
might be attracted toward assertive, raucous musical styles which In conclusion, this research has reported data from a large sam-
partly serve a ‘display’ function, whereas feminine cultures ple showing that musical taste was not related to education and
might place greater value upon music with stronger empathic employment, but was related to the region in which respondents
qualities or which emphasizes interpersonal relations. Since the were located. Participants’ ratings of the uses of music and ratings
arguably rather similar cultures sampled in the present research of their typical emotional reactions to music were subject to a
nonetheless gave rise to significant effects of region, it would three-way interaction between region, education, and employment.
seem that culture could be an important variable. Moreover, As such, the data provide some support for the theories proposed
given that the cultural dimensions highlighted here have been by notable sociologists, and subsequently supported by data col-
implicated in preferences for a range of consumer goods, we lection, that reactions to music ought to be mediated by social
might expect them to extend to media other than music. Finally, class. However, the interaction between variables indicative of
it would also be very interesting to test notions such as those social class and the region in which participants were located

© 2013 The Scandinavian Psychological Associations

Scand J Psychol 54 (2013) Musical taste, class and culture 441

suggests that culture may be an interesting topic for future quanti- Gans, H. J. (1974). Popular culture and high culture: An analysis and
tative research; and it is possible to derive some detailed hypothe- evaluation of taste. New York: Basic Books.
Hutchison, T. W. & Wotring, C. E. (1993). Musical preferences among
ses that might specify the relationship between culture and
clientele of a non-thematic night club: A study of diversification.
responses to music. Popular Music and Society, 17, 43–62.
Klein, N. (2000). No logo. London: Flamingo.
North, A. C. (2010). Individual differences in musical taste. American
REFERENCES Journal of Psychology, 123, 199–208.
North, A. C. & Hargreaves, D. J. (1997). Liking, arousal potential, and
Adorno, T. W. (1941). On popular music. Zeitschrift f€ur Sozialforschung,
the emotions expressed by music. Scandinavian Journal of Psychol-
9, 17–49.
ogy, 38, 45–53.
Berlyne, D. E. (1971). Aesthetics and psychobiology. New York: Apple-
North, A. C. & Hargreaves, D. J. (2007). Lifestyle correlates of musical
ton-Century Crofts.
preference: 3. Travel, money, education, employment, and health.
Bourdieu, P. (1971). Intellectual field and creative project. In M. F. D.
Psychology of Music, 35, 473–497.
Young (Ed.), Knowledge and control. London: Collier-Macmillan.
North, A. C., Hargreaves, D. J. & O’Neill, S. A. (2000). The importance
Bourdieu, P. (1984). Distinction: a social critique of the judgement of
of music to adolescents. British Journal of Educational Psychology,
taste. London: Routledge.
70, 255–272.
DeNora, T. (2000). Music in everyday life. Cambridge: Cambridge
Pegg, C. (1984). Factors affecting the musical choices of audiences in East
University Press.
Suffolk, England. In R. Middleton & D. Horn (Eds.), Popular music 4:
DiMaggio, P. & Useem, M. (1978). Social class and arts consumption:
performers and audiences. Cambridge: Cambridge University Press.
The origins and consequences of class differences in exposure to the
Russell, J. A. (1978). Evidence of convergent validity on the dimensions
arts in America. Theory and Society, 5, 141–161.
of affect. Journal of Personality and Social Psychology, 36,
Dixon, R. D. (1981). Musical taste cultures and taste publics revisited:
1152–1168.
A research note of new evidence. Popular Music and Society, 8, 2–9.
Shavitt, S., Lee, A. Y. & Johnson, T. P. (2008). Cross-cultural consumer
Fox, W. S. & Wince, M. H. (1975). Musical taste cultures and taste pub-
psychology. In C. Haugvedt, P. M. Herr & F. R. Kardes (Eds.),
lics. Youth and Society, 7, 198–224.
Handbook of consumer psychology (pp. 1103–1131). New York:
Franc
es, R. (1967). Communication persuasive et communication esthe-
Psychology Press.
tique. Journal de Psychologie, 4, 415–430.
Skipper, J. K. (1975). Musical tastes of Canadian and American college
Frith, S. (1978). The sociology of rock. London: Constable.
students: An examination of the massification and Americanization
Frith, S. (1981). Sound effects: Youth, leisure, and the politics of rock
theses. Canadian Journal of Sociology, 1, 49–59.
‘n’ roll. New York: Pantheon.
Frith, S. (1990). What is good music? Canadian University Music Received 7 January 2013, accepted 9 May 2013
Review, 10, 97–98.

© 2013 The Scandinavian Psychological Associations