R.L.

Zweigenhaft:
Journalof IndividualD
© 2008
ifferences A Do
Hogrefe
2008; Re 29(1):45–55
& Vol. Mi
Huber En core
Publishers

A Do Re Mi Encore
A Closer Look at the Personality
Correlates of Music Preferences
Richard L. Zweigenhaft
Guilford College, Greensboro, NC, USA

Abstract. Through the development of the Short Test of Musical Preferences (STOMP) and a larger theory of music preferences,
Rentfrow and Gosling (2003, 2006) have helped guide the way toward understanding the role of music in people’s lives, and the
relationship between music preferences and personality. The four music dimensions they established in their 2003 study provide a
broad-brush look at some of the relationships between music preferences and personality. This study of 83 undergraduates at Guilford
College in Greensboro, NC, used the NEO-PI, rather than the Big Five Inventory, which allowed us to examine the six facets that make
up each of the Big Five traits as well as those traits themselves, and it looked separately at the music genres that make up the four music
dimensions identified by Rentfrow and Gosling (2003). The findings provide general support for Rentfrow and Gosling’s work, but
they also demonstrate that the personality patterns for the specific music genres differ considerably from one another, even those that
fall within the same broad music dimensions. The Openness trait was by far the most robust of the Big Five traits assessed by the
NEO-PI, and preferences for some music genres (e.g., folk, international music, and rap/hip-hop) were far more revealing of personality
than others (e.g., classical, rock, and electronic).

Keywords: music preferences, music and personality, music genres

For many years, psychologists have been interested in the
personalities of highly creative people, including, of course,
musicians (e.g., Barron, 1968, 1969). Researchers have stud-
ied the personalities of music majors (e.g., Bourke & Francis,
2000, Shuter, 2000), those who play certain instruments (e.g.,
Ciuffardi and Noemi, 2000), professional women musicians
(e.g., Stremikis, 2002), and Australian rock and pop musi-
cians (Gillespie & Myors, 2000).
Social scientists also have written about devoted fans of
various music genres and music groups. For example, var-
ious sociologists (e.g., Weinstein, 1991) and psychologists
(e.g., Arnett, 1996) have studied metalheads, fans of heavy
metal music. In a series of studies that drew on question-
naires, interviews, as well as observations of concerts, Ar-
nett (1996) found that metalheads scored higher on sensa-
tion-seeking than a control group, they were more likely
than other Americans of their age to be from divorced and
dysfunctional families, and they were very much alienated
from mainstream culture. Similarly, sociologist Rebecca
Adams and her students followed the Grateful Dead on one
of their tours, gathering data about deadheads through in-
terviews, surveys, and participant observation (Adams &
Sardiello, 2000). McGown and Dulaney (1999), using a
measure of the Grateful Dead experience (GDE) and an
assessment of personality traits, found that those with the
Grateful Dead experience were significantly more likely
than most people to be outgoing, open to new and different
experiences, and agreeable.
More recently, in an effort to expand the focus from spe-
cific groups of especially devoted fans, like deadheads and
metalheads, to a broader look at a wide range of music
preferences, Rentfrow and Gosling (2003) developed the
Short Test of Music Preferences (STOMP). The 2003 ver-
sion of the STOMP asked respondents to rate 14 different
music genres. A factor analysis indicated that these 14 gen-
res clustered into four general dimensions that they called
reflective and complex (including preferences for blues,
jazz, classical, and folk music), Intense and rebellious
(rock, alternative, and heavy metal), upbeat and conven-
tional (country, sound track, religious, and pop music), and
energetic and rhythmic (rap/hip-hop, soul/funk, and elec-
tronica/dance music). In a series of studies, they found that
those with preferences for each of these four dimensions
shared similar personality characteristics. For example,
those who preferred reflective and complex music tended
to score high on Emotional Stability and Openness.
Rentfrow and Gosling (2003) assessed personality using
the Big Five Inventory, a 44-item measure that provides
scores for each of the five broad personality domains gen-
erally referred to as the Big Five (Neuroticism, Extraver-
sion, Openness, Agreeableness, and Conscientiousness).
Unlike the 240-item NEO-PI itself, however, on which the
Big Five Inventory is based, this shorter measure does not
yield data on the 30 facets that make up the Big Five traits.
For example, each participant in Rentfrow and Gosling’s
(2003) studies received a score on Neuroticism (on the Big

© 2008 Hogrefe & Huber Publishers Journal of Individual Differences 2008; Vol. 29(1):45–55
DOI 10.1027/1614-0001.29.1.45
46 R.L. Zweigenhaft: A Do Re Mi Encore
Five Inventory it is called Emotional Stability), but not on
the six facets that constitute Neuroticism – anxiety, angry
hostility, depression, self-consciousness, impulsiveness,
and vulnerability. The studies that have been done of met-
alheads and deadheads do not lead one to assume that either
group is more or less neurotic (or emotionally stable) than
other people, but they do suggest that metalheads would
score higher and deadheads would score lower on angry
hostility, and that both metalheads and deadheads might
score higher on impulsiveness. As Rentfrow and Gosling
concluded, “future research that includes narrower facets
of personality is needed to provide a finer grained picture
of the effects of personality on music preferences” (2003,
p. 1251).
Similarly, Rentfrow and Gosling (2003) demonstrated
the broad personality patterns that correlated with prefer-
ences for each of the four music dimensions, but they did
not include information on personality patterns for those
who prefer the various genres that fall within each dimen-
sion. They reported, for example, the correlations between
the Big Five trait scores and scores on reflective and com-
plex, but they did not report separate correlations between
these traits (or their facets) and preferences for the four
music genres that factored onto reflective and complex
(blues, classical, folk, and jazz). Even though preferences
for classical and folk music both loaded on the same factor
and, thus, intercorrelated, it is possible – indeed, it is likely
– that those who prefer classical music differ in personality
from those who prefer folk music. Moreover, Rentfrow and
Gosling have since added at least nine new music genres
to their initial 14, indicating that the original 14 did not
cover certain genres important to music fans, and under-
scoring the need to consider the specific genres separately.
The primary purpose of this study was to extend the
work of Rentfrow and Gosling (2003) by exploring the re-
lationship between music preferences and personality us-
ing finer gradations for each of the two. By using the longer
and more complex NEO-PI instead of the shorter and sim-
pler Big Five Inventory, we were able to look at scores for
the Big Five traits and for the thirty facets that make up
those traits. In addition, we looked not only at Rentfrow
and Gosling’s (2003) four music dimensions, but also at the
individual music genres.
Methods
Between the spring of 2001 and the spring of 2007, students
taking certain psychology courses at Guilford College, a
small liberal arts college in Greensboro, North Carolina,
were given the option of taking the NEO-PI. About one
fourth of the students in these classes chose to do so. This
entailed taking the test on their own (this generally took
between 45 min and 1 h), scoring it (about 15 to 20 min),
and then going over the meaning of the scores with the
instructor (this usually took between 30 and 45 min). Most
Journal of Individual Differences 2008; Vol. 29(1):45–55
students used their NEO-PI scores as part of the self-anal-
yses they wrote for an Introduction to Personality course.
All agreed to allow the instructor to use the NEO-PI data
for this project on music preferences.
As of the fall of 2004, when this study of music prefer-
ences and personality was designed, 49 students had taken
the NEO-PI; some were still students at the college, but
others had graduated or transferred. Those who were still
Guilford students, and those for whom addresses could be
obtained, were contacted and asked to fill in the STOMP;
21 did so. Between the spring of 2004 and the spring of
2007, another 62 students took both the NEO-PI and the
STOMP (as of the spring of 2004, all students took a re-
vised version of the STOMP that included an additional
seven musical genres: bluegrass, funk, gospel, internation-
al, oldies, opera, and punk; an even more recent version of
the STOMP includes reggae and new age). Thus, this study
included NEO-PI scores and STOMP scores for 83 current
or former students; for 64 students, the expanded 21-genre
version of the STOMP was used.
The students fell into three groups: 51 (or 61.4%) were
traditional-aged undergraduates; 14 (or 16.9%) were part
of the college’s adult education program, which means they
were Greensboro residents and older than 23; and 18 (or
21.7%) were part of an early college program that consists
of gifted high-school students who have been accepted into
an on-campus program that allows them to enroll in college
courses even though they are still in high school. The av-
erage age for the early college students was 16.6; for the
traditional aged students, it was 19.4; and for the adult stu-
dents, it was 36.6. As is generally true of the psychology
majors at the college, most were female (66, or 80%) and
most were white (69, or 83%).
Results
Gender, Race, and Age Differences
After checking to make sure there were no outliers (there
were none), we looked to see if there were meaningful dif-
ferences between the men and women, the white students
and the students of color, and the traditional and nontradi-
tional age college students in our sample.
Gender
As noted, 66 of the students in our sample were women,
and 17 were men. When we compared these two subgroups
in terms of music preferences, using independent sample
t-tests, we found significant differences on four of the 21
music genres. The men and women differed significantly
on only one of the original 14 music genres: The women
gave higher ratings to pop, Mf = 4.89, SD = 1.58, than did
the men, Mm = 3.82, SD = 1.67, t = 2.46, df = 81, p < .01.
© 2008 Hogrefe & Huber Publishers
 R.L. Zweigenhaft: A Do Re Mi Encore 47

Table 1. Mean scores and ANOVAs, music preferences for early college students, traditional aged students, and adult
students
Early college Traditional aged Adult students F df p<
n = 18 n = 51 n = 14
Blues 4.71ab 5.45a 4.57b 7.81 2 .001
b b
Country 3.00 3.71 4.79 3.87 2 .03
Folk 3.44a 4.76a 4.07 4.38 2 .02
Heavy metal 2.89 3.45c 2.00c 3.72 2 .03
c c
Rap/hip-hop 3.75 4.71 2.79 5.55 2 .01
Religious 3.89 2.91c 5.36c 10.27. 2 .001
Soulfunk 4.03b 4.50 5.43b 3.12 2 .05
Funk 3.43a 4.53a 3.86 3.29 2 .04
a c b,c
Punk 4.43 4.40 2.50 6.86 2 .002
Gospel 3.71 3.58c 5.14c 3.83 2 .03
a
difference between early college and traditional aged students, p < .05, using Tukey; bdifference between early college and adult students, p <
.05, using Tukey; cdifference between traditional aged students and adult students, p < .05, using Tukey.

The women also gave higher ratings than the men on three
of the subsequently added music genres: international (Mf
= 4.74, SD = 1.57, Mm = 3.54, SD = 1.76, t = 2.39, df = 62,
p < .02), gospel (Mf = 4.22, SD = 1.78, Mm = 2.92, SD =
2.10, df = 62, t = 2.25, p < .03), and punk (Mf = 4.23, SD
= 1.86, Mm = 3.08, SD = 1.61, t = 2.04, df = 62, p < .05).
Race
students). It appears that these three student subgroups have
notably different tastes in music, and we will need to keep
this in mind as we examine the relationship between music
preferences and personality.
The Four Music Dimensions and the Big Five
Traits

Similarly, when we compared the 64 white students with the
13 students of color, we found that these two groups differed
in their ratings of three of the 21 music genres: The students
of color gave higher ratings to religious music (Msofc = 4.57,
SD = 2.17, Mwh = 3.33, SD = 1.94, t = 2.15, df = 81, p < .04),
to soulfunk (Msofc = 5.32, SD = 1.64, Mwh = 4.40, SD = 1.60,
t = 1.96, df = 81, p < .05)-and to gospel (Msofc = 5.33, SD =
1.30, Mwh = 3.63, SD = 1.89, t = 2.95, df = 62, p < .005).
Age
As noted, in addition to the 51 traditional age students, our
sample also included 14 adult education students (over 23
years of age) and 18 high school students in an early college
program. When we ran one-way analyses of variance com-
paring these three groups, we found significant differences
in the ratings on 10 of the 21 music genres. These findings,
which can be seen in Table 1, indicate that the traditional
age students gave significantly higher ratings than at least
one of the other two groups on blues, folk, heavy metal,
rap/hip-hop, and funk; the adult students gave significantly
higher ratings than at least one of the other two groups on
country, religious, soulfunk, and gospel; the early college
students’ ratings fell between the other two groups on all
genres except for punk, which they rated slightly but not
significantly higher than the traditional age students (both
groups rated punk significantly higher than did the adult
First, we looked to see if the four categories of music pref-
erence that were identified in Rentfrow and Gosling’s
(2003) factor analysis demonstrated the same personality
patterns for this sample, based as it was on the NEO-PI
itself (rather than the much shorter Big Five Inventory).
The results can be seen in Table 2, which includes the cor-
relations Rentfrow and Gosling (2003) obtained in two
studies, both based on data from undergraduates at the Uni-
versity of Texas at Austin (their first sample included 1,704
students taking psychology courses in the fall of 2001, and
their second sample included 1,383 students taking psy-
chology courses in the spring of 2002), and the 83 students
in this current study.
As can be seen, for the first domain of music preference,
reflective and complex, Rentfrow and Gosling (2003) found
only one correlation with a Big Five trait that was statistically
significant in both their samples: Those high in this domain
scored higher on Openness (r = .44 and r = .41). The data are
similar in the current sample (r = .34).
Rentfrow and Gosling (2003) also found significant pos-
itive correlations between scores on the second domain of
music preference, Intense and rebellious, and Openness (r
= .18, r = .15), and so did we in the current study (r = .15).
For the third domain of music preference, upbeat and con-
ventional, Rentfrow and Gosling (2003) found significant
correlations on four of the Big Five traits (Extraversion, r
= .24 and r = .15; Agreeableness, r = .23 and r = .24; Open-
ness, r = –.14 and r = –.08; and Conscientiousness, r = .15

© 2008 Hogrefe & Huber Publishers Journal of Individual Differences 2008; Vol. 29(1):45–55
48 R.L. Zweigenhaft: A Do Re Mi Encore

Table 2. Correlates of the four music preference dimensions with the Big Five personality traits
Reflective & Complex Intense & Rebellious Upbeat & Conventional Energetic & Rhythmic
R&G1 R&G2 Z R&G1 R&G2 Z R&G1 R&G2 Z R& 1 R&G2
N= N= N = 83 N = 1704 N = N = 83 N = N= N = 83 N = N= N = 83
Big Five 1,704 1,383 1,383 1,704 1,383 1,704 1,383
Neuroticism –.08* –.04 –.18 .01 .01 –.11 .07 .04 .09 –.01 .01 .02
Extraversion .01 –.02 .00 0 .08* –.04 .24* .15* .09 .22 .19* .22*
Openness .44* .41* .35* .18* .15* .15 –.14* –.08* –.36*ab .03 .04 .32*ab
Agreeableness .01 .03 .09 –.04 .01 –.03 .23* .24* .13 .08* .09* –.07
Conscientiousness –.02 –.06 –.01 –.04 –.03 –.10 .15* .18* .23 0 –.03 –.12
* p < .05; *p < .05, two-tailed. aZ score comparisons significantly different from R&G1; bZ score comparisons significantly different from
R&G2.

Table 3. Correlations between the four music preferences, the Big Five, and the 30 facets
Reflective & Complex Intense & Rebellious Upbeat & Conventional Energetic & Rhythmic
N = 83 N = 83 N = 83 N = 83
Neuroticism –.18 –.11 .09 .02
Anxiety –.15 –.02 .18 .02
Angry hostility –.28* –.21 .03 –.02
Depression –.11 –.08 .12 –.04
Self-consciousness –.08 .01 .05 –.10
Impulsiveness –.06 –.08 –.13 .13
Vulnerability –.23* –.12 .07 .05
Extraversion .00 –.04 .09 .22
Warmth .05 –.03 .10 .17
Gregariousness –.13 –.04 .07 .19
Assertiveness .20 –.11 .11 .04
Activity –.04 –.18 .11 .00
Excitement seeking –.17 .09 .01 .28*
Positive emotions .08 .10 –.03 .23*
Openness .35* .15 –.36* .32*
Fantasy .23* .02 –.22* .24*
Esthetics .32* .06 –.23* .32*
Feelings .11 .08 –.19 .23*
Actions .26* .10 –.02 .22*
Ideas .28* .09 –.26* .09
Values .15 .25* –.44* .17
Agreeableness .09 –.03 .13 –.07
Trust .22* .03 .01 –.07
Straightforwardness –.02 –.08 .07 –.29*
Altruism .14 .07 .20 .19
Compliance –.11 –.08 .04 –.05
Modesty .09 –.03 .17 –.10
Tendermindedness .12 –.06 .04 .11
Conscientiousness –.01 –.10 .23* –.12
Competence .09 .03 .12 .01
Order –.15 –.16 .17 –.08
Dutifulness .12 –.15 .24* –.08
Achievement striving –.05 –.16 .24* –.15
Self-discipline .01 .01 .19 –.18
Deliberation –.03 .00 .08 –.05
*p < .05, two-tailed.

Journal of Individual Differences 2008; Vol. 29(1):45–55 © 2008 Hogrefe & Huber Publishers
 R.L. Zweigenhaft: A Do Re Mi Encore
and r = .18). In our sample, we found the same patterns for
all four: Agreeableness (r = .13), Openness (r = –.36), Con-
scientiousness (r = .23), and Extraversion (r = .09). Com-
parisons between the correlations in this study and the cor-
relations in their studies, based on conversions to z scores,
revealed that the correlation in this study between this do-
main and Openness was significantly more negative than
the two correlations they obtained (z = 1.86, N = 1,787, p
< .04 and z = 2.09, N = 1,466, p < .02).
Finally, on the fourth dimension of music preferences, en-
ergetic and rhythmic, Rentfrow and Gosling (2003) found sig-
nificant patterns that held for both of their samples on two of
the Big Five traits: Those who scored high on this dimension
scored high on Extraversion (r = .22, r = .19) and Agreeable-
ness (r = .08, r = .09). The students in our study demonstrated
a similar pattern for Extraversion (r = .22), but the opposite
pattern for Agreeableness (r = –.07); comparisons based on
conversions to z scores approached but did not reach conven-
tional levels of significance (z = 1.27, N = 1,787, p < .11, and
z = 1.19, N = 1,466, p < .12). In addition, although Rentfrow
and Gosling (2003) found no meaningful relationship between
preference for energetic and rhythmic music and Openness (r
= .03 and r = .04), for the current sample we found a significant
positive correlation (r = .32); conversions to standard scores
revealed that these correlations differed significantly (z = 2.48,
N = 1,787, p < .01, and z = .09, N = 1,466, p < .02).
The current data, therefore, show considerable but not
total agreement with the patterns reported in Rentfrow and
Gosling’s (2003) study. The correlations found in the cur-
rent study correlate with those of the first Rentfrow and
Gosling sample at r = .78, and with those of the second
Rentfrow and Gosling sample at r = .76. The weakest cor-
relations were on the energetic and rhythmic dimension (r
= .40 and r = .47), as was evidenced by the fact that they
found no significant relationships with Openness in their
two samples but we found a significant positive relation-
ship, and they found a positive relationship with Agree-
ableness, but we found a negative relationship. Similarly,
when we ran reliability analyses for the genres that made
up each of Rentfrow and Gosling’s (2003) four music do-
mains, we found the following Cronbach’s αs: reflective
and complex, .98; intense and rebellious, .94; upbeat and
conventional, .92; and energetic and rhythmic, .64.
When we performed a factor analysis similar to the one
employed by Rentfrow and Gosling (2003) – that is, a prin-
cipal components factor analysis using a varimax rotation
– on the ratings of the 83 students in this study on these 14
music genres, we obtained five, not four factors, but they
were quite similar to those that Rentfrow and Gosling
found. Two of the factors consisted of the very same gen-
res: the first factor included pop, religious, soundtracks,
and country (the factor they called upbeat and convention-
al), and the second factor included heavy metal, rock, and
alternate (what they called intense and rebellious). The
third factor in this factor analysis included jazz and classi-
cal (two of the four genres that loaded on their reflective
and complex factor). The fourth factor included folk, blues,
© 2008 Hogrefe & Huber Publishers
49
and electronic (the first two of these were on their reflective
and complex factor, and the third was on what they called
energetic and rhythmic). Finally, the fifth factor included
rap/hip-hop and soul/funk (both of which fell on their en-
ergetic and rhythmic factor). Given how much smaller our
sample was than theirs, this factor analysis appears gener-
ally to confirm their work, but it also suggests that for dif-
ferent populations (theirs was in Texas, ours in North Car-
olina), and at different times (theirs was done in 2001 and
2002, ours between 2004 and 2007), factor analyses on the
ratings of the 14 music genres might lead to slightly differ-
ent clusters. This, too, demonstrates the need to look sepa-
rately at the different genres rather than relying solely on
the factors that emerged in their research
The Four Music Dimensions and the 30
Facets Within the Big Five Traits
As we have indicated, the NEO-PI includes six facets for
each of the Big Five traits. When we ran reliability analyses
using Cronbach’s αs for the scores on these five sets of
facets, they were as follows: Neuroticism, .83; Extraver-
sion, .80; Openness, .72; Agreeableness, .71; and Consci-
entiousness, .87. Although these facets intercorrelate with
one another, they tap distinctly different dimensions of per-
sonality. For example, a person who scores high on Neu-
roticism because of a high score on angry hostility is likely
to be quite different from someone who scores equally high
on Neuroticism but who has a low score on angry hostility.
Similarly, a person who scores high on Neuroticism who
has a high score on self-consciousness differs from a person
with an equally high Neuroticism score but a low score on
self-consciousness. Table 3 includes the correlations be-
tween scores on the four musical preference dimensions
and the 30 facets assessed by the NEO-PI.
Reflective and Complex
As can be seen in Table 3, in our study two of the six facets,
angry hostility (r = –.28) and vulnerability (r = –.23) cor-
related significantly; high ratings on this dimension of mu-
sic were significantly related to lower scores on angry hos-
tility and vulnerability.
Rentfrow and Gosling (2003) did not find meaningful
relationships between this music dimension and the trait of
Extraversion, nor did we. There were no significant rela-
tionships with any of the six Extraversion facets (though
one, assertiveness, approached significance, r = .20, p <
.07). Rentfrow and Gosling (2003) found strong positive
correlations between this music domain and Openness (r =
.44 and r = .41), as did we (r = .35). Preferences for this
domain correlated positively with all six of the Openness
facets, but only four of the six reached conventional levels
of significance: openness to fantasy (r = .34), openness to
Journal of Individual Differences 2008; Vol. 29(1):45–55
50 R.L. Zweigenhaft: A Do Re Mi Encore
esthetics (r = .40), openness to actions (r = .26), and open-
ness to ideas (r = .28).
The data in the current study suggest a stronger relation-
ship between preference for reflective and complex music
and the fourth of the Big Five traits, Agreeableness, than
was apparent in Rentfrow and Gosling’s (2003) earlier
work. Although they found virtually no correlation be-
tween these two (r = .01 and r = .03), we found a slightly,
though not significantly, stronger correlation of r = .09. In
our study, there was a significant correlation between pref-
erences for this domain and ratings for the trust facet (r =
.22).
Like Rentfrow and Gosling (2003), we found no mean-
ingful patterns between this music preference dimension
and the trait of Conscientiousness, nor did preference for
this music dimension correlate significantly with any of the
six facets that constitute this personality trait.
Intense and Rebellious
Rentfrow and Gosling (2003) found no significant pattern
between scores on preferences for intense and rebellious
music and Neuroticism, nor did we. We did, however, find
one near-significant relationship: those who prefer this
kind of music scored lower on angry hostility (r = –.21, p
< .06). Preference for intense and rebellious music did not
correlate with any of the other traits, and with only one of
the other 24 facets; students who preferred this domain also
scored high on the openness to values facet of the Openness
trait (r = .25).
Upbeat and Conventional
Neither Rentfrow and Gosling (2003) nor we found signif-
icant correlations between ratings on upbeat and conven-
tional music and scores on Neuroticism (the correlations in
their samples were r = .07 and r = .04; in this study it was
r = .09). There was a positive correlation between ratings
on this music dimension and scores on one of the six facets,
but it did not reach statistical significance: Those who pre-
ferred upbeat and conventional music were more likely to
score high on Anxiety (r = .18, p < .11).
Rentfrow and Gosling (2003) found positive correla-
tions between ratings of upbeat and conventional music and
Extraversion, as did we, although the pattern was not sig-
nificant in our data for the trait or for the facets that make
up the trait. As noted, they found a negative relationship
between this music domain and Openness (r = –.14 and r
= –.08), and we found a significantly stronger negative re-
lationship (r = –.36). Our more detailed look at the six fac-
ets revealed negative correlations with all six facets, four
of which were statistically significant: openness to fantasy
(r = –.22), openness to esthetics (r = –.23), openness to
ideas (r = –.26), and openness to values (r = –.44); one of
Journal of Individual Differences 2008; Vol. 29(1):45–55
the other two facets, openness to feelings, approached con-
ventional levels of significance (r = .19, p < .08).
Rentfrow and Gosling (2003) found positive correla-
tions between a preference for this music dimension and
Agreeableness (r = .23 and r = .24), as did we (r = .13).
The correlations for the six Agreeableness facets indicate
that the stronger relationships were for modesty (r = .17)
and altruism (r = .20), although neither reached conven-
tional levels of significance. Similarly, in all three studies
there were positive correlations between preference scores
for this music dimension and Conscientiousness (r = .15, r
= .18, r = .23). The data for the six facets indicate that the
relationship is based largely on dutifulness (r = .24) and
achievement striving (r = .24). .
Energetic and Rhythmic
Rentfrow and Gosling (2003) found that those who pre-
ferred energetic and rhythmic music scored significantly
higher on Extraversion (r = .22 and r = .19) and Agreeable-
ness (r = .08 and r = .09). We, too, found a positive corre-
lation for Extraversion (r = .22), but, as noted above, we
found a negative rather than a positive correlation with
Agreeableness (r = –.07). A closer look at the correlations
for the six facets within Extraversion reveals that two, ex-
citement seeking (r = .28) and positive emotions (r = .23),
were significant. As for the facets that make up Agreeable-
ness, one of the six was significant: Those who preferred
this music domain scored significantly lower on straight-
forwardness (r = –.29).
As noted above, whereas Rentfrow and Gosling (2003)
found no relationship between preference for this music
dimension and Openness, we found a positive relationship
(r = .32) that was statistically significant in our study and
significantly stronger than the correlations found in Rent-
frow and Gosling’s (2003) study. A more careful look at
the correlations for the six facets indicates that four were
significant: openness to fantasy (r = .24), openness to es-
thetics (r = .32), openness to feelings (r = .23), and open-
ness to actions (r = .22).
The Specific 21 Musical Genres
As noted, Rentfrow and Gosling (2003) determined that the
original 14 musical genres fell into the four music dimen-
sions we have examined thus far in this paper; they subse-
quently have identified another nine music genres (Gos-
ling, 2006, personal communication), seven of which we
included in this study. We looked to see if there were dis-
tinctive patterns for each of the music genres. As can be
seen in Tables 4–7, the patterns for the various music pref-
erence genres within the four dimensions are sometimes
similar, but not the same; in some cases the differences are
dramatic. These tables include the correlations with each
of the Big Five traits (whether they were statistically sig-
© 2008 Hogrefe & Huber Publishers
 R.L. Zweigenhaft: A Do Re Mi Encore 51

Table 4. Correlations between reflective and complex gen- Table 5. Correlations between Intense and Rebellious gen-
res, traits and facets res, traits, and facets
Blues Classical Folk Jazz Alternate Heavy metal Rock
N = 83 N = 83 N = 83 N = 83 N = 83 N = 83 N = 83
Neuroticism –.12 –.16 –.08 –.17 Neuroticism .05 –.19 –.08
Angry Hostility –.25* Angry hostility –.23*
Vulnerability –.22* Extraversion –.03 –.08 .03
Extraversion .02 –.05 –.02 .07 Openness .08 .19 .04
Assertiveness .32* Values .26* .24*
Openness .30* .04 .40* .27* Agreeableness –.01 –.09 .04
Fantasy .32* Conscientiousness –.01 –.18 .01
Esthetics .23* .29* .32* Order –.22*
Feelings .20 Achievement striving –.22*
Actions .26* .22* *p < .05, two-tailed.
Ideas .29* .22*
Values .25* .27* tive, r = –.28). The ratings for heavy metal correlated sig-
Agreeableness .07 .10 .03 .07 nificantly, and negatively, with one of the Neuroticism fac-
Trust .23*
ets, angry hostility (r = –.23). Ratings for heavy metal also
correlated negatively with Conscientiousness (r = –.18),
Modesty .23*
and significantly so on two of the Conscientiousness facets,
Conscientiousness .02 .03 –.16 .09 order (r = –.22) and achievement striving (r = –.22). There-
Dutifulness fore, despite their reputation as angry and rebellious, those
*p < .05, two-tailed. who preferred heavy metal music scored low, not high, on
angry hostility. Furthermore, although Arnett (1996) re-
nificant or not) and, for the 30 facets that make up those ported that the metalheads in his studies were prone to sen-
five traits, only those correlations that were significant at sation seeking and risk taking, we found only a moderate
the p < .05 level. and nonsignificant relationship for the excitement seeking
Table 4 includes the data for the four genres in the re- facet of the NEO-PI (r = .12, p < .26). Similarly, although
flective and complex domain (blues, classical, folk, and there was a significant negative correlation between pref-
jazz). As can be seen, there were positive and significant erences for heavy metal music and achievement striving,
correlations between the ratings for blues, folk, and jazz these data do not reveal the extreme alienation that Arnett
and scores on the Openness trait, and on some or most of found in his interviewees.
the facets that make up Openness. This was not the case, Four music preference genres factored onto the third
however, for ratings for classical music, the fourth genre in music dimension, upbeat and conventional: country, pop,
this dimension (r = .04). Ratings for folk music correlated religious, and soundtracks. As can be seen in Table 6, there
negatively with angry hostility (r = –.25), and ratings for was one consistent pattern: High scores on these four gen-
jazz correlated negatively with vulnerability (r = –.22), res correlated negatively with Openness. All four of the
positively with assertiveness (r = .32) and positively with correlations with this trait were negative, and there was a
modesty (r = .23). These patterns reveal little information significant negative correlation between each of these mu-
about the personalities of those who like classical music, sic genres and at least one of the six Openness facets. In
more information about those who like the other three gen- addition, ratings for pop correlated positively with Consci-
res in this domain, and the most information about those entious (r = .22), and ratings for religious music correlated
who like jazz (they are not only open, as are those who positively with both the trait of Conscientious (r = .21, p <
prefer the blues and folk music, but they manage to be both .06) and the dutifulness facet (r = .27). There was a positive
assertive and modest). correlation between ratings for soundtracks and both the
The second music dimension, intense and rebellious, in- Agreeableness trait (r = .26) and the modesty facet (r = .28).
cludes alternative, heavy metal, and rock. As Table 5 re- Finally, the fourth music dimension, energetic and rhyth-
veals, only one personality dimension correlated with pref- mic, included three music genres, which Rentfrow and
erences for alternative music (those who rated it high Gosling (2003) called electronic, rap/hip-hop, and
scored high on openness to values) and none correlated soul/funk. As Table 7 indicates, a strong personality profile
with preferences for rock music. The personality correlates emerged for rap/hip-hop, but not for the other two. Ratings
for heavy metal are more revealing, and are intriguing. Rat- for rap/hip-hop correlated significantly with impulsiveness
ings for heavy metal correlated negatively with Neuroti- (r = .22), Extraversion (r = .28), gregariousness (r = .32),
cism (r = –.19 – only one other correlation, between ratings excitement seeking (r = .42), positive emotions ((r = .28),
for international music and Neuroticism, was more nega- Openness ((r = .38), openness to esthetics (r = .33), open-

© 2008 Hogrefe & Huber Publishers Journal of Individual Differences 2008; Vol. 29(1):45–55
52 R.L. Zweigenhaft: A Do Re Mi Encore

Table 6. Correlations between Upbeat and Conventional Table 7. Correlations between Energetic and Rhythmic
genres, traits, and facets genres, traits, and facets
Country Pop Religious Sound- Electronic Rap/hip-hop Soul/funk
tracks N = 83 N = 83 N = 83
N = 83 N = 83 N = 83 N = 83 Neuroticism –.10 .06 .07
Neuroticism .08 .06 .02 .11 Impulsiveness .22*
Impulsiveness –.23* Extraversion .01 .28* .14
Extraversion .09 .17 .00 –.01 Warmth .22*
Openness –.11 –.31* –.36* –.26* Gregariousness .32*
Esthetics –.28* –.23* Excitement-seeking .42*
Feelings –.27* Positive emotions .28*
Ideas –.30* Openness .04 .38* .22*
Values –.22* –.25* –.54* –.24* Esthetics .33* .22*
Agreeableness .02 .01 .11 .26* Feelings .32*
Modesty .28* Actions .22*
Conscientiousness .07 .22 .21 .17 Values .34*
Dutifulness .27* Agreeableness –.16 –.03 .06
*p < .05, two-tailed.
Straightforwardness –.26*
Altruism .22*
Conscientiousness –.09 –.19 .07
Self-discipline –.24*
*p < .05, two-tailed.

Table 8. Correlations between newly added genres, traits, and facets

Bluegrass International Oldies Opera Punk Gospel Funk
n = 64 n = 64 n = 64 n = 64 n = 64 n = 64 n = 64
Neuroticism –.11 –.28* .05 –.14 –.03 –.02 –.11
Anxiety –.25*
Angry hostility –.28*
Depression –.26*
Vulnerability –.25*
Extraversion –.02 .08 .02 .16 .06 .15 –.01
Assertiveness .26*
Openness .37* .43* .22 .37* .34* .08 .38*
Fantasy .35* .38* .32* .27* .26* .34*
Esthetics .28* .34* .42*
Feelings .32* .25* .25*
Actions .34* .33*
Values .29* .36* .28* .42*
Agreeableness .05 .02 –.02 .05 .16 .26* .01
Compliance –.29*
Conscientiousness –.07 .00 –.18 –.09 –.14 .18 –.38*
Competence –.25*
Order –.25* –.31*
Dutifulness .26* –.32*
Achievement striving –.34*
Self-discipline –.32*
Deliberation –.27*
*p < .05, two-tailed.

Journal of Individual Differences 2008; Vol. 29(1):45–55 © 2008 Hogrefe & Huber Publishers
 R.L. Zweigenhaft: A Do Re Mi Encore
ness to feelings (r = .32), openness to values (r = .34), and
correlated negatively with straightforwardness (r = –.26)
and self-discipline (r = –.24). Ratings for electronic music
did not correlate significantly with any of the five traits or
30 facets, and ratings for soul/funk correlated with one of
the Extraversion facets (warmth, r = .22), one of the Agree-
ableness facets (altruism, r = .22), with Openness (r = .22)
and with two of the openness facets (openness to esthetics,
r = .22, and openness to actions, r = .22).
What about the seven music preference categories Rent-
frow and Gosling added, which we were able to include for
the last 64 of the 83 students who participated in this study?
Table 8 includes the correlations between preference scores
for each of these genres and scores on the Big Five traits
and the 30 facets. Four of these seven factor onto the re-
flective and complex dimension: bluegrass, opera, interna-
tional, and oldies. As was the case for those who preferred
three of the original four genres in this music domain
(blues, folk, and jazz, but not classical), we found that those
who preferred these four music genres scored high on
Openness (and at least one of the openness facets). Most
striking in Table 8 are the data for international music. Rat-
ings for this genre correlated negatively with Neuroticism
(r = –.28), and with three of the Neuroticism facets (angry
hostility, r = –.28; depression, r = –.26; and vulnerability,
r = –.25). They also correlated positively with assertiveness
(r = 26). Also noteworthy in Table 8 is that ratings for opera
correlated negatively with order (r = –.27) and ratings for
oldies correlated negatively with compliance (r = –.29).
One of the added genres, punk, factored onto the second
dimension, intense and rebellious. Ratings for punk corre-
lated positively with Openness (r = .34) and two of the
Openness facets (openness to fantasy, r = .26, and openness
to values, r = .28).
Preferences for gospel music factored onto the upbeat
and conventional dimension. Ratings for gospel music cor-
related positively with the trait of Agreeableness (r = .26)
and the Conscientiousness facet of dutifulness (r = .26).
Finally, ratings for funk, which factored onto the ener-
getic and rhythmic dimension, correlated positively with
Openness (r = .38) and with all of the Openness facets ex-
cept for openness to ideas, and negatively with Conscien-
tiousness (r = –.38) and five of the six Conscientiousness
facets.
When we step back from all of these findings and ask
which students with which music preferences are likely to
be the least and most Neurotic, we see that the most nega-
tive correlations with Neuroticism were with ratings for
international music (r = –28 and (surprisingly) heavy metal
(r = –.19). The most positive correlations with Neuroticism
were with ratings for country (r = .08) and soundtracks (r
= .11). The highest positive correlation for Extraversion
was with ratings for rap/hip-hop (r = .28) and most negative
correlation was with ratings for heavy metal (r = –.08). The
most positive correlation with Openness was with ratings
for international music (r = .43), and the most negative cor-
relations were with ratings for religious music (r = –.36).
© 2008 Hogrefe & Huber Publishers
53
The strongest positive correlations for Agreeableness were
with ratings for gospel (r = .26) and soundtracks (r = .26),
and the strongest negative correlation was with ratings for
electronic music (r = –.16). Finally, the strongest positive
correlation for Conscientiousness was with ratings for pop
(r = .22) and the strongest negative correlation was with
ratings for funk (r = –.38)
Grades
Rentfrow and Gosling (2003) asked participants to rate
themselves on a variety of dimensions, one of which was
“intelligent.” For a subsample of their participants, they
administered the 50-item Wonderlic IQ test to assess verbal
and analytic reasoning ability. They found that those who
preferred reflective and complex music, and those who pre-
ferred intense and rebellious music, saw themselves as
more intelligent, and they scored higher on the verbal sec-
tion of the Wonderlic (but not the analytical section). In
contrast, those who preferred upbeat and conventional mu-
sic saw themselves as less intelligent in one of their two
studies (r = –.05, p < .05) but not in the other study (r =
–.02, ns), and they scored lower on the verbal section of
the Wonderlic (r = –.18) but not the analytical section (r =
.02).
We included no self-assessment of intelligence, nor did
we administer a test of cognitive abilities, but we did look
at the students’ final grades in the course in which they
were enrolled when they participated in this research.
When we looked at the correlations between scores on mu-
sical preferences and their grades (one student withdrew
from the class and received no grade), we found no corre-
lations that reached statistical significance, but some sug-
gestive patterns: ratings for rock correlated with high
grades (r = .19, n = 82, p < .09) as did ratings for punk (r
= .19, n = 63, p < .13), and ratings for country correlated
negatively with grades (r = –.15, n = 82, p < .17), as did
ratings for rap/hip-hop (r = –.14, n = 82, p < .20).
Discussion and Conclusion
These findings support Rentfrow and Gosling’s (2003)
general contention that music preferences are related to
personality patterns. They suggest, however, that a more
nuanced look at both the facets that make up the Big Five
traits, and the music genres that make up the four music
preference dimensions identified by Rentfrow and Gosling
(2003) enhance understanding of those personality pat-
terns.
Both their findings and these more current findings in-
dicate that some music domains, and some music genres,
are much more likely to reveal personality than others. Ta-
ble 4 shows, for example, that how the respondents felt
about folk music, one of the reflective and complex genres,
Journal of Individual Differences 2008; Vol. 29(1):45–55
54 R.L. Zweigenhaft: A Do Re Mi Encore
correlated significantly with nine traits and facets, but how
they felt about classical music, also in the reflexive and
complex domain, did not correlate with any of the Big Five
traits or 30 facets. Similarly, as can be seen in Table 7, how
they felt about rap/hip-hop, which falls in the energetic and
rhythmic domain, correlated with 11 traits or facets, but
how they felt about electronic music, which is in the same
domain, did not correlate with any.
Along the same lines, some traits and facets are more
likely to be revealed, especially Openness and the six facets
subsumed by it. More than 60% of the significant correla-
tions between ratings of music genres and scores on traits
and facets were for the trait of Openness and the six open-
ness facets (57 of 93, or 61%); Conscientiousness was a
distant second (14 of 93, or 15%), followed by Neuroticism
(10 of 93, or 11%), Extraversion (7 of 93, or 8%), and,
finally, Agreeableness (5 of 93, or 5%).
Our focused look at the 21 music genres raises some
intriguing issues. For example, as has been noted, in his
research on metalheads, Arnett (1996), using a version of
Zuckerman, Eysenck, and Eysenck’s (1978) scale, found
that his participants were especially high sensation seekers.
Not only did they score higher on this paper and pencil
scale than a control group of nonmetalheads, but in his in-
terviews with metalheads Arnett found that they were es-
pecially likely to have engaged in a variety of risky behav-
iors in the previous few months (for example, to have driv-
en over 80 mph, to have engaged in unprotected sex, and
to have used illegal drugs). In this study, however, we found
only a moderate, and nonsignificant, correlation with the
NEO-PI scale that measures the facet of excitement seeking
(r = .12, p < .26). This suggests that there may be a mean-
ingful distinction between those who prefer certain kinds
of music, as assessed by the STOMP, and those who iden-
tify strongly with that kind of music, as perhaps is revealed
by someone calling himself or herself a metalhead or a
deadhead. Arnett notes that there is an important distinction
between what he calls a “taste culture” and a “subculture”
(1996, p. 179), and it appears that the STOMP assesses the
former but not the latter.
So, too, was it surprising that those who indicated pref-
erences for the other three intense and rebellious music
genres – rock and alternative from the original 14 genres,
and punk from the subsequently added genres – did not
come across as either intense or rebellious. They tended to
be open (ratings for alternate and punk correlated positively
with at least one of the Openness facets), but none, for ex-
ample, showed any tendency toward angry hostility (in
fact, the correlations were negative: for rock, r = –.19, p <
.08, for alternative, r = –.05, and for punk, r = –.20, p <
.11) and very little toward excitement seeking (here the cor-
relations were positive: r = .07, r = .14, r = .01), two per-
sonality characteristics one might associate with intensity
or rebelliousness. In addition, although one might expect
ratings on these genres to correlate negatively with scores
on the traits of Agreeableness and Conscientiousness, they
did not. There were no consistent patterns, and no signifi-
Journal of Individual Differences 2008; Vol. 29(1):45–55
cant correlations, between ratings of on these genres and
scores on impulsiveness, self-discipline, or dutifulness.
Moreover, as we have noted, ratings for rock and punk cor-
related positively with grades, not exactly indicating rebel-
lious academic behavior. It is likely that with the passage
of time, rock and punk have become less rebellious and
more mainstream. In fact, rock music was the highest rated
of all 21 genres in this study, with a mean of 5.65, well
above the average rating for the genres, which was 4.20.
The label for intense and rebellious may need to be
changed.
As we have noted, the trait of Openness is the most ro-
bust of the five traits, and ratings for religious music cor-
related negatively with Openness (r = –.36); the strongest
correlation was with the facet that assesses openness to val-
ues (r = –.54, the strongest of all the correlations reported
in Tables 4–8). This facet is defined in the manual in the
following way: “Openness to Values means the readiness
to reexamine social, political, and religious values” (Costa
and McCrae, 1992, p. 17). North Carolina is in what is often
called “the Bible belt,” and those who are religious in North
Carolina – at least, those in this study who indicate they
like religious music – appear not only to be sure of their
faith, but not prone to question their views or seek out al-
ternative perspectives. Interestingly, they are also unlikely
to be open to their feelings (r = –.27), defined in the manual
as “receptivity to one’s own inner feelings and emotions
and the evaluation of emotion as an important part of life”
(Costa & McCrae, 1992, p. 17). When we performed a re-
gression analysis using preferences for religious music as
the dependent variable, and age, gender, and race as the
independent variables, the regression was highly signifi-
cant (F = 6.39, df = 3, p < .001), with age as the key pre-
dictive variable (r = .41, p < .001) and race as the second
most predictive variable (r = .19, p < .06); gender did not
predict at all (r = .02, ns). The older respondents and the
students of color in our sample were more likely to prefer
religious music than the younger respondents and the white
respondents. Religious music was also one of the two low-
est ranked genres (M = 3.54); only heavy metal had a lower
average rating by the students in this study (M = 3.08).
Interestingly, the same pattern did not hold for gospel mu-
sic, which falls on the same upbeat and conventional di-
mension (the correlation between preferences for gospel
music and the Openness trait was positive, not negative, r
= .08, and gospel music had a higher average rating, M =
3.95).
In their development of the STOMP and a larger theory
of music preferences, Rentfrow and Gosling (2003, 2006)
have helped guide the way toward understanding the role
of music in people’s lives, and the relationship between
music preferences and personality. The four music dimen-
sions they established in their 2003 study provide a broad
look at some of the relationships between music preferenc-
es and personality, but the findings in this study indicate
that it is important to examine specific music genres as well
as broad music dimensions. These findings also suggest
© 2008 Hogrefe & Huber Publishers
 R.L. Zweigenhaft: A Do Re Mi Encore
that the relationship between music preferences and per-
sonality may vary for different groups based on such fac-
tors as geography (those who prefer country music in Texas
may differ from those who prefer country music in North
Carolina) and age. Moreover, the data in this study raise
some intriguing questions for future research. Why, for ex-
ample, was there a negative relationship between the rat-
ings for opera and the Conscientiousness facet that assesses
order? Why was there not the expected relationship be-
tween ratings for rock music (a genre that falls on the ques-
tionably labeled intense and rebellious dimension) and an-
gry hostility, and why was there a positive correlation be-
tween ratings on rock and grades? Perhaps future research
that combines the qualitative methods used in the earlier
research on metalheads (e.g., Arnett, 1996) and deadheads
(e.g., Adams and Sardiello, 2000) with the quantitative
benefits provided by the STOMP can answer such ques-
tions.
References
Adams, R.G. and Sardiello, R. (Eds.). (2000). Deadhead social
science: “You ain’t gonna learn what you don’t want to know.”
Walnut Creek, CA: Alta Mira Press.
Arnett, J.J. (1996). Metalheads: Heavy metal music and adoles-
cent alienation. Boulder, CO: Westview Press.
Barron, F. (1968). Creativity and personal freedom. Princeton,
NJ: Van Nostrand.
Barron, F. (1969). Creative person and creative process. New
York: Holt, Rinehart and Winston.
Bourke, R., & Francis, L.J. (2000). Personality and religion
among music students. Pastoral Psychology, 48, 437–444.
Ciuffardi, L., & Noemi, V. (2000). Personality features associated
with the performance of certain musical instruments. Interdis-
ciplinaria, 17, 1–20.
© 2008 Hogrefe & Huber Publishers
55
Costa, P.T., & McCrae, R.R. (1992). NEO-PI-R professional man-
ual. Odessa, FL: Psychological Assessment Resources.
Gillespie, W., & Myors, B. (2000). Personality of rock musicians.
Psychology of Music, 28, 154–165.
McCown, W., & Dulaney, W. (1999). The Grateful Dead experi-
ence: A factor analytic study of the personalities of people who
identify with the Grateful Dead. In Robert G. Weiner (Ed.),
Perspectives on the Grateful Dead (pp. 113–118). Westport,
CT: Greenwood Press.
Rentfrow, P.J., & Gosling, S.D. (2003). The do re mi’s of everyday
life: The structure and personality correlates of music prefer-
ences. Journal of Personality and Social Psychology, 84,
1236–1256.
Rentfrow, P.J., & Gosling, S.D. (2006). Message in a ballad: The
role of music preferences in interpersonal perception. Psycho-
logical Science, 17, 236–242.
Shuter, D. (2000). Profiling music students: Personality and reli-
giosity. Psychology of Music, 28, 190–196.
Stremikis, B.A. (2002). The personal characteristics and environ-
mental circumstances of successful women musicians. Cre-
ativity Research Journal, 14, 85–92.
Weinstein, D. (1991). Heavy metal: A cultural sociology. New
York: Lexington.
Zuckerman, M., Eysenck, S. & Eysenck, H.J. (1978). Sensation-
seeking in England and America: Cross-cultural, age, and sex
comparisons. Journal of Consulting and Clinical Psychology,
46, 139–149.
Richard Zweigenhaft
Guilford College
5800 W. Friendly Ave.
Greensboro, NC 27410
USA
Tel. +1 336 316-2269
Fax +1 336 316-2467
E-mail rzweigen@guilford.edu
Journal of Individual Differences 2008; Vol. 29(1):45–55