https://doi.org/10.1177/0305735616658957

Psychology of Music2017, Vol. 45(3) 338 –354

© The Author(s) 2016Reprints and permissions:

sagepub.co.uk/journalsPermissions.navDOI: 10.1177/0305735616658957

journals.sagepub.com/home/pom

Personality, uses of music, and music preference: The influence of openness to experience and extraversion

Elizabeth J. Vella1 and Gregory Mills2

AbstractThe purpose of this study was to ascertain whether uses of music partially mediate the link between personality and music preference. Undergraduate students (N = 122) completed the following scales: The Brief Big Five Inventory, The Uses of Music Inventory, The Short Test of Music Preference, The Life Orientation Test Revised, The Beck Depression Inventory, and the Perceived Stress Scale. Openness to experience positively predicted preferences for reflective-complex (RC; e.g., jazz/blues) and intense-rebellious (IR; e.g., rock/metal) music and was inversely related to upbeat-conventional (UC; e.g., country/pop) music, whereas extraversion was positively related to preferences for energetic-rhythmic (ER; e.g., rap/soul) and UC genres. A link between trait optimism and ER music preference was fully mediated by the more prominent extraversion trait. The relationship between openness to experience and RC music preference was partially mediated by cognitive uses of music, with a marginally significant analysis indicating partial mediation of emotional uses of music for openness to experience and IR music preference. Trait neuroticism, perceived stress, and depression scores all correlated positively with emotional uses of music. The current findings support studying personality contextually alongside uses of music when investigating music preference and shed light on how negative affect may inform emotional uses of music.

Keywordsfunctions of music, mood regulation, personality, preference, stress

It is a commonly held belief that music affects mood. Music therapies have been employed in various contexts to reduce anxiety and induce a state of calm. However, variations in music type may induce differential effects on mood. For example, directors often select musical pieces

1Department of Psychology, University of Southern Maine, USA2Department of Urology, Maine Medical Partners, USA

Corresponding author:Elizabeth J. Vella, Department of Psychology, University of Southern Maine, 96 Falmouth Street, Portland, ME 04103, USA. Email: elizabeth.vella@maine.edu

658957 POM0010.1177/0305735616658957Psychology of MusicVella and Millsresearch-article2016

Article

Vella and Mills 339

to “set a mood” for a movie scene. Music is an essential component of many social activities and differences in music preferences tend to vary widely from person to person. Pertinent factors such as personality and music preference have been known to elicit instrumental influences in predicting musical mood induction (e.g., Kreutz, Ott, Teichmann, Osawa, & Vaitl, 2008). The relationship between personality and artistic preferences has received considerable attention from psychologists over the years, in an effort to further delineate the predictive utility of traits on specific factions of human behavior. Understanding the relationships between music prefer-ence and personality may inform applications of music therapies.

Early research on the associations between personality and music preference focused sub-stantively on the extraversion construct, noting significant positive associations with prefer-ences for music characterized by strong rhythms and fast tempos (e.g., Cattell & Saunders, 1954). More recently, with the advent of the widely supported five factor model of personality, findings have pointed toward the prominence of openness to experience as a more robust trait in predicting aesthetic preferences (e.g., Cleridou & Furnham, 2014; Rawlings, Barrantes i Vidal, & Furnham, 2000). In a study investigating the link between personality and music pref-erence, Rentfrow and Gosling (2003) derived the following four genres of music preference via factor analysis of their Short Test of Music Preference (STOMP): reflective-complex (RC; blues, jazz, classical, and folk); intense-rebellious (IR; rock, alternative, and heavy metal); upbeat-conventional (UC; country, soundtrack, religious, and pop); and energetic-rhythmic (ER; rap/hip-hop, soul/funk, and electronic/dance). Significant positive correlations have been observed between openness to experience and preferences for RC and IR genres of music, with openness to experience inversely linked to UC music preference, whereas extraversion has been positively linked to ER music preference (Rentfrow & Gosling, 2003; Zweigenhaft, 2008). A recent inves-tigation among a predominately male undergraduate sample in Germany replicated these aforementioned associations between openness to experience and preferences for RC, IR, and UC genres of music, in addition to the link between extraversion and ER preference, further extending their findings to suggest additional relationships between extraversion and increased preference for IR music, and neuroticism being positively predictive of UC preference and nega-tively predictive of IR preference (Langmeyer, Guglhor-Rӧdan, & Tarnai, 2012). Although inconsistencies persist across samples regarding a purported positive link between both extra-version and agreeableness in terms of UC preference (e.g., Zweigenhaft, 2008), a clear pattern remains across studies regarding the importance of openness to experience and, perhaps to a lesser degree, extraversion.

However provocative, it is important to keep in mind that the magnitude of correlation between Big Five personality traits and music preference variables tends to cluster in the weak to moderate range (e.g., Langmeyer et al., 2012), indicating that other factors are at play in predicting music preferences, such as how listeners use music. Chamorro-Premuzic and Furnham (2007) developed an inventory to assess common ways in which individuals use music in their daily living, reporting evidence that indicates listeners using music as a means of emotion regulation, cognitive stimulation, and as a background stimulus. Indeed, evidence suggests associations between uses of music and preferences, whereby cognitive uses of music positively predict preferences for IR and RC music and inversely predict preferences for ER and UC music (Getz, Marks, & Roy, 2014). Further, Chamorro-Premuzic and Furnham (2007) hypothesized and found evidence for individual differences in uses of music, whereby individu-als rating high in trait neuroticism were found to use music as a medium for emotion regula-tion, whereas openness to experience correlated positively with cognitive uses of music. These findings provide a sense of coherence between trait and behavior; individuals experiencing a high degree of negative affect as a function of trait neuroticism may be predicted to engage in

340 Psychology of Music 45(3)

music listening as a means of regulating mood states, just as those fascinated by aesthetic beauty and art may use music as a means of cognitive stimulation by actively appreciating the complexity of musical compositions. It has also been hypothesized that trait extraversion would correlate positively with background uses of music (e.g., music listening at work and during social events as a means of maintaining an ambient degree of auditory stimulation), however such predictions have been met with mixed support (e.g., Chamorro-Premuzic & Furnham, 2007; Chamorro-Premuzic, Swami, & Cermakova, 2012).

Researchers have begun to propose and test direct and indirect models of associations among personality, uses of music, and music preference, in which uses of music may serve as interme-diary connections between trait and preference. Results have indicated a direct effect between emotional uses of music and preferences for sad compositions (minor harmonies with slow tempos), whereas background uses of music may partially mediate the link between extraver-sion and preferences for happy compositions (major harmonies, faster tempos, and dance-like rhythms) (Chamorro-Premuzic, Fagan, & Furnham, 2010). A recent study also reported evi-dence indicating that individuals rating high on trait optimism or perceived stress endorse emo-tional uses of music for their UC preference (Getz et  al., 2014). So, despite being inversely correlated constructs, stress and optimism appear to influence emotional uses of music prefer-ence independently.

Theoretical orientations designed to explain the relationships among these variables have offered that music preferences serve to reinforce self-views and, as such, signify a reflection of a social identity (e.g., Rentfrow & Gosling, 2003; Greenberg et al., 2016). To this end, music pref-erence can be regarded as a form of self-expression, a medium for expressing personality traits while simultaneously making a statement to others about oneself. Further, individual personal-ity traits may very well resonate with uses of music alongside preferences in a coherent fashion: someone rating high in openness to experience may be drawn to cognitive uses of music when listening to their preferred genre of RC compositions. In this way, cognitive uses of music may partially mediate the relationship between openness to experience and RC preference.

Another potentially useful framework for studying personality ties to music preference con-cerns the circumplex model of emotion (Russell, 1980), whereby emotions are conceptualized in a two-dimensional affective space of arousal and valence to create four quadrants of high arousal/positive valence (e.g., joviality), low arousal/positive valence (e.g., calm/relaxation), low arousal/negative valence (e.g., sadness), and high arousal/negative valence (e.g., anger). Punkanen, Eerola, and Erkkilä (2011) used the circumplex model in combination with the behavioral activation system of approach motivation (e.g., Carver & White, 1994) to hypothe-size that clinically depressed adults would display significantly lower preference ratings for music characterized as high energy and/or angry relative to their non-depressed counterparts, since these types of music are incongruent with the disengaged approach qualities of depres-sion. Results supported this hypothesis, providing evidence that complements characteriza-tions of depression in terms of impaired approach motivation and behavior (e.g., McFarland, Shankman, Tenke, Bruder, & Klein, 2006).

The current study

The purpose of the current study is to bring contemporary models to bear concerning the inter-relationships among personality traits, uses of music, and music preferences. Although past evidence suggests reliable associations between particular traits and music preferences (e.g., openness to experience and RC/IR preference), the potential mediating role for uses of music remains comparatively unexplored. Further, it is unknown whether the personality trait of

Vella and Mills 341

optimism is uniquely predictive of music preference as suggested by previous research (e.g., Getz et al., 2014) or if this variance is secondary to a more prominent overarching personality trait, such as extraversion. Finally, the current study seeks to utilize the circumplex model of emotion alongside approach motivation conceptualizations of behavior activation to guide the testing of models predicting congruence between trait and preference.

Emotion regulation: Stress, depression, and neuroticism

Given the ubiquitous nature of music listening in contemporary culture, it is theoretically rea-sonable to propose that individuals may commonly use music listening as a medium of emotion regulation during periods of emotional strain, for example following stressful life experiences or elevated symptoms of depression. As such, emotional uses of music may represent a common coping modality for acute shifts in negative affective states. Likewise, personality traits charac-terized by high levels of negative affect are also anticipated to predict emotional uses of music. To this end, recent ratings of perceived stress and depressive symptoms are hypothesized to cor-relate positively with emotional uses of music, in addition to trait neuroticism. Finally, in regard to the circumplex model of emotion and in alignment with previous research (e.g., Punkanen et al., 2011), it is expected that trait neuroticism would inversely correlate with music prefer-ences characterized as high energy/angry (e.g., IR), given the incongruence of these music genres with the disengaged approach qualities of neuroticism.

Openness to experience

Individuals rating high in openness to experience have been described as sensitive to beauty and art, intellectually curious, and to have an emotionally complex life (Costa & McCrae, 1992). Therefore, it stands to reason that openness to experience as a trait would predict preferences for musical stimuli outside the pop culture palate and perhaps disliking the con-ventional styles of music. Given the preponderance of evidence to date regarding the predic-tive associations pertaining to the openness to experience trait, it is expected that openness will be positively associated with both RC and IR music preferences and inversely associated with UC preference. Moreover, in light of intellectual curiosity coupled with sensitivity to art, it is expected that cognitive uses of music will mediate the relationship between open-ness to experience and RC music preference. Likewise, regarding emotional complexity com-bined with sensitivity to art, the circumplex model may suggest that emotional uses of music at least partially mediate the relationship between openness to experience and IR music preference.

Extraversion

Costa and McCrae (1992) have described the extraversion construct as a personality dimen-sion comprising a host of traits, including sociability, activity, and tendencies to experience joy and pleasure. Thus, trait extraversion may be conceptualized as an approach-motivated construct tending towards high arousal/positive valence on the circumplex model. As such, main effects are hypothesized between extraversion and preferences for ER and UC music. It is also anticipated that the extraverts will employ background uses of music as a means to maintain a preferred level of arousal; an exploratory ancillary hypothesis will investigate whether background uses of music mediates the relationship between extraversion and ER/UC preferences.

342 Psychology of Music 45(3)

Optimism

Characterized as a general expectation of positive outcomes in life, trait optimism may be con-ceptualized similarly to extraversion in terms of approach motivation tending towards high arousal/positive valence on the cirucumplex model. To this end, it is likewise expected that opti-mism will positively correlate with ER and UC music preferences. However, given the demon-strated significant positive correlation between trait optimism and extraversion (e.g., Sharpe, Martin, & Roth, 2011), the variance pertaining to optimism and music preference may be sec-ondary to a more prominent overarching relationship of extraversion; as such, an exploratory mediational model will be tested to ascertain whether trait optimism remains a significant pre-dictor of music preference following inclusion of extraversion as a mediator variable.

Methods

Participants

The participants in the current study included 122 undergraduate psychology majors at a mid-sized public university, who earned extra course credit for their participation (32 men, 90 women; M = 21.45 years, SD = 5.75 years; range: 18–53 years). Participants were recruited for participation through fliers posted on an introductory psychology subject pool corkboard and through announcements made in psychology courses on campus. This study received approval from the campus institutional review board and all participants signed and received a personal copy of the informed consent document prior to their participation. The sample con-sisted of Caucasian (89%), African American (3%), Hispanic (2%), Native American (1%), and multi-racial (5%) ethnicities, which roughly approximated the respective campus student pop-ulation base rates.

Materials

The Brief Big Five Personality Inventory is a 44-item Likert-type assessment of the five basic personality dimensions, including Openness to Experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism (John & Srivastava, 1999). Sample items include “I see myself as someone who …” is talkative (extraversion); worries a lot (neuroticism); and is sophisticated in art, music, or literature (openness to experience). Participants rate themselves on a scale ranging from 1 = Disagree Strongly to 5 = Agree Strongly. Internal consistencies for these scales have been found to fluctuate between .75–.9 in American and Canadian samples, with average alphas > .8, and test-retest reliabilities ranging from .8–.9 across three months (John & Srivastava, 1999). Cronbach’s alpha internal consistencies were moderate to high for the cur-rent sample: extraversion = .84, neuroticism = .68, and openness to experience = .69.

The 15-item Uses of Music Inventory was administered to measure emotional (e.g., listening to music really affects my mood), cognitive (e.g., listening to music is an intellectual experience for me), and background (e.g., I enjoy listening to music while I work) uses of music (Chamorro-Premuzic & Furnham, 2007). Respondents rated their agreement with each item on a scale from 1 = Strongly Disagree to 5 = Strongly Agree. Internal consistencies for the uses of music subscales were found to range from .76–.85 among a sample of 341 British and American university students (Chamorro-Premuzic & Furnham, 2007). The current sample demon-strated smaller internal consistencies in the moderate to high range, a factor likely impacted by the comparatively smaller sample size and the fact that each subscale includes 5 items: emo-tional = .45, cognitive = .75, and background = .62.

Vella and Mills 343

Music preferences were measured by use of the 14-item STOMP inventory (Rentfrow & Gosling, 2003). A principal components factor analysis with an eigenvalue criterion of 1 on Varimax rotation yielded a four factor solution with loadings above .5: RC Factor 1 (jazz, blues, classical, and folk); IR Factor 2 (rock, alternative, and heavy metal); ER Factor 3 (rap/hip-hop, R&B/soul, and dance/electronica); and UC Factor 4 (soundtrack, pop, and country). Religious music failed to load on any of the factors in this factor analysis and as such was not incorpo-rated in the scoring of music preferences in this sample. The four factors collectively accounted for 63% of the variance in music preference, with the RC, IR, ER, and UC factors accounting for 23%, 15%, 14%, and 10% of the total variance, respectively. Table 1 features the music prefer-ence factor scores pertaining to this principal components analysis. This four factor solution resonates with previous reports on this inventory (e.g., Langmeyer et al., 2012). Rentfrow and Gosling (2003) reported stable test-retest reliabilities of these music preferences across 3 weeks in a sample of 1,704 undergraduate students, rs .77–.89. Cronbach’s alpha internal consisten-cies for the music preferences in the current sample are as follows: RC = .76, IR = .72, ER = .55, and UC = .61.

Depression symptoms in the past two weeks were measured via administration of the 21- item Beck Depression Inventory II, a widely used depression assessment tool with demonstrated high internal consistency among college students (α = .93; Beck, Steer, & Brown, 1996). Internal consistency for this inventory in the current sample was high (α = .88). Respondents rate themselves for each item on a 4-point scale, ranging from 0 to 3, with the range of total scores being 0–63.

Trait optimism was assessed with the Life Orientation Test–Revised, a 10-item inventory designed to ascertain degree of expectations for positive outcomes in life (Scheier, Carver, & Bridges, 1994). Participants rate their agreement on each item using a 5-point Likert-type scale, ranging from 0 = Strongly Disagree to 4 = Strongly Agree. Sample items include “in uncer-tain times, I usually expect the best” and “I’m always optimistic about my future”. This inven-tory featured adequate internal consistency in the current sample (α = .75).

Table 1. Music preference factor scores (Varimax rotated pattern).

Music genre RC IR ER UC

Jazz .801 .064 .180 −.175Blues .735 .013 .140. −.265Classical .732 .168 −.200 .097Folk .664 .303 −.173 .110Religious .388 −.303 .239 .273Rock .071 .841 −.007 .021Alternative .220 .752 .016 .207Heavy Metal .069 .736 .020 −.278Rap/Hip-Hop −.243 .007 .781 .113R&B/Soul .147 −.327 .674 .176Dance/Electronica .207 .296 .633 −.131Soundtrack .174 −.074 −.146 .802Pop −.351 .119 .368 .640Country −.381 −.033 .354 .607

Note. N = 122. Factor loadings > .50 are featured in boldface and represent genres included in each of the four derived music preference factor scores. RC = reflective-complex; IR = intense-rebellious; ER = energetic-rhythmic; and UC = upbeat-conventional.

344 Psychology of Music 45(3)

Participants completed the Perceived Stress Scale, a 10-item inventory designed to tap into how unpredictable, uncontrollable, and overloaded respondents found their lives in the past month (Cohen, Kamarck, & Mermelstein, 1983). Sample items include, “how often have you been nervous or ‘stressed’?” and “how often have you felt difficulties were piling up so high that you could not overcome them?” Participants respond to each item on a 5-point Likert-type scale ranging from 0 = never to 4 = very often, with total scores ranging from 0–40. The perceived stress scale demonstrated high internal consistency in the current sam-ple (α = .89).

Procedure

Following informed consent procedures, participants were assigned a subject ID number and sent a link to the questionnaires corresponding to the current study via campus email. The questionnaires were administered online through the primary investigator’s secure socket layer encrypted account with Survey Monkey. The ordering of questionnaire presentation was as fol-lows: Beck Depression Inventory II, Uses of Music Inventory, the Brief Five Factor Inventory, the Life Orientation Test–Revised, the Perceived Stress Scale, and the STOMP.

Analytic strategy

All statistical analyses were conducted using the Statistical Package for Social Sciences, v. 19.0, with alpha levels set to .05 for detection of significant effects (SPSS, 2011). Preliminary analy-ses explored significant demographic differences in music preference and uses of music as a function of age via Pearson correlation and gender via independent sample t-test, effects that would be entered as covariates for subsequent analyses on respective music preferences. Pearson correlation was also used to test the emotion regulation hypothesis that recent bouts of emotional strain via depression and perceived stress would positively correlate with emo-tional uses of music, in addition to chronic tendencies toward negative affect as indicated by trait neuroticism. Multiple regression analyses were conducted to test specific hypotheses per-taining to the influences of personality and uses of music on music preferences factors, treated as outcome variables. First, separate regression models tested for the hypothesized associations of openness to experience as a predictor of preferences for both RC and IR. Similar analyses were then conducted for both extraversion and optimism as separate and combined predictors of ER and UC music preferences. The regression equation for the association between a trait and preference including a single covariate is y = a + b1x1 + b2x2, whereby “y” represents a music preference criterion (e.g., RC), “a” represents the intercept, b1x1 represents the slope per-taining to the covariate (e.g., age), and b2x2 represents the slope pertaining to the trait predictor variable (e.g., openness to experience). For each significant regression model, both standard-ized and unstandardized betas were provided alongside Y-predicted regression equations, to facilitate interpretation of the criterion.

A subsequent set of regressions evaluated uses of music as potential mediators between per-sonality and music preferences. Guidelines and procedures were followed for testing mediation via multiple regression analysis as suggested by Baron and Kenny (1986), entering the hypoth-esized mediator variable (e.g., cognitive uses of music) in a base model including a predictor variable (e.g., openness to experience) for estimating a criterion (e.g., RC preference). Full medi-ation is indicated when the base predictor variable becomes non-significant in the presence of the mediator, whereas partial mediation may be evident under the circumstances of a decre-mented base predictor beta weight that remains significant in the presence of a purported

Vella and Mills 345

mediator. Under these latter circumstances, the Aroian version of the Sobel test for partial mediation was employed in accord with recommendations (Baron & Kenny, 1986), to ascertain whether change in unstandardized beta weights in the presence of a hypothesized mediator represents a significant difference.

Results

Descriptive statistics pertaining to the primary variables under study are featured in Table 2. Preliminary correlational analyses indicated age to be positively associated with RC music preference, r = .19, p < .05, but unrelated to other music preferences, rs < |.12|, p > .05. Exploratory gender analyses indicated that men scored higher than women on IR music pref-erence, t(120) = 2.54, p = .01, whereas women scored higher than men in preference for UC music, t(120) = –3.63, p < .001. Further, a marginal effect was also apparent for women to display higher ER preference scores relative to men, t(120) = –1.93, p < .06. Gender differ-ences were also noted with regard to uses of music, whereby women were found to be more likely than men to endorse employing emotional uses of music, t(120) = –3.31, p = .001, whereas men were more likely than women to endorse employing cognitive uses of music, t(120) = 2.22, p = .029.

Table 2. Descriptive statistics of psychosocial variables, uses of music, and music preferences.

Variable Mean Standard Deviation (SD) Min. Max.

Openness to experience 3.6 .581 1.8 4.9Extraversion 3.4 .81 1.63 5.0Neuroticism 2.99 .645 1.63 4.75Depression 8.39 .637 0.0 32.0Optimism 14.9 4.06 5.0 23.0Perceived stress 17.89 6.92 5.0 39.0Emotional uses of music 3.75 .652 1.6 5.0Cognitive uses of music 2.51 .862 1.0 5.0Background uses of music 3.49 .765 1.4 5.0RC music preference 3.04 .892 1.0 5.0IR music preference 3.52 .985 1.0 5.0ER music preference 3.75 .787 1.67 5.0UC music preference 3.69 .860 1.33 5.0 Male FemaleUses of music Mean SD Mean SDEmotional 3.44 .591 3.86 .638**Cognitive 2.8 1.06 2.41 .76*Background 3.34 .79 3.54 .753 Male FemaleMusic preference Mean SD Mean SDReflective Complex 3.24 .742 2.97 .933Intense-Rebellious 3.9 1.03 3.39 .099*Energetic-Rhythmic 3.52 .884 3.83 .738†

Upbeat-Conventional 3.24 .974 3.85 .760**

Note. N = 122 (32 men, 90 women).†p < .06. *p < .05. **p ⩽ .001.

346 Psychology of Music 45(3)

Emotion regulation: Stress, depression, and neuroticism

In accord with the emotion regulation hypothesis, individuals experiencing recent bouts of emotional strain by way of depression and perceived stress were found to endorse emotional uses of music: r = .28, p = .002 for depression and r = .2, p = .028 for perceived stress. Interestingly, both of these forms of emotional strain were positively associated with back-ground uses of music as well: r = .31, p < .001 for depression and r = .19, p = .036 for perceived stress. Furthermore, in support of an emotion regulation model, trait neuroticism also corre-lated positively with emotional uses of music, r = .2, p = .025, and marginally with background uses of music, r = .16, p = .084. However, contrary to expectation, trait neuroticism was unre-lated to IR music preference (r = .06, p > .5), failing to support the circumplex model expecta-tion of a significant inverse association between the disengaged approach qualities of trait neuroticism with preferences for high energy/angry music. Table 3 depicts intercorrelations of psychosocial variables with uses of music and music preference variables.

Personality associations with music preference factors

Multiple regression analyses evaluated openness to experience as a predictor of RC and IR music preferences, with age and gender (men = 1, women = 2) entered as covariates, respec-tively. As hypothesized, openness to experience was positively linked to both RC (b = .532, SE = .133; β = .346, p < .001; R2 = .149, ΔR2 = .115) and IR (b = .385, SE = .148; β = .228, p = .01; R2 = .102, ΔR2 = .051) music preferences. Further, openness to experience was inversely linked to UC music preference (b = –.492, SE = .121; β = –.332, p < .001; R2 = .208, ΔR2 = .11). Similarly, entering gender as a covariate, extraversion was likewise observed as a significant predictor of preferences for ER (b = .336, SE = .084; β = .346, p < .001; R2 = .144, ΔR2 = .114) and UC (b = .241, SE = .092; β = .227, p = .01; R2 = .148, ΔR2 = .049) music. Importantly, a model estimating UC music preference, including gender as a covariate, involving simultaneous entry of both openness to experience (b = –.583, SE = .117; β = –.394, p < .001) and extraver-sion (b = .327, SE = .086; β = .308, p < .001) remained significant with neither trait decre-menting in strength, suggesting that each trait adds unique variance in predicting UC preference (R2 = .295, ΔR2 = .197).

Analyses pertaining to the optimism construct yielded mixed results. Contrary to expecta-tion, optimism was unrelated to UC music preference: β = .111, p = .2; in accord with expecta-tion, a model including gender as a covariate indicated a significant effect for trait optimism on ER music preference (b = .035, SE = .017; β = .179, p = .048; R2 = .062, ΔR2 = .032). However, simultaneous entry of both extraversion and optimism in a model estimating ER music prefer-ence, including gender as a covariate, yielded a sustained significant effect for extraversion (b = .317, SE = .093; β = .326, p < .001) and a non-significant effect for optimism (b = .009, SE = .018; β = .048, p = .6), suggesting that the relationship between trait optimism and ER music preference is fully mediated by extraversion (R2 = .146, ΔR2 = .116). Indeed, following criteria delineated by Baron and Kenny (1986), each condition of mediation was satisfied. Figure 1 displays the fully mediated model. Table 4 lists Y-predicted regression equations for significant personality associations with music preference factors.

Uses of music as a mediator between personality and music preference

Hypothesized relationships stipulated that cognitive uses of music would at least partially medi-ate the relationship between openness to experience and RC music preference. Indeed, multiple

Vella and Mills 347

Tab

le 3

. In

terc

orre

latio

ns o

f psy

chos

ocia

l var

iabl

es w

ith u

ses

of m

usic

and

mus

ic p

refe

renc

es.

Mea

sure

12

34

56

78

91

01

11

21

3

1.

Neu

roti

cism

—−

.26

*−

.01

−.6

5**

.55

*.6

5**

.2†

.1.1

6.0

1.0

6−

.09

−.0

42

. Ex

trav

ersi

on—

.18

†.4

2**

−.2

4*

−.1

8†

.09

−.1

6.0

3−

.06

.09

.37

**.2

8*

3.

Ope

nn

ess

to e

xper

ien

ce—

.16

.1−

.11

.2†

.32

**.2

1†

.37

**.2

5*

−.0

3−

.36

**4

. O

ptim

ism

—−

.5**

−.5

8**

−.0

4−

.12

−.0

6−

.05

−.0

6.2

†.1

55

. D

epre

ssio

n—

.69

**.2

8*

.15

.31

**.1

2−

.05

−.0

6−

.14

6.

Per

ceiv

ed S

tres

s Sc

ale

—.2

†.0

4.1

9†

−.0

4−

.07

−.0

4.0

37

. Em

otio

nal

use

s of

mu

sic

—.0

7.2

8**

.02

.14

.07

.18

. C

ogn

itiv

e u

ses

of m

usi

c—

.19

†.3

9**

.13

−.1

7−

.46

**9

. B

ackg

rou

nd

use

s of

mu

sic

—.0

5.1

4−

.04

−.1

21

0.

Ref

lect

ive-

Com

plex

mu

sic

pref

eren

ce—

.3**

.00

4−

.32

**1

1.

Inte

nse

-Reb

ellio

us

mu

sic

pref

eren

ce—

−.1

−.0

81

2.

Ener

geti

c-R

hyt

hm

ic m

usi

c pr

efer

ence

—.3

**1

3.

Upb

eat-

Con

ven

tion

al m

usi

c pr

efer

ence

—

Not

e. N

= 1

22 (

32 m

en, 9

0 w

omen

).† p

⩽ .0

5. *

p ⩽

.01.

**p

⩽ .0

01.

348 Psychology of Music 45(3)

regression including gender as a covariate revealed openness to experience to positively predict cognitive uses of music (b = .457, SE = .127; β = .308, p < .001; R2 = .133, ΔR2 = .094); like-wise, cognitive uses of music with age as a covariate predicted RC music preference (b = .426, SE = .085; β = .411, p < .001; R2 = .202, ΔR2 = .168). Simultaneous inclusion of both openness to experience and cognitive uses of music in a model estimating RC music preference, after con-trolling for participant age, reduced the strength of openness to experience as a predictor from β = .346, p < .001 to β = .227, p = .01, suggesting a partially mediated effect (R2 = .245, ΔR2 = .211). Using the unstandardized regression coefficients and corresponding standard errors for openness to experience (predictor) on cognitive uses of music (mediator) and cognitive uses of music (mediator) on RC music preference (criterion), the Aroian version of the Sobel test was significant, z = 2.88, SE = .067, p = .004, thereby confirming a partially mediated effect (see Figure 2).

In addition to cognitive uses of music mediating the link between openness to experience and RC music preference, we hypothesized emotional uses of music to at least partially mediate the association between openness to experience and IR music preference. First, multiple regression including gender as a covariate indicated a positive relationship between openness to experience and emotional uses of music (b = .258, SE = .096; β = .23, p = .008; R2 = .136, ΔR2 = .052). Emotional uses of music was also found to positively predict IR music preference, including gender as a covariate (b = .345, SE = .137; β = .228, p = .01; R2 = .099, ΔR2 = .048). A multiple regression model involving simultaneous inclusion of openness to experience and emotional uses of music to predict IR preference, including gender as a covariate, reduced the strength of openness to experience as a predictor from β = .228, p = .01 to β = .186, p = .038, suggesting another partially mediated effect (R2 = .131, ΔR2 = .080). However, the Aroian version of the Sobel test to evaluate the sig-nificance of change in prediction following mediator inclusion using unstandardized regres-sion coefficients and standard errors was only marginally significant, z = 1.79, SE = .05, p = .07. Figure 3 features the marginally significant partial mediation model. See Table 4 for regression equations pertaining to personality as a predictor of uses of music and uses of music as a predictor of music preferences.

Figure 1. The relationship between trait optimism and energetic-rhythmic music preference is fully mediated by extraversion. N = 122. All regressions incorporate gender as a covariate: β = .173, p = .056 for ER preference; β = .215, p = .017 for extraversion.

Vella and Mills 349

Exploratory ancillary hypotheses proposed that extraverts may employ background uses of music for their preferred genres, whereas emotional uses of music may partially mediate the relationship between optimism and musical preferences. However, extraversion was unrelated to background uses of music, r = .03, p = .75, and optimism was likewise unrelated to emo-tional uses of music, r = –.04, p = .65.

Figure 2. The relationship between openness to experience and reflective-complex music preference is partially mediated by cognitive uses of music. N = 122. Age entered as a covariate for models predicting reflective-complex music preference, β = .185, p = .04. Gender entered as a covariate for model predicting cognitive uses of music, β = –.173, p = .046. Aroian version of the Sobel test for partial mediation was significant, z = 2.88, p = .004.

Table 4. Y-predicted regression equations for significant models including personality, music preference factors, and uses of music.

Predictor Covariate Criterion Equation (y = a + b1x1 + b2x2)

Personality predictors of music preferenceOpenness Age RC RC = .741 + .018(age) + .532(openness)Openness Gender IR IR = 2.94 – .461(gender) + .385(openness)Openness Gender UC UC = 4.94 + .559(gender) – .492(openness)Extraversion Gender ER ER = 2.03 + .155(gender) + .336(extraversion)Extraversion Gender UC UC = 1.98 + .42(gender) + .241(extraversion)Optimism Gender ER ER = 2.76 + .269(gender) + .035(optimism)Personality predictors of music usesOpenness Gender CUOM CUOM = 1.46 – .337(gender) + .457(openness)Openness Gender EUOM EUOM = 2.03 + .455(gender) + .258(openness)Uses of music predictors of music preferenceCUOM Age RC RC = 1.22 + .035(age) + .426(CUOM)EUOM Gender IR IR = 3.36 – .651(gender) + .345(EUOM)

Note. N = 122. Y-predicted regression equations include the following terms: y = criterion, a = constant, b1 = covariate slope, x1 = covariate value, b2 = predictor slope, and x2 = predictor value. RC = reflective-complex; IR = intense-rebel-lious; UC = upbeat-conventional; and ER = energetic-rhythmic. CUOM = cognitive uses of music. EUOM = emotional uses of music.

350 Psychology of Music 45(3)

Discussion

The current study sought to replicate and extend previous research concerning provocative associations among personality, uses of music, and music preference, incorporating theoretical orientations of the circumplex model of emotion alongside approach motivation conceptual-izations of behavioral activation to test for coherence between self-views and music preference. To this end, the results highlight the prominence of openness to experience and extraversion for predicting music preference in divergent ways, with the former trait predicting preferences for RC and IR music, and the latter trait predictive of ER and UC music preferences. Another key contribution of the current study concerns mediational analyses regarding uses of music, whereby individuals rating high in openness to experience may use music for cognitive pur-poses when listening to preferred RC compositions. Recent findings indicated that openness to experience broadly predicted listener preferences for music perceived as rating high in “depth”, a generic music attribute factor characterized as intelligent, sophisticated, inspiring, complex, and thoughtful (Greenberg et al., 2016); these findings resonate with the current results that cognitive uses of music partially mediate the relationship between openness to experience and RC music preference.

The emotion regulation hypothesis was supported with evidence indicating recent symp-toms of depression and stress correlating positively with emotional uses of music, in addition to trait neuroticism. The circumplex model of emotion was supported in regard to approach moti-vation for extraversion: personality dimension characterized by sociability, activity, and tenden-cies to experience positive emotions was predictive of preference for music associated with high arousal/positive valence (ER and UC). To a lesser extent, the circumplex model prediction of emotional uses of music partially mediating the relationship between openness to experience and IR music preference offered only marginally significant support, whereas the circumplex expectation of disengaged approach qualities of neuroticism being inversely linked to IR music preference, was unsupported.

Figure 3. Marginally signficant partial mediation model: the relationship between openness to experience and intense-rebellious music preference, including emotional uses of music as a mediator. N = 122. Gender entered as a covariate in all models: β = .289, p = .001 for emotional uses of music and β = –.226, p = .012 for intense-rebellious music preference. Aroian version of the Sobel test for partial mediation was marginally signficant, z = 1.79, p = .07.

Vella and Mills 351

Preliminary analyses investigated age and gender associations with music preferences and uses of music, indicating a positive relationship between age and RC music preference, in addi-tion to gender differences whereby men preferred IR music to a larger degree than women, whereas women preferred UC music to a higher degree than men, as well as a marginally sig-nificant tendency for women to prefer ER music. In terms of music uses, although age was unrelated to this construct, women were observed as more likely to endorse emotional uses of music, whereas men endorsed cognitive uses of music to a higher degree than women. These findings have been echoed in the recent literature concerning age-related associations with music preference (Bonneville-Roussy, Rentfrow, Xu, & Potter, 2013) as well as gender differ-ences in music preference (Clark & Giacomantonio, 2013) and uses of music (Chamorro-Premuzic et al., 2012).

Many of the hypothesized relationships were supported by the current findings. The emo-tion regulation hypothesis was widely supported by evidence indicating a positive relationship between emotional strain and emotional uses of music across multiple constructs for trait neuroticism, perceived stress, and recent depression symptoms, in addition to a positive link between these constructs and background uses of music. Although previous findings have reported both neuroticism and perceived stress as positive predictors of emotional uses of music (Chamorro-Premuzic et al., 2012; Getz et al., 2014), the current findings offer new-found evidence that these results extend to depressive symptoms as well as to background uses of music. Collectively, these findings assert that under periods of emotional strain, individuals may be prone to use their music listening for the purposes of emotion regulation and as a background stimulus that may serve as a source of company, thereby supporting music ther-apy applications.

With regard to the associations between personality traits and music preferences, the results of the current study revealed, as expected, openness to experience as positively linked to prefer-ences for both RC and IR music, while being inversely related to UC music, confirming and strengthening previous accounts that individuals rating high on openness to experience prefer music stimuli outside the mainstream palate. Further, in support of circumplex models of emo-tion as related to behavioral activation and personality, extraversion was found to be positively associated with preference for ER and UC music (consistent with Rentfrow & Gosling, 2003 and Zweigenhaft, 2008) and optimism was positively associated with ER music preference. However, simultaneous entry of both extraversion and optimism in a model estimating ER music prefer-ence resulted in a significant effect for extraversion and a non-significant effect for optimism. Therefore, the predicted relationship between optimism and ER music preference was fully mediated by the more prominent role of extraversion, suggesting that music preference asso-ciations of optimism may be secondary to the foundational extraversion trait. Contrary to Getz et al. (2014), the results of the current study do not support an association between trait opti-mism and preferences for UC music, which may reflect a spurious finding in the prior research.

As previously stated, the strength of the correlations between Big Five personality traits and music preference variables tend to be in the weak to moderate range (e.g., Greenberg et  al., 2016; Langmeyer et al., 2012), a general finding corroborated by current research evidence that openness to experience and extraversion account for anywhere between 5% to 12% of the variance in music preferences, indicating that other factors are at play in predicting music pref-erence, such as uses of music. The results of the current report offer some evidence that uses of music partially mediate the relationship between openness to experience and music preference. As hypothesized, cognitive uses of music partially mediated the relationship between openness to experience and RC music preference. A similar relationship was also apparent for emotional uses of music to partially mediate the link between openness to experience and IR music

352 Psychology of Music 45(3)

preference, although the Sobel test for partial mediation was only marginally significant (p = .06), requiring replication in subsequent research to confirm a reliable association.

The results of the current study suggest that individuals rating high in openness to experi-ence, the trait most commonly associated with artistic/music preference in the extant litera-ture, may be prone to use music in particular ways congruent with their preferred genres, appreciating the composition complexity and musician technique when listening to preferred RC music and potentially regulating emotional experience when listening to preferred IR music. These findings parallel previous work investigating the purported links among personality, uses of music, and music preference (e.g., Getz et al., 2014) and add to theoretical orientations put forth by Chamorro-Premuzic and Furnham (2007), regarding the interplay between uses of music and preference. However, unlike Chamorro-Premuzic et al. (2010), the current study did not confirm an association between extraversion and background uses for listener-preferred music, suggesting that this association may be unreliable. That said, Chamorro-Premuzic et al. (2010) had their participants listen to musical clips previously rated for emotional content in the lab rather than completing a music preference inventory, a methodological distinction that may account for this discrepancy.

On the whole, evidence in support of circumplex predictions concerning trait predictors of music preference is mixed and seems more cogent with regard to approach-motivated qualities of disposition (e.g., extraversion in relation to ER/UC preference and emotional uses of music as a potential partial mediator between openness to experience and IR preference) than disen-gaged trait qualities (e.g., neuroticism unrelated to IR preference). In the current study, coher-ence between trait and preference was more evident with respect to approach-motivated dispositions concerning extraverted sociability and preferences for musical qualities character-ized as energetic, rhythmic, and upbeat or trait qualities reflecting artistic interests outside the mainstream palate and emotional uses of music for genres characterized as intense and rebel-lious. Interestingly, Punkanen et al. (2011) used the circumplex model to predict that clinically depressed participants would display significantly lower preferences for music rated as high energy/angry relative to their non-depressed counterparts due to the disengaged approach qualities of the clinical condition. A similar hypothesis in relation to trait neuroticism was unsupported in the current study. Indeed, depressive symptoms and neuroticism were both unrelated to IR music preference (see Table 3). It is possible that the findings reported by Punkanen et al. (2011) do not extend to non-clinical samples and/or that methodological dis-tinctions account for the current discrepancy in findings. Punkanen et al. (2011) had their participants listen to a variety of orchestral musical excerpts from original motion picture scores that vary in arousal and valence and to rate their preference following each musical piece rather than complete a brief questionnaire regarding genre preferences.

As with any other, the current study is not without limitations. The cross-sectional nature of the current design limits the precision with which we may effectively estimate relationships among the variables under study. A more compelling methodology would incorporate longitu-dinal assessments capable of capturing trait specific contextual effects of music uses for pre-ferred genres. Although the results of the study resonate with previous findings in this area (e.g., Rentfrow & Gosling, 2003; Zweigenhaft, 2008), it is possible that the sample under study may constrain the generalizability of the results. The participants under study were exclusively undergraduate and predominately female (74%) and Caucasian (89%), which may not reflect the general population. Also worthy of mention as a potential limitation to the current study concerns recent changes to the assessment of music preferences in the literature, moving away from the four factor STOMP model of RC, IR, UC, and ER preferences to a more comprehensive five factor model entitled MUSIC, which accounts for sub-genres described as mellow (e.g., soft

Vella and Mills 353

rock), unpretentious (e.g., bluegrass), sophisticated (e.g., classical), intense (e.g., metal), and contemporary (e.g., rap) (e.g., Rentfrow, Goldberg, & Levitin, 2011; Rentfrow et al., 2012). As such, the current findings may be limited to the extent that the MUSIC model offers more assess-ment precision than afforded by the STOMP model.

Future research in this area should address these limitations by examining music prefer-ences in the context of a more representative participant sample over time. The current report adds to the extant literature in a number of ways: demonstrating that recent bouts of stress and depressive symptoms, as well as trait neuroticism, inform emotional uses of music; supporting use of circumplex orientations of emotion for linking trait to preference in regard to approach-motivated dispositions; and showing that personality traits in addition to the different uses of music both have predictive value with regard to music preferences in a way that supports study-ing these constructs in concert, particularly in reference to openness to experience. Continued examination of the uses of music during certain activities and/or emotional contexts could provide valuable insight into the relationship between personality and music preferences.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychologi-cal research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. doi: 10.1037/0022–3514.51.6.1173

Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation.

Bonneville-Roussy, A., Rentfrow, P. J., Xu, M. K., & Potter, J. (2013). Music through the ages: Trends in musical engagement and preferences from adolescence through middle adulthood. Journal of Pesonality and Social Psychology, 105(4), 703–717. doi: 10.1037/a0033770

Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The BIS/BAS scales. Journal of Personality and Social Psychology, 67(2), 319–333. doi: 10.1037/0022–3514.67.2.319

Cattell, R. B., & Saunders, D. R. (1954). Musical preferences and personality diagnosis: I. A fac-torization of one hundred and twenty themes. Journal of Social Psychology, 39(1), 3–24. doi: 10.1080/00224545.1954.9919099

Chamorro-Premuzic, T., Fagan, P., & Furnham, A. (2010). Personality and uses of music as predictors of preferences for music consensually classified as happy, sad, complex, and social. Psychology of Aesthetics, Creativity, and the Arts, 4(4), 205–213. doi: 10.1037/a0019210

Chamorro-Premuzic, T., & Furnham, A. (2007). Personality and music: Can traits explan how people use music in daily life? British Journal of Psychology, 98, 175–185. doi: 10.1348/ 000712606X111177

Chamorro-Premuzic, T., Swami, V., & Cremakova, B. (2012). Individual differences in music consump-tion are predicted by uses of music and age rather than emotional intelligence, neruoticism, extra-version or openness. Psychology of Music, 40(3), 285–300. doi: 10.1177/0305735610381591

Clark, S. S., & Giacomantonio, S. G. (2013). Music preferences and empathy: Toward predicting prosocial behavior. Psychomusicology, 23(3), 177–186. doi: 10.1037/a0034882

Cleridou, K., & Furnham, A. (2014). Personality correlates of aesthetic preferences for art, architecture, and music. Empirical Studies of the Arts, 32(2), 231–255. doi: 10.2190/EM.32.2.f

Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 386–396.

354 Psychology of Music 45(3)

Costa, P. T., & McCrae, R. R. (1992). Normal personality assessment in clinical practice: The NEO person-ality inventory. Psychological Assessment, 4(1), 5–13. doi: 10.1037/1040–3590.4.1.5

Getz, L. M., Marks, S., & Roy, M. (2014). The influence of stress, optimism, and music training on music uses and preferences. Psychology of Music, 42(1), 71–85. doi: 10.1177/0305735612456727

Greenberg, D., Kosinski, M., Stillwell, D. J., Monteiro, B. L., Levitin, D. J., & Rentfrow, P. J. (2016). The song is you: Preferences for musical attribute dimensions reflect personality. Social Psychological and Personality Science. Advance online publication. doi: 10.1177/1948550616641473

John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theo-retical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality theory and research (pp. 102–138). New York, NY: Guilford Press.

Kreutz, G., Ott, U., Teichmann, D., Osawa, P., & Vaitl, D. (2008). Using music to induce emotions: Influences of musical preference and absorption. Psychology of Music, 36(1), 101–126. doi: 10.1177/0305735607082623

Langmeyer, A., Guglhor-Rӧdan, A., & Tarnai, C. (2012). What do music preferences reveal about person-ality? A cross-cultural replication using self ratings and ratings of music samples. Journal of Individual Differences, 33(2), 119–130. doi: 10.1027/1614–0001/a000082

McFarland, B. R., Shankman, S. A., Tenke, C. E., Bruder, G. E., & Klein, D. N. (2006). Behavioral activa-tion deficits predict the six-month course of depression. Journal of Affective Disorders, 91, 229–234. doi: 10.1016/j.jad.2006.01.012

Punkanen, M., Eerola, T., & Erkkilä, J. (2011). Biased emotional preferences in depression: Decreased liking of angry and energetic music by depressed patients. Music and Medicine, 3(2), 114–120. doi: 10.1177/1943862110395597

Rawlings, D., Barrantes i Vidal, N., & Furnham, A. (2000). Personality and aesthetic preference in Spain and England: Two studies relating sensation seeking and openness to experience to liking for paint-ings and music. European Journal of Personality, 14(6), 553–576.

Rentfrow, P. J., Goldberg, L. R., & Levitin, D. J. (2011). The structure of musical preferences: A five-factor model. Journal of Personality and Social Psychology, 100(6), 1139–1157. doi: 10.1037/a0022406

Rentfrow, P. J., Goldberg, L. R., Stillwell, D. J., Kosinski, M., Gosling, S. D., & Levitin, D. J. (2012). The song remains the same: A replication and extension of the MUSIC model. Music Perception, 30(2), 161–185. doi: 10.1525/mp.2012.30.2.161

Rentfrow, P. J., & Gosling, S. D. (2003). The do re mi’s of everyday life: The structure and personality correlates of music preferences. Journal of Personality and Social Psychology, 84(6), 1236–1256. doi: 10.1037/0022–3514.84.6.1236

Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39(6), 1161–1178.

Scheier, M. F., Carver, C. S., & Bridges, M. W. (1994). Distinguishing optimism from neuroticism (and trait anxiety, self-mastery, and self-esteem): A re-evaluation of the Life Orientation Test. Journal of Personality and Social Psychology, 67, 1063–1078. doi: 10.1037/0022–3514.67.6.1063

Sharpe, J. P., Martin, N. R., & Roth, K. A. (2011). Optimism and the big five factors of personality: Beyond neuroticism and extraversion. Personality and Individual Differences, 51(8), 946–951. doi: 10.1016/j.paid.2011.07.033

SPSS, Inc. (2011). SPSS for Windows software (Version 19.0.1). Chicago, IL: Author.Zweigenhaft, R. L. (2008). A do re mi encore: A closer look at personality correlates of music preference.

Journal of Individual Differences, 29(1), 45–55. doi: 10.1027/1614–0001.29.1.45