The mechanism of music for reducing psychological stress: Music preference as a mediator

The Arts in Psychotherapy 48 (2016) 62–68

Contents lists available at ScienceDirect

The Arts in Psychotherapy

The mechanism of music for reducing psychological stress: Music preference as a mediator Jun Jiang a , Daphne Rickson b , Cunmei Jiang c,∗ a b c

Education College, Shanghai Normal University, Shanghai, China New Zealand School of Music, Wellington, New Zealand Music College, Shanghai Normal University, Shanghai, China

a r t i c l e

i n f o

Article history: Received 22 August 2015 Received in revised form 7 February 2016 Accepted 16 February 2016 Available online 27 February 2016 Keywords: Valence Arousal Music preference Psychological stress Mediator

a b s t r a c t In order to examine the mechanisms through which music might alleviate psychological stress, a study of the effects of music listening following induced stress was conducted. Female music education students (N = 200) were randomly assigned to eight groups, after experiencing induced stress via a mental arithmetic test. Individuals in each group listened through headphones to one piece of music classified in terms of the levels of arousal and valence of music, and familiarity. Participants rated their tension and state anxiety levels before and after music listening, as well as their levels of valence and arousal for music, music preference, and familiarity, after listening. The results revealed that the levels of arousal and valence, and the degree of music preference predicted tension and state anxiety levels, and the effects of music valence and arousal on stress reduction were partially mediated by music preference. The most important factor in reducing stress was the degree of liking for the music, but not the degree of familiarity with the music. Our findings have important implications for individuals, and clinicians, who use music to reduce stress. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Psychological stress (hereafter referred to as stress) is one of the most critical problems in modern society, and has become a great risk to human health. Listening to music is a non-invasive intervention, which can be used to alleviate stress because of its close correlation with emotion (e.g., Bradt, Dileo, & Shim, 2013; Linnemann, Ditzen, Strahler, Doerr, & Nater, 2015; Robb, Nichols, Rutan, Bishop, & Parker, 1995; Thoma et al., 2015). Not all music is appropriate for stress reduction however (Chafin, Roy, Gerin, & Christenfeld, 2004; Yehuda, 2011). Knowing how various factors associated with music listening can impact on stress reduction would enable music listeners to more readily choose music appropriate for this purpose, and increase efficacy. Music is considered to have properties that induce arousal (low arousal vs. high arousal) and valence (low pleasure vs. high pleasure), which are in turn mediated by listeners’ music training background, music preference, and familiarity with the piece. More specifically, low-arousal music is considered to be more effective than high-arousal music in alleviating tension or state

∗ Corresponding author at: Music College, Shanghai Normal University, 100 E. Guilin Road, Shanghai 200234, China. E-mail address: [email protected] (C. Jiang). http://dx.doi.org/10.1016/j.aip.2016.02.002 0197-4556/© 2016 Elsevier Ltd. All rights reserved.

anxiety (Fisher & Greenberg, 1972; Gan, Lim, & Haw, 2015; Iwanaga, Kobayashi, & Kawasaki, 2005; Iwanaga & Moroki, 1999; Lingham & Theorell, 2009; Sandstrom & Russo, 2010); high-pleasure music tends to reduce stress more effectively than low-pleasure music (Sandstrom & Russo, 2010); musically trained listeners were found to have a lower state anxiety score after listening to low-arousal music compared with untrained listeners (Smith & Morris, 1977); familiar music can make listeners less anxious (Sung, Lee, Li, & Watson, 2012), calm and tranquil (Margounakis & Politis, 2012). Furthermore, it has been suggested that music preference correlates negatively with state anxiety (Smith & Morris, 1977), and positively with relaxation (Stratton & Zalanowski, 1984). The effect of preferred music listening on stress reduction has been reported for college students (Davis & Thaut, 1989; Iwanaga & Moroki, 1999; Jeong, 2008), air traffic controllers (Lesiuk, 2008), and patients (Clark et al., 2006; Rosenow & Silverman, 2014). There is also a positive correlation between years of music training and perceived and felt emotions (Castro & Lima, 2014; Lima & Castro, 2011; Rawlings & Leow, 2008). In these studies, the longer the music training listeners have, the more intensely they were able to perceive and experience the emotional impact of the music. In contrast, other studies found no association between arousal (Rohner & Miller, 1980; Stratton & Zalanowski, 1984), music preference (Sandstrom & Russo, 2010), familiarity (Chafin et al., 2004; Hatta & Nakamura, 1991), music training (Knight & Rickard, 2001;

J. Jiang et al. / The Arts in Psychotherapy 48 (2016) 62–68

Laohawattanakun et al., 2011) and stress reduction. These contradictory findings suggest the relationships amongst these factors are complex. Indeed, several studies have observed the interactions between arousal and familiarity (Iwanaga, Ikeda, & Iwaki, 1996), between arousal and music preference (Jiang, Zhou, Rickson, & Jiang, 2013), and between music type and music training (Wang, 2014). For example, Jiang et al. (2013) found that listening to lowarousal music induced significantly lower tension and state anxiety levels than listening to high-arousal music when music was not preferred. However, there was no significant difference between low-arousal and high-arousal music for reducing tension and state anxiety levels when music was preferred. The aim of the present study was to examine the mechanisms through which music might alleviate psychological stress. Specifically we aimed to determine how the factors influence stress reduction and which factor plays the most important role in stress reduction. In order to examine the role of music training on stress reduction, we included participants with different lengths of music training ranging from 0.5 to 14 years. Potential participants who experienced induced stress after a mental arithmetic test were included in this study.

2. Method 2.1. Participants Two hundred and eighty female undergraduates majoring in music education were recruited for this study. This study was approved by the Ethics Committee of Shanghai Normal University. All participants had normal hearing, volunteered for the research and provided written informed consents. Eighty participants were excluded because they did not experience induced stress after a mental arithmetic test. Consequently, 200 participants took part in this experiment and were assigned to eight experimental groups (n = 25), each listening to one piece of music: familiar high arousal-high pleasure music (FHHM), familiar high arousal-low pleasure music (FHLM), familiar low arousal-low pleasure music (FLLM), familiar low arousal-high pleasure music (FLHM), unfamiliar high arousal-high pleasure music (UHHM), unfamiliar high arousal-low pleasure music (UHLM), unfamiliar low arousal-low pleasure music (ULLM), and unfamiliar low arousal-high pleasure music (ULHM). Table 1 presents the participants’ characteristics and the means and standard deviations of tension and state anxiety levels before and after the mental arithmetic task (stressor). A one-way analysis of variance (ANOVA) indicated that there were no significant differences in age, F(7,192) = 1.31, p = .249, 2p = .05, and length of music training, F(7,192) = 0.85, p = .549, 2p = .03, across the eight groups. To assess the efficacy of the stressor, we conducted a two-way ANOVA on tension or state anxiety level, with time (prestressor and poststressor) as the within-subjects variable and group (FHHM, FHLM, FLLM, FLHM, UHHM, UHLM, ULLM and ULHM) as the between-subjects variable. For the tension level, There was a significant main effect of time, F(1,192) = 1007.81, p < .001, 2p = .84, but the main effect of group, F(7,192) = 0.55, p = .797, 2p = .02, and the time by group interaction, F(7,192) = 0.43, p = .884, 2p = .02, were not significant. For the state anxiety level, the main effect of time was significant, F(1,192) = 604.13, p < .001, 2p = .76, although the main effect of group, F(7,192) = 1.27, p = .266, 2p = .04, and the interaction between time and group, F(7,192) = 0.49, p = .844, 2p = .02, did not reach significance. These results indicated that the mental arithmetic task increased participants’ stress levels, and the eight groups did not differ in the tension and state anxiety levels prior to listening to music.

63

2.2. Stimuli A pretest study was conducted to select music excerpts for this study. Thirty two music excerpts were selected by the experimenters, each representing one of the four quadrants of Russell’s (1980) circumplex model. In this model, emotional states reflect a mixture of two core dimensions, valence and arousal representing pleasantness (hedonic tone) and excitation (intensity). There are thus four quadrants in the circumplex model including high-pleasure and high-arousal emotions, low-pleasure and high-arousal emotions, low-pleasure and low-arousal emotions, and high-pleasure and low-arousal emotions. The heuristic value of the two-dimensional model was confirmed measuring emotional expression and induction through music (e.g., Eerola & Vuoskoski, 2011; Ritossa & Rickard, 2004; Schubert, 1999; Vieillard et al., 2008). Twenty female undergraduates majoring in music education who did not participate in the formal experiment listened to the excerpts, and rated each on valence, arousal, and familiarity. In terms of valence, the excerpts with a mean rating higher than five were considered as high-pleasure music, a mean rating lower than three being lowpleasure music. Similarly, the excerpts with a mean arousal rating higher than five were high-arousal music, while those with a mean rating lower than three were considered as low-arousal music. Familiar excerpts were those with a familiarity rating higher than five, while unfamiliar excerpts were those with a familiarity rating lower than three. In order to ensure the participants in the FHHM, FHLM, FLLM, and FLHM groups were familiar with music excerpts, all of the participants were asked to listen repeatedly to the four pieces of music rated ‘familiar’ in the pre-test (Victory, Mars—the Bringer of War, Ase’s Death, and Tempo di Bolero moderato assai), and as far as possible to remember them within two weeks prior to the formal experiment. Based on the ratings, the most representative eight excerpts (see Table 2) were selected as the music stimuli. All music excerpts were instrumental music composed by Western composers to avoid the influence of lyrics on listeners’ emotional responses to music (Ali & Peynircioglu, 2006; Brattico et al., 2011; Hunter, Schellenberg, & Schimmack, 2010; Stratton & Zalanowski, 1994). The duration of all excerpts was 4:35 with a 1s fade-in and/or a 1-s fade-out (Kreutz, Ott, Teichmann, Osawa, & Vaitl, 2008; Sandstrom & Russo, 2010).

2.3. Measures A mental arithmetic test as a stressor was administered to induce participants’ stress. There were 50 items (e.g., 1. 8026 − 37 = ? 10. 7386 × 2 = ?) in the test, and participants were required to complete the items within 5 min. The Tension Rating Scale was used to measure the levels of tension participants felt. They were required to rate their own levels of tension on a 4-point Likert scale with 1 being “not at all”, 2 being “slightly”, 3 being “moderately”, and 4 being “very much”. Higher ratings indicate more tense participants felt. The State Anxiety Inventory, a subscale of State-Trait Anxiety Inventory (STAI; Spielberger, 1983), was used to measure the state anxiety levels. STAI is a 40-item self-report questionnaire with 20 of the items making up the State Anxiety Inventory (SAI), and the other 20 items making up the Trait Anxiety Inventory (TAI). The STAI uses a 4-point Likert scale response format with a range from 1 (“not at all” for the SAI or “almost never” for the TAI) to 4 (“very much so” for the SAI or “almost always” for the TAI). Higher scores indicate higher levels of anxiety. Test–retest stability coefficients for the SAI range from .16 to .62, and correlations for the TAI range from .73 to .86 (Spielberger, 1983). The STAI has also demonstrated good concurrent and construct validity (Spielberger, Sydeman, Owen, & Marsh, 1999).

64

J. Jiang et al. / The Arts in Psychotherapy 48 (2016) 62–68

Table 1 Participants’ characteristics and means and standard deviations of tension and state anxiety levels before and after the stressor for the eight groups. Group

FHHM FHLM FLLM FLHM UHHM UHLM ULLM ULHM

Age

20.28 (1.31) 20.60 (1.83) 20.64 (1.38) 20.52 (1.45) 20.28 (1.57) 19.88 (1.24) 20.36 (1.22) 19.76 (1.16)

Length

4.92 (3.51) 4.00 (2.67) 4.58 (3.14) 4.92 (2.86) 4.98 (3.12) 5.70 (3.54) 4.86 (3.36) 3.88 (2.96)

Tension

State anxiety

Prestressor

Poststressor

Prestressor

Poststressor

1.24 (0.52) 1.24 (0.44) 1.24 (0.44) 1.24 (0.52) 1.28 (0.46) 1.12 (0.33) 1.28 (0.46) 1.24 (0.44)

2.88 (0.83) 2.92 (0.70) 2.72 (0.79) 2.80 (0.82) 3.04 (0.79) 2.72 (0.68) 2.92 (0.86) 3.00 (0.82)

34.12 (8.29) 39.64 (7.43) 36.36 (8.26) 35.24 (7.58) 37.20 (6.49) 36.08 (7.99) 38.12 (7.18) 37.36 (6.47)

49.08 (9.12) 55.56 (9.78) 51.16 (11.07) 53.52 (11.20) 53.00 (10.63) 51.52 (9.54) 52.40 (9.74) 54.12 (10.66)

FHHM, familiar high arousal-high pleasure music; FHLM, familiar high arousal-low pleasure music; FLLM, familiar low arousal-low pleasure music; FLHM, familiar low arousal-high pleasure music; UHHM, unfamiliar high arousal-high pleasure music; UHLM, unfamiliar high arousal-low pleasure music; ULLM, unfamiliar low arousal-low pleasure music; ULHM, unfamiliar low arousal-high pleasure music; Length, length of music training.

Table 2 Music characteristics and mean ratings of valence, arousal and familiarity for each excerpt (SDs in parentheses). Music

Composer

Excerpt

Valence

Arousal

Familiarity

Victory Mars, the Bringer of War Ase’s Death Tempo di Bolero moderato assai Acroyahsi Threnody for the Victims of Hiroshima Lachrimae Antiquae Nature’s Path

T. Huljic´ G. Holst E. Grieg M. Ravel Y. Chryssomallis K. Penderecki

0:03–4:38 3:53–8:27 0:00–4:35 0:06–4:41 3:46–8:21 0:02–2:30 3:30–5:37 0:00–4:35 0:11–4:46

6.70 (0.47) 1.65 (0.75) 1.30 (0.47) 5.75 (0.44) 6.45 (0.51) 1.05 (0.22)

6.30 (0.73) 6.05 (0.51) 1.50 (0.51) 1.85 (0.37) 6.10 (0.72) 6.60 (0.50)

5.85 (0.67) 5.70 (0.66) 5.65 (0.67) 5.75 (0.44) 2.05 (0.94) 1.15 (0.37)

2.35 (0.75) 6.10 (0.64)

1.75 (0.64) 1.45 (0.60)

1.50 (0.69) 1.25 (0.44)

J. Dowland D. May

Valence, arousal, music preference and familiarity were assessed with the following questions: “Rate how pleasant the music sounds” (from 1 = very unpleasant to 7 = very pleasant), “Rate how stimulating the music sounds” (from 1 = very calming to 7 = very stimulating), “Rate how much you like the music you just heard” (from 1 = strongly dislike to 7 = strongly like), and “Rate how much you are familiar with the music you just heard” (from 1 = very unfamiliar to 7 = very familiar) on a 7-point Likert scale.

and music preference ratings, but not with length of music training and familiarity rating. In addition, there were significant correlations between arousal and music preference ratings, and between valence and music preference ratings. Neither length of music training nor familiarity rating was correlated with music preference rating, in either the tension, or the anxiety rating. Because stress reduction efficacy does not appear to be affected by length of music training and familiarity, these two factors were not entered into the subsequent mediation and dominance analyses.

2.4. Procedure 3.2. Mediation analysis Participants were tested individually in a quiet classroom. First, they rated their levels of tension and completed the SAI. Then they performed a mental arithmetic test presented on paper by verbally reporting answers within 5 min. After completing the mental arithmetic test, they were asked to rate their tension levels and complete the SAI, before listening to a piece of music through AKG K512 MKII stereo headphones connected with a laptop computer (Acer Aspire 4720Z) at a desired loudness level. Finally, the levels of tension and state anxiety were immediately reobtained after listening to music. Meanwhile, participants also rated the levels of music valence, music arousal, music preference and familiarity. The experiment lasted approximately 25 min. 3. Results A correlation analysis was used to investigate which factors could predict stress reduction. Based on the correlation results, mediation analysis was used to reveal how the factors influence stress reduction, and dominance analysis was finally employed to determine the dominant factor in predicting stress reduction. 3.1. Correlation analysis Table 3 shows Spearman’s correlations for arousal, valence, length of music training, music preference, familiarity, tension and state anxiety after listening to music. As can be seen, both tension and anxiety ratings were associated with arousal, valence,

According to appraisal theory, emotions are evoked by evaluations of events (e.g., Lazarus, 1982; Moors, 2013; Roseman, 1991; Smith, Haynes, Lazarus, & Pope, 1993). In other words, the influence of a stimulus on emotion is mediated by cognitive appraisals such as preference and familiarity (Scherer, 2001), and thus music preference and familiarity may represent cognitive appraisals. However, familiarity was correlated neither with tension nor with state anxiety, as revealed by the correlation analyses. It was thus hypothesized that music preference may mediate the relationship between music and stress. The structural equation modeling (SEM) was used to test whether music preference partially (Model 1) or completely (Model 2) mediates valence and arousal’s effects on tension or state anxiety. Model fit was assessed with the chisquare (2 ) test and several additional fit indices (West, Taylor, & Wu, 2012), including the comparative fit index (CFI), Tucker-Lewis index (TLI), the standardized root-mean-square residual (SRMR), and the root-mean-square error of approximation (RMSEA) with its 90% confidence interval (CI). A good-fitting model is indicated by CFI and TLI values above .95, an RMSEA value below .06, an SRMR value smaller than .08, and a nonsignificant 2 test (Hu & Bentler, 1999; West et al., 2012). As illustrated in Table 4, the SEM revealed that Model 1 had a good fit to the data in tension and state anxiety. However, Model 2 did not fit the data in tension and state anxiety. Moreover, the indirect effects of valence (indirect effect = −.44, SE = .04, 95% CI = [−.52, −.35], p < .001) and arousal (indirect effect = .22, SE = .03,

J. Jiang et al. / The Arts in Psychotherapy 48 (2016) 62–68

65

Table 3 Correlations for the variables. Variable

1

1. Arousal 2. Valence 3. Length of music training 4. Music preference 5. Familiarity 6. Tension 7. State anxiety

–

**

.005 .002 −.292** .003 .467** .345**

2

3

4

5

6

7

– −.006 .662** .034 −.542** −.545**

– .010 −.029 −.008 −.056

– .082 −.780** −.668**

– −.041 −.093

– .759**

–

p < .01.

Table 4 Goodness-of-fit summaries for the two models. Model

Tension

1 2 a b

State anxiety



df

p

CFI

TLI

RMSEA

SRMR

2

df

pa

CFI

TLI

RMSEAb

SRMR

0.01 33.94

1 3

.937 <.001

1.00 0.93

1.02 0.85

0.00 [0.00, 0.06] 0.23 [0.16, 0.30]

0.00 0.07

0.01 23.22

1 3

.937 .001

1.00 0.94

1.02 0.88

0.00 [0.00, 0.06] 0.18 [0.12, 0.26]

0.00 0.06

2

a

b

Bollen–Stine bootstrap p value based on 5000 resamplings. 90% CI in brackets.

Fig. 1. SEM depicting the relationship between valence and arousal and tension (a) and state anxiety (b) with music preference as a partial mediator. ***p< .001.

66

J. Jiang et al. / The Arts in Psychotherapy 48 (2016) 62–68

Table 5 Dominance analysis of music preference (X1 ), arousal (X2 ) and valence (X3 ) in predicting tension and state anxiety. Subset model

Tension

State anxiety Additional contribution of

2

R Null (no predictors) Models with one predictor X1 X2 X3

.651 .242 .295

Models with two predictors X1 , X2 X1 , X3 X2 , X3

.700 .651 .534

Models with three predictors X1 , X2 , X3 Overall average

Additional contribution of

X1

X2

X3

.651

.242

.295

.049

.000 .292

.458 .356

.239 .006 .055

.172

.706

2

R

.510 .171 .301 .540 .517 .469

X1

X2

X3

.510

.171

.301

.030

.007 .298

.369 .216

.168 .024 .047

.095

.564 .410

.147

95% CI = [.16, .29], p < .001) on tension through music preference were significant. Likewise, the indirect effects of valence (indirect effect = −.32, SE = .05, 95% CI = [−.43, −.23], p < .001) and arousal (indirect effect = .17, SE = .03, 95% CI = [.11, .23], p < .001) on state anxiety through music preference were significant. Model 1 with standardized regression coefficients is depicted in Fig. 1. As can be seen, valence and arousal not only directly but also indirectly through music preference influenced tension (Fig. 1a) and state anxiety (Fig. 1b). The results suggest that the effects of valence and arousal on stress are partly mediated by music preference. 3.3. Dominance analysis Standardized regression coefficients may give an indication of the relative importance of each predictor in the structural equation modeling (SEM), with higher values for regression coefficients indicating more importance. As shown in Fig. 1, music preference might be more important than valence and arousal for reducing stress. In order to determine which factor plays the most important role in the stress reduction, dominance analysis was employed in the present study. Dominance analysis is considered to be a new and reliable method for comparing the relative importance of predictors (Azen, 2013; Azen & Budescu, 2003; Budescu, 1993), and has been widely applied in previous research (e.g., Kim, Petscher, Schatschneider, & Foorman, 2010; Suh, Diener, Oishi, & Triandis, 1998; Tighe & Schatschneider, 2014). Table 5 summarizes the results of dominance analysis for tension and state anxiety levels. The first column identifies the variables in each subset model. The columns labeled R2 represent the variance in tension or state anxiety levels explained by the model appearing in the corresponding row. Columns labeled Xi show the increase in R2 as a result of adding that particular predictor into the row model. In addition, the last row labeled overall average was obtained by averaging the values in the corresponding column above it. As can be shown, music preference explained 41.0% of the variance in tension level, while arousal and valence accounted for 14.7% and 14.9% of the variance, respectively. Regarding the state anxiety level, 29.9% of the variance was due to music preference, whereas 10.6% and 15.9% was attributed to arousal and valence. These results confirm that music preference is the most important predictor of stress level. 4. Discussion It has been suggested that stress reduction efficacy would be affected by many factors such as music arousal and valence, listeners’ music preference, familiarity, and music training. Previous research, however, has produced conflicting results. Based on this,

.149

.299

.106

.159

the purpose of the present study was to explore the mechanisms of music for reducing stress to determine how the factors influence stress reduction, and which factor is most salient. Our results indicated that music preference, arousal and valence did predict stress reduction, and the effects of valence and arousal on stress reduction were partially mediated by music preference. Music preference plays a critical role when music is used for stress reduction. These results have important implications for those wanting to use music to reduce stress. Our finding is consistent with previous evidence that music preference is more important than arousal (Jiang et al., 2013), and arousal is more important than valence (Sandstrom & Russo, 2010). This may be due to the fact that music preference is positively correlated with the intensity of felt happiness (Hunter et al., 2010; Kreutz et al., 2008) and peace (Kreutz et al., 2008). The more preference listeners have for the music, the more intense happiness or peace they feel. Nonetheless, Sandstrom and Russo (2010) did not find a significant correlation between liking music and stress levels. The conflicting findings may be attributed to difference of point of Likert scale the two studies used. A 4-point Likert scale was used for rating preference in the study of Sandstrom and Russo (2010), whereas a 7-point Likert scale was used in the present study. In Sandstrom and Russo’s (2010) study, the restricted rating range (4point scale) may have pointed to nonsignificance even when there was an actual correlation between the two variables (Kantowitz, Roediger, & Elmes, 2014; Tabachnick & Fidell, 2013). On the other hand, the relative importance of music preference on stress reduction could be also explained by appraisal theory. The theory claims that it is evaluations and interpretations of events, not events per se that determine one’s emotional responses (e.g., Lazarus, 1982; Moors, 2013; Roseman, 1991; Smith et al., 1993). If music is evaluated as preferred, then it could induce positive emotions and contribute to relaxation. On the contrary, if music is not liked, then it could cause negative emotions. This explanation has been supported by neuroscientific evidence that intense pleasure in response to preferred music is associated with dopamine activity in the mesolimbic reward system (Montag, Reuter, & Axmacher, 2011; Salimpoor, Benovoy, Larcher, Dagher, & Zatorre, 2011; Salimpoor et al., 2013). Our finding that arousal influences stress reduction, to a certain extent, is in line with past research that found low-arousal music has larger effect than high-arousal music in reducing stress (e.g., Gan et al., 2015; Iwanaga & Moroki, 1999; Jiang et al., 2013; Lingham & Theorell, 2009; Sandstrom & Russo, 2010). However, other studies did not find significant differences in relaxation levels induced by high-arousal and low-arousal music listening (Rohner & Miller, 1980; Stratton & Zalanowski, 1984). This discrepancy may result

J. Jiang et al. / The Arts in Psychotherapy 48 (2016) 62–68

from the difference in methodology between the studies. Unlike the attentive music listening in the present study, Rohner and Miller (1980) did not require participants to listen to the music. Rather, they told them that the music was to help pass the time. Compared with attentive music listening, passive music listening may have weaker effects on stress reduction. In the study by Stratton and Zalanowski (1984) mentioned above, both the low-arousal and high-arousal music were participant-preferred, whereas in the present study preferred and nonpreferred music was included. If there is no difference between high-arousal and low-arousal music in reducing stress when participants like the music (Jiang et al., 2013), then the effect of arousal should not have been observed by Stratton and Zalanowski (1984). Although Sandstrom and Russo (2010) reported that highpleasure music was marginally better than low-pleasure music in reducing state anxiety, the effect was not significant. This is different from our finding which indicated that participants feel less state anxiety when listening to music rated as pleasurable, than music that is not. One possible explanation is the difference in duration of music clips between the two studies. In our study, the duration of music clips was 4:35, while 2-min music clips were used in Sandstrom and Russo’s (2010) study. When the focus is on felt emotions, the induction of an emotional response and the subsequent self-reporting of the experienced emotion may require a longer duration (Eerola & Vuoskoski, 2013). It may be difficult to make a realistic evaluation within a relative short clip (Ritossa & Rickard, 2004). Another possible explanation relates to sample size. In the present study, there were 200 participants, and each group contained 50 listeners. However, Sandstrom and Russo (2010) included 63 participants, each group contained at most 22 listeners. This relatively small sample size may not have the power to detect the effect of valence. Familiarity ratings did not predict stress level in the present study, which is consistent with some previous studies (Chafin et al., 2004; Hatta & Nakamura, 1991), but not with others (Margounakis & Politis, 2012; Sung et al., 2012). The conflicting findings may be attributed to the relationship between familiarity and preference. It has been reported that familiarity is positively correlated with music preference (Rawlings & Leow, 2008; Ritossa & Rickard, 2004; Tan, Yowler, Super, & Fratianne, 2012). That is, familiar music is also preferred. The effect of familiarity reported in Sung et al.’s (2012) study may be due to the role of music preference, since the researcher reported the familiar music was also preferred music for participants. Likewise, the effect of familiarity in the study by Margounakis and Politis (2012) may be attributed to the confusion of familiarity and preference, although the researchers have not reported the degree of liking for music excerpts. Previous studies (Knight & Rickard, 2001; Laohawattanakun et al., 2011) have reported that music training had no significant effects on stress reduction. Our findings reinforce that even for listeners with music training, the length of music training does not affect the efficacy of music for stress reduction. It is, however, contradictory to the finding of Smith and Morris (1977), who demonstrated that the effect of low-arousal music on state anxiety reduction is greater for music majors than for nonmusic majors. The discrepancy may be due to the difference in music preference. Unlike the music major participants in the present study, music major listeners in the study of Smith and Morris (1977) liked low-arousal music more than nonmusic majors. It is therefore probable that the effect of music training is attributable to the effect of music preference. Indeed, music emotion studies also reveal music training background would not affect intensity of felt emotion for listeners (Harrer & Harrer, 1977; Kreutz et al., 2008). In conclusion, although valence and arousal influence psychological stress, these effects are partially mediated by music

67

preference. Our findings suggest that music preference plays a critical role on the potential for music to reduce stress, provide a deeper understanding of stress-reducing effects of music, and have important implications for individuals and clinicians who wish to use music to reduce stress.

Acknowledgement This research was supported by a grant from the National Natural Science Foundation of China (31470972) to C.J.

References Ali, S. O., & Peynircioglu, Z. F. (2006). Songs and emotions: Are lyrics and melodies equal partners? Psychology of Music, 34(4), 511–534. Azen, R. (2013). Using dominance analysis to estimate predictor importance in multiple. In Y. Petscher, C. Schatschneider, & D. L. Compton (Eds.), Applied quantitative analysis in education and the social sciences (pp. 34–64). New York, NY: Routledge. Azen, R., & Budescu, D. V. (2003). The dominance analysis approach for comparing predictors in multiple regression. Psychological Methods, 8(2), 129–148. Bradt, J., Dileo, C., & Shim, M. (2013). Music interventions for preoperative anxiety. Cochrane Database of Systematic Reviews, 6 http://dx.doi.org/10.1002/14651858. CD006908.pub2 Brattico, E., Alluri, V., Bogert, B., Jacobsen, T., Vartiainen, N., Nieminen, S., et al. (2011). A functional MRI study of happy and sad emotions in music with and without lyrics. Frontiers in Psychology, 2, 308. http://dx.doi.org/10.3389/fpsyg.2011. 00308 Budescu, D. V. (1993). Dominance analysis: A new approach to the problem of relative importance of predictors in multiple regression. Psychological Bulletin, 114(3), 542–551. Castro, S. L., & Lima, C. F. (2014). Age and musical expertise influence emotion recognition in music. Music Perception, 32(2), 125–142. Chafin, S., Roy, M., Gerin, W., & Christenfeld, N. (2004). Music can facilitate blood pressure recovery from stress. British Journal of Health Psychology, 9(3), 393–403. Clark, M., Isaacks-Downton, G., Wells, N., Redlin-Frazier, S., Eck, C., Hepworth, J. T., et al. (2006). Use of preferred music to reduce emotional distress and symptom activity during radiation therapy. Journal of Music Therapy, 43(3), 247–265. Davis, W. B., & Thaut, M. H. (1989). The influence of preferred relaxing music on measures of state anxiety, relaxation, and physiological responses. Journal of Music Therapy, 26(4), 168–187. Eerola, T., & Vuoskoski, J. K. (2011). A comparison of the discrete and dimensional models of emotion in music. Psychology of Music, 39(1), 18–49. Eerola, T., & Vuoskoski, J. K. (2013). A review of music and emotion studies: Approaches, emotion models, and stimuli. Music Perception, 30(3), 307–340. Fisher, S., & Greenberg, R. P. (1972). Selective effects upon women of exciting and calm music. Perceptual and Motor Skills, 34(3), 987–990. Gan, S. K.-E., Lim, K. M.-J., & Haw, Y.-X. (2015). The relaxation effects of stimulative and sedative music on mathematics anxiety: A perception to physiology model. Psychology of Music, http://dx.doi.org/10.1177/0305735615590430 Harrer, G., & Harrer, H. (1977). Music, emotion and autonomic function. In M. Critchley, & R. A. Henson (Eds.), Music and the brain (pp. 202–216). London, England: Butterworth-Heinemann. Hatta, T., & Nakamura, M. (1991). Can antistress music tapes reduce mental stress? Stress Medicine, 7(3), 181–184. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. Hunter, P. G., Schellenberg, E. G., & Schimmack, U. (2010). Feelings and perceptions of happiness and sadness induced by music: Similarities, differences, and mixed emotions. Psychology of Aesthetics, Creativity, and the Arts, 4(1), 47–56. Iwanaga, M., Ikeda, M., & Iwaki, T. (1996). The effects of repetitive exposure to music on subjective and physiological responses. Journal of Music Therapy, 33(3), 219–230. Iwanaga, M., Kobayashi, A., & Kawasaki, C. (2005). Heart rate variability with repetitive exposure to music. Biological Psychology, 70(1), 61–66. Iwanaga, M., & Moroki, Y. (1999). Subjective and physiological responses to music stimuli controlled over activity and preference. Journal of Music Therapy, 36(1), 26–38. Jeong, H. C. (2008). The effect of music therapy on the physiological and psychological status of women college students based on their preference of music. Journal of Korean Academy of Adult Nursing, 20(2), 321–330. Jiang, J., Zhou, L., Rickson, D., & Jiang, C. (2013). The effects of sedative and stimulative music on stress reduction depend on music preference. The Arts in Psychotherapy, 40(2), 201–205. Kantowitz, B. H., Roediger, H. L., & Elmes, D. G. (2014). Experimental psychology (10th ed.). Stamford, CT: Cengage Learning.

68

J. Jiang et al. / The Arts in Psychotherapy 48 (2016) 62–68

Kim, Y.-S., Petscher, Y., Schatschneider, C., & Foorman, B. (2010). Does growth rate in oral reading fluency matter in predicting reading comprehension achievement? Journal of Educational Psychology, 102(3), 652–667. Knight, W. E. J., & Rickard, N. S. (2001). Relaxing music prevents stress-induced increases in subjective anxiety, systolic blood pressure, and heart rate in healthy males and females. Journal of Music Therapy, 38(4), 254–272. 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. Laohawattanakun, J., Chearskul, S., Dumrongphol, H., Jutapakdeegul, N., Yensukjai, J., Khumphan, N., et al. (2011). Influence of music training on academic examination-induced stress in Thai adolescents. Neuroscience Letters, 487(3), 310–312. Lazarus, R. S. (1982). Thoughts on the relations between emotion and cognition. American Psychologist, 37(9), 1019–1024. Lesiuk, T. (2008). The effect of preferred music listening on stress levels of air traffic controllers. The Arts in Psychotherapy, 35(1), 1–10. Lima, C. F., & Castro, S. L. (2011). Emotion recognition in music changes across the adult life span. Cognition and Emotion, 25(4), 585–598. Lingham, J., & Theorell, T. (2009). Self-selected “favourite” stimulative and sedative music listening – How does familiar and preferred music listening affect the body? Nordic Journal of Music Therapy, 18(2), 150–166. Linnemann, A., Ditzen, B., Strahler, J., Doerr, J. M., & Nater, U. M. (2015). Music listening as a means of stress reduction in daily life. Psychoneuroendocrinology, 60, 82–90. Margounakis, D., & Politis, D. (2012). Exploring the relations between chromaticism, familiarity, scales and emotional responses in music. In Paper presented at the 19th Colloquium on Music Informatics. Montag, C., Reuter, M., & Axmacher, N. (2011). How one’s favorite song activates the reward circuitry of the brain: Personality matters!. Behavioural Brain Research, 225(2), 511–514. Moors, A. (2013). On the causal role of appraisal in emotion. Emotion Review, 5(2), 132–140. Rawlings, D., & Leow, S. H. (2008). Investigating the role of psychoticism and sensation seeking in predicting emotional reactions to music. Psychology of Music, 36(3), 269–287. Ritossa, D. A., & Rickard, N. S. (2004). The relative utility of ‘pleasantness’ and ‘liking’ dimensions in predicting the emotions expressed by music. Psychology of Music, 32(1), 5–22. Robb, S. L., Nichols, R. J., Rutan, R. L., Bishop, B. L., & Parker, J. C. (1995). The effects of music assisted relaxation on preoperative anxiety. Journal of Music Therapy, 32(1), 2–21. Rohner, S. J., & Miller, R. (1980). Degrees of familiar and affective music and their effects on state anxiety. Journal of Music Therapy, 17(1), 2–15. Roseman, I. J. (1991). Appraisal determinants of discrete emotions. Cognition and Emotion, 5(3), 161–200. Rosenow, S. C., & Silverman, M. J. (2014). Effects of single session music therapy on hospitalized patients recovering from a bone marrow transplant: Two studies. The Arts in Psychotherapy, 41(1), 65–70. Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39(6), 1161–1178. Salimpoor, V. N., Benovoy, M., Larcher, K., Dagher, A., & Zatorre, R. J. (2011). Anatomically distinct dopamine release during anticipation and experience of peak emotion to music. Nature Neuroscience, 14(2), 257–262. Salimpoor, V. N., van den Bosch, I., Kovacevic, N., McIntosh, A. R., Dagher, A., & Zatorre, R. J. (2013). Interactions between the nucleus accumbens and auditory cortices predict music reward value. Science, 340(6129), 216–219.

Sandstrom, G. M., & Russo, F. A. (2010). Music hath charms: The effects of valence and arousal on recovery following an acute stressor. Music and Medicine, 2(3), 137–143. Scherer, K. R. (2001). Appraisal considered as a process of multilevel sequential checking. In K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal processes in emotion: Theory, methods, research (pp. 92–120). New York, NY: Oxford University Press. Schubert, E. (1999). Measuring emotion continuously: Validity and reliability of the two-dimensional emotion-space. Australian Journal of Psychology, 51(3), 154–165. Smith, C. A., Haynes, K. N., Lazarus, R. S., & Pope, L. K. (1993). In search of the “hot” cognitions: Attributions, appraisals, and their relation to emotion. Journal of Personality and Social Psychology, 65(5), 916–929. Smith, C. A., & Morris, L. W. (1977). Differential effects of stimulative and sedative music on anxiety, concentration and performance. Psychological Reports, 41(3f), 1047–1053. Spielberger, C. D. (1983). Manual for the state-trait anxiety inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D., Sydeman, S. J., Owen, A. E., & Marsh, B. J. (1999). Measuring anxiety and anger with the state-trait anxiety inventory (STAI) and the state-trait anger expression inventory (STAXI). In M. E. Maurish (Ed.), The use of psychological testing for treatment planning and outcome assessment (2nd ed., pp. 993–1021). Mahwah, NJ: Lawrence Erlbaum Associates. Stratton, V. N., & Zalanowski, A. H. (1984). The relationship between music, degree of liking, and self-reported relaxation. Journal of Music Therapy, 21(4), 184–192. Stratton, V. N., & Zalanowski, A. H. (1994). Affective impact of music vs. lyrics. Empirical Studies of the Arts, 12(2), 173–184. Suh, E., Diener, E., Oishi, S., & Triandis, H. C. (1998). The shifting basis of life satisfaction judgments across cultures: Emotions versus norms. Journal of Personality and Social Psychology, 74(2), 482–493. Sung, H., Lee, W., Li, T., & Watson, R. (2012). A group music intervention using percussion instruments with familiar music to reduce anxiety and agitation of institutionalized older adults with dementia. International Journal of Geriatric Psychiatry, 27(6), 621–627. Tabachnick, B. G., & Fidell, L. S. (2013). Using multivariate statistics (6th ed.). Boston, MA: Pearson Education. Tan, X., Yowler, C. J., Super, D. M., & Fratianne, R. B. (2012). The interplay of preference, familiarity and psychophysical properties in defining relaxation music. Journal of Music Therapy, 49(2), 150–179. Thoma, M. V., Zemp, M., Kreienbühl, L., Hofer, D., Schmidlin, P. R., Attin, T., et al. (2015). Effects of music listening on pre-treatment anxiety and stress levels in a dental hygiene recall population. International Journal of Behavioral Medicine, 22(4), 498–505. Tighe, E. L., & Schatschneider, C. (2014). A dominance analysis approach to determining predictor importance in third, seventh, and tenth grade reading comprehension skills. Reading and Writing, 27(1), 101–127. Vieillard, S., Peretz, I., Gosselin, N., Khalfa, S., Gagnon, L., & Bouchard, B. (2008). Happy, sad, scary and peaceful musical excerpts for research on emotions. Cognition and Emotion, 22(4), 720–752. Wang, W. (2014). A study of the type and characteristics of relaxing music for college students. In Paper presented at the 167th Acoustical Society of America Meeting. West, S. G., Taylor, A. B., & Wu, W. (2012). Model fit and model selection in structural equation modeling. In R. H. Hoyle (Ed.), Handbook of structural equation modeling (pp. 209–231). New York, NY: Guilford Press. Yehuda, N. (2011). Music and stress. Journal of Adult Development, 18(2), 85–94.