Reasons for personal music listening: A mobile experience sampling study of emotional outcomes

Abstract

An important aspect of researching everyday music use is determining the reasons people have for listening to music. While this has been the focus of an extensive body of research, findings have been inconsistent, and the frequencies and affective outcomes of different reasons for listening remain unclear. Emotional reasons for listening are of particular interest, as these have been consistently shown to be of central importance to everyday music use. The current study aimed to provide empirical data to clarify the frequencies of reasons for listening, and their affective outcomes, by using the experience sampling method (ESM). Participants (N = 327; mean age 21.02 years, SD = 6.18) used the MuPsych app, a mobile ESM designed for the real-time and ecologically valid measurement of personal music listening. Results revealed that emotional reasons were most frequently used only when the listener was in a negative mood. Listening to cope with a situation or forget problems was associated with negative affective states and poor emotional health and well-being. It was concluded that personal music listening is utilised to fulfil specific emotional needs, which are determined by initial mood, and influenced by emotional health.

Keywords

emotion regulation experience sampling methodology individual differences listening reasons personal music listening smartphone

An important component of understanding how music is used in everyday life is clarifying the various reasons for listening to music. The reasons people have for music listening directly relate to the role music plays in their lives, and how they use it to fulfil particular needs. This article will give an overview of the various methodological approaches to assessing reasons for listening, and address the limitations of these with an alternative method, developed specifically for personal music listening on mobile devices.

The majority of previous research has relied on questionnaires, requiring participants to subjectively and retrospectively report their various reasons for listening. This method has revealed that music is most frequently used to improve mood, to relax or “chill out”, and to express, release, and influence emotions (Juslin & Laukka, 2004; Schafer & Sedlmeier, 2009). Studies have often focused on a particular age group, with reasons for listening assessed separately for adolescent (North, Hargreaves, & O’Neill, 2000; Tarrant, North, & Hargreaves, 2000), university student (Chamorro-Premuzic & Furnham, 2007), and elderly populations (Laukka, 2007). Questionnaires have also been used to determine the reasons for listening within specific environments. For example, while in the workplace, music is used most frequently to improve mood, relax, and increase happiness (Haake, 2011). While travelling, music is used for enjoyment, to pass time, to create or accentuate emotions, and to fade out external noises (Heye & Lamont, 2010). Along with quantitative questionnaires, reasons for listening have also been assessed through the use of qualitative interviews. One such study revealed seven main themes in the functions of music, relating to background, memories, diversion, emotion, self-regulation, reflection, and social bonding (Boer & Fischer, 2012).

Several reviews have attempted to collate these varied results on reasons for listening and music functions. For example, Laiho’s (2004) review of the literature identified four primary functions of music for adolescent listeners: those related to emotions, identity, interpersonal relationships, and agency. The most comprehensive review of the literature was performed by Schafer, Sedlmeier, Stadtler, and Huron (2013), incorporating over 500 proposed musical functions, from both theoretical and empirical studies. This extensive body of research was shown to converge on four major dimensions: social functions (such as those relating to identity or personality), emotional functions (such as mood enhancement), cognitive or self-related functions (such as escapism), and arousal-related functions (such as stress reduction: Schafer et al., 2013). This major review of music uses and functions produced a refined list of 129 distinct items, which were then analysed empirically through factor analysis. This involved 834 participants rating these items as reasons they would listen to music (from 0 “not at all” to 6 “fully agree”). Principal component analysis revealed three distinct dimensions: arousal and mood regulation (M = 3.78), self-awareness (M = 3.59), and social relatedness (M = 2.01).

Several additional factor analyses have been utilised to investigate the various uses of music within different demographics. One such factor analysis focused on the reasons adolescents have for listening to music, and produced three factors: fulfilling emotional needs, creating external impression, and enjoyment (North et al., 2000). A similar analysis on adolescent listeners revealed three main factors of fulfilling emotional needs, fulfilling social needs, and self-actualisation (Tarrant et al., 2000). For university undergraduates, the most important factors were (in order of importance): positive mood management, diversion, negative mood management, interpersonal relationships, personal identity, and surveillance (Lonsdale & North, 2011). Another analysis of this demographic, however, found the underlying factors of music use to be emotional, cognitive, and background (Chamorro-Premuzic & Furnham, 2007). In an elderly population, analysis has shown enjoyment to be the most prominent factor in music use, followed by relaxation and company, mood regulation, then identity and agency (Laukka, 2007). It is clear from this overview that previous studies have produced a broad range of reasons for listening, across various methodological approaches and demographic groups, with several inconsistencies that remain to be clarified through empirical analysis.

Despite the apparent inconsistencies in previous research, one prominent observation – regardless of methodology – is the importance of emotional reasons for listening. Literature reviews and factor analyses investigating functions of music have consistently determined emotional functions to be the most important, while social and other functions tend to be of lesser importance (Laiho, 2004; Lonsdale & North, 2011; Schafer et al., 2013; Tarrant et al., 2000). In particular, research has repeatedly shown emotion regulation functions to be crucial, with regulation described as the most prominent function of music known from research on Western populations (Juslin & Laukka, 2004; Lonsdale & North, 2011; Saarikallio & Erkkila, 2007; Schafer & Sedlmeier, 2009; Schafer, Tipandjan, & Sedlmeier, 2012). Responses to questionnaires have indicated that 93% of people use music to change their mood, with 49% doing so often (Juslin & Laukka, 2004). Emotional regulation functions include reaching a desired mood, or leaving an undesired one, reducing stress, increasing energy, venting strong emotions, and lifting spirits (Denora, 1999; Wells & Hakanen, 1991). In both adolescent and adult populations, the main goals for regulation are mood control and mood improvement, and these are reached using at least seven different regulatory strategies (Saarikallio, 2011; Saarikallio & Erkkila, 2007). These regulatory functions of music play an important role in everyday listening, and can be used successfully to change emotional states (van Goethem & Sloboda, 2011).

A consistent finding emerging from the literature is that emotional reasons for listening are used more frequently by certain demographic groups. One clear finding is that adolescents report emotional reasons more often, and emotion regulation appears to be of heightened importance to this age group (e.g., Laiho, 2004; Lonsdale & North, 2011). A similar differentiation is evident by gender, with female listeners using music more to fulfil emotional needs, and to manipulate or enhance certain moods (Boer et al., 2012; North et al., 2000; Sloboda, O’Neill, & Ivaldi, 2001; Wells & Hakanen, 1991). Emotional reasons also appear to be utilised more often by those low in emotional stability (Chamorro-Premuzic & Furnham, 2007). Furthermore, direct negative associations have been observed between various emotional reasons for listening and psychological functioning (Thoma, Scholz, Ehlert, & Nater, 2012). Many of these demographic findings with regard to the frequencies of emotional reasons remain to be confirmed through empirical and ecologically valid data collection.

While the existing literature on reasons for music use is extensive, an overwhelming majority of this research relies on retrospective reporting (whether by questionnaires or interviews). Retrospective reports introduce recall biases, which reflect reconstructive memory of unique or personally relevant events (Sloboda et al., 2001). Furthermore, these reports are influenced by commonly held beliefs of situational appropriateness or experimenter demand (Gibson, Aust, & Zillmann, 2000). This style of measurement does not take into account the social context of listening, which is a critical aspect in determining how people engage with music (Juslin, Liljeström, Västfjäll, Barradas, & Silva, 2008). As these previous studies have relied on a single report from each participant, they are unable to capture how frequently participants utilise different reasons for listening, and under which circumstances.

The experience sampling method (ESM: Csikszentmihalyi & LeFevre, 1989) offers an approach with minimal reliance on retrospective reporting. This method captures current subjective experience, through the use of repeated questionnaires across natural everyday functioning. Importantly, this allows for a more ecologically valid assessment of how reasons for listening are used within different social contexts. ESM studies of everyday music use typically allow listeners to select their reason or reasons for listening from a set list. Those that have allowed the selection of multiple reasons have revealed enjoyment to be the most frequently reported, with passing time and creating an atmosphere also common (Greasley & Lamont, 2011; North, Hargreaves, & Hargreaves, 2004). However, these studies have differed in the reasons available for selection, producing notably different sets of results. For example, Greasley and Lamont (2011) found relaxation to be the second most common function, reported in 42% of music episodes, while this option was not included in the list presented by North et al. (2004). Similarly, while habit was the third most common function reported by North et al. (2004), this was not an option in the more recent study. In a third ESM study, in which participants were forced to select a single reason from a list, the most common listening motive was to get some company, followed by music chosen by others, then relax, to get energised, and to pass the time (Juslin et al., 2008). The clear differences in these three sets of results may be due to the limited number of possible reasons, with each of these studies presenting a list of 15 reasons or less (Greasley & Lamont, 2011; Juslin et al., 2008; North et al., 2004). In contrast, when an ESM study allowed for open-ended responses, the most frequently reported reasons were nostalgic in nature, followed by reasons for mood regulation (Sloboda et al., 2001). This demonstrates that even with the temporal and ecological advantages of the ESM, inconsistencies are still apparent, and valid study design is essential in ensuring the accurate measurement of reasons for listening.

To investigate reasons for listening, the current study employed mobile ESM (m-ESM), as developed and tested by Randall and Rickard (2013). The m-ESM utilises event-based experience sampling, capturing the real-time state of the listener during music episodes. This real-time measurement effectively eliminates recall bias, and has major temporal advantages over previous ESM studies, which have relied on random sampling and some degree of retrospective reporting. Unlike these previous studies, the m-ESM observes only personal music listening, defined as listening to music on personal mobile devices, which is rapidly growing as a crucial component of everyday music use. This style of music engagement gives listeners complete control over what they listen to in their given social context, allowing them to fulfil their specific needs (Hargreaves & North, 1999). Mobile devices offer listeners the unprecedented ability to control and adjust their mood throughout their day, as they move between different environments (Bull, 2005). As personal music listening allows listeners complete control over why they listen to their music, it is ideal for observing reasons for listening and the associated emotional outcomes.

The current study sought to utilise the m-ESM to clarify inconsistencies in the literature on the relative frequencies of reasons for listening. The first aim was therefore to determine the frequencies of reasons for personal music listening, and explore how emotional reasons relate to demographic and individual variables. It was hypothesised that emotional reasons for listening would be used more frequently than non-emotional reasons. Furthermore, it was hypothesised that emotional reasons for listening would be used more frequently by female, adolescent, and emotionally unstable listeners.

The current study also sought to extend research in this field to include affective outcomes of reasons for listening, with a particular focus on emotional reasons. No study to date has provided empirical links between reasons for personal music listening and affective outcomes. As this listening style is becoming increasingly popular, it is essential to understand when these reasons produce hedonic benefit, and when they might have undesirable emotional outcomes. The second aim was therefore to determine the changes in emotional valence and arousal associated with emotional reasons for personal music listening. As this aim was exploratory, no explicit hypotheses were set.

Method

Participants

Participants were 594 users of MuPsych, who downloaded and used the app for the two-week collection period. Those with fewer than five completed music experience reports were excluded from analysis, producing a final set of 327 participants (249 females, 78 males; aged 13–51 years, M = 21.02 years; SD = 6.18). Participants varied on level of musical training (9.79% none; 20.18% primary school; 63.61% secondary school; 6.42% tertiary) and musical instrument playing (22.94% none; 30.28% 1–3 years; 14.07% 4–6 years; 32.36% > 6 years). Participants were recruited in a variety of ways, including promotion through online and printed articles, and through social media. Others were sourced from the Undergraduate Psychology Student Participant Pool at Monash University, with these students receiving course credit for first year psychology courses. No participant received any monetary incentive for participation.

Materials

All elements of the study were presented to participants through the MuPsych app, which was developed and tested by Randall and Rickard (2013). MuPsych is a smartphone app that collects data through event-based experience sampling reports (ESRs), with music ESRs automatically administered when participants listen to music on their own device. The app also collects control data through non-music ESRs, which are presented at times of no music listening. Finally, a series of psychological surveys are made available to participants at set times over the collection period, but these can be completed at any time of convenience.

Music ESRs assess the affective state of the listener at the moment they begin playing music, and then again after three minutes, if music is still playing at this time. This short listening period was selected primarily to ensure that the change in affective state was due to the particular piece of music the participant selected for their specific reason for listening, within their listening context (Randall & Rickard, 2013). Music ESRs measure affective state using two 7-point slider scales of current mood (valence) and energy level (arousal). These two dimensions have been demonstrated to be efficient and reliable measures of music-induced emotion, explaining a high proportion of variance (Thoma, Ryf, Mohiyeddini, Ehlert, & Nater, 2012; Vuoskoski & Eerola, 2011).

Immediately following the three-minute assessment of affective state, music ESRs present a series of additional screens, measuring variables such as the context of the listener, and their subjective ratings of the music. Included within this series of screens are two list-format screens to assess the primary reason for listening. These are presented in a branching format, with the first screen requiring a selection from three categories: “Emotional reason”, “Social reason”, or “Other reason”. Upon selection, participants are then presented with one of three corresponding lists of reasons for listening, from which they select a primary reason. This branching format was utilised to reduce the number of items on any one screen, in order to improve answering efficiency, and minimise both response time and recall bias. The 33 primary reasons for listening were compiled from the four relevant ESM studies of everyday music use, and cross-checked against a comprehensive review of listening reasons, to ensure all relevant items were included (Greasley & Lamont, 2011; Juslin et al., 2008; North et al., 2004, Schafer et al., 2013, Sloboda et al., 2001). All 33 reasons can be seen in Table 1, with the text presented to listeners shown in the “Reason” column, and the labels used in analyses shown in the “Label” column. All ESR lists present items in a randomised order, and include a help icon, to provide users with explanatory text when required.¹

Table 1.

The relative frequencies of all reasons for listening across initial level of valence.

Reason	Label	Frequency
		Initial valence
		All (%)	Negative (%)	Positive (%)
Emotional reasons (total)		31.82	59.36	26.02
To relax/de-stress	Relax	6.64	9.24	6.08
For enjoyment	Enjoy	4.04	1.69	5.03
To raise energy/get pumped up	Raise	3.11	1.84	3.40
To improve my mood	Improve	3.00	7.65	1.89
To reflect on my emotions	Reflect	2.78	7.96	1.59
To cope with a situation	Cope	2.51	7.98	1.38
To maintain/enhance a mood	Maintain	2.48	2.00	3.02
To release/vent emotion	Release	2.15	6.68	0.69
To forget my problems	Forget	2.10	6.32	1.04
To express my feelings	Express	1.22	3.23	0.75
To feel strong emotions	Strong	0.91	3.52	0.55
Social reasons (total)		10.02	5.75	11.03
To entertain	Entertain	2.82	0.64	3.55
To create an atmosphere	Atmosphere	2.34	0.74	3.08
To block out other people	Block	1.44	1.10	1.13
To reduce loneliness	Loneliness	1.10	1.63	0.83
To dance	Dance	0.74	0.09	1.05
To socialise	Socialise	0.52	0.26	0.51
To feel in control	Control	0.52	0.85	0.38
To assert personality	Personality	0.24	0.28	0.27
To express my identity/values	Express	0.18	0.00	0.15
To build self-esteem	Self-esteem	0.13	0.16	0.07
Other reasons (total)		58.17	34.89	62.95
For background music	Background	19.23	11.08	21.95
For entertainment	Entertainment	13.08	4.62	15.81
To reduce boredom	Boredom	8.05	4.06	8.17
Out of habit	Habit	4.24	4.12	4.07
To help concentrate	Concentrate	3.53	1.71	3.80
For a distraction	Distraction	3.12	4.61	2.11
To appreciate the music as art	Appreciate	1.98	1.05	2.45
To be creative/to imagine	Creative	1.57	0.59	1.84
To bring back memories	Memories	1.47	1.69	1.16
To understand my thoughts	Understand	1.00	2.01	0.66
To analyse the music	Analyse	0.96	0.00	0.89
To identify with the lyrics/artist	Identify	0.83	0.63	0.66

From the battery of psychological surveys presented through MuPsych, the current study utilised measures of emotional health, well-being, and emotional stability. Emotional health indicators were assessed using the Depression Anxiety and Stress Scale (DASS), with three 7-item subscales of depression, anxiety, and stress (Lovibond & Lovibond, 1995). Well-being was assessed using the 5-item Satisfaction With Life Scale (SWLS: Diener, Emmons, Larsen, & Griffin, 1985), and the Positive and Negative Affect Schedule (PANAS), with 14-item subscales of Positive and Negative affect (Watson, Clark, & Tellegen, 1988). An additional 8 items were added to the PANAS-20 in response to criticisms that low arousal affect items are not duly represented in the original questionnaire. Emotional stability was assessed using a subscale of the International Personality Item Pool (IPIP: Goldberg, 1992). Finally, the Music Use questionnaire (MUSE: Chin & Rickard, 2012) was used to assess musical training and instrument playing. Previous assessment has determined that surveys presented in MuPsych produce similar Cronbach’s alpha scores to those published from the standard questionnaires (Randall & Rickard, 2013).

Procedure

Participants were instructed to download the MuPsych app to their own mobile device (iPhone or iPod touch)² and use it as their personal music player for a two-week period. All elements of the study were presented electronically through the mobile interface, including consent and demographics forms, music and non-music ESRs, and all psychological surveys. Music ESRs were presented at times when participants selected to listen to music, and in accordance with a set of timing rules, including a three-hour gap between completed ESRs, and a maximum of two per day. These rules were refined through pilot testing to ensure ecological measurement of music experiences with minimal intrusion (Randall & Rickard, 2013). Non-music ESRs were presented at times relative to the completion of music ESRs (approximately 24 hours later). These timing rules were selected to reflect listening patterns and waking times of participants more accurately than arbitrary time windows (Randall & Rickard, 2013). Participants were also required to complete a series of psychological surveys, the availability of which were staggered over the two-week collection period (four from the first day, and four more after a week). The timing of survey presentation was determined from pilot testing feedback, and was to avoid overwhelming participants with too many surveys to complete at a single time.

Data analyses

All analyses were performed on the participant level, by creating aggregate scores for each individual, rather than analysing on an ESR level. This approach is recommended for ESM (Hektner, Schmidt, & Csikszentmihalyi, 2007), and has been utilised in previous ESM studies of music use (Juslin et al., 2008), as data points at the ESR level are not independent.

For the first aim, the relative frequency of each reason for listening was averaged at the participant level, producing relative percentages for each individual. These scores were then converted to arcsin values, to allow for normal distribution of data. Using these scores, paired samples t-tests were performed to determine differences in frequencies between emotional reasons and other reasons. In addition, one-way ANOVAs were used to determine any differences in frequency scores between emotional reasons for listening. Using two-tailed Pearson correlations (with an alpha value of .05), arcsin scores were then correlated with individual scores of emotional stability, emotional health, and well-being. To test specific hypotheses, independent samples t-tests (with an alpha value of .05) were used to determine differences in these scores between both females and males, and adolescents and adults.

For the second aim, aggregate valence and arousal scores for each reason used were created at the participant level. For analysis purposes, these scores were centred to zero, with a range of −3 to +3. To assess change in valence and arousal for each reason, related-sample t-tests (with Bonferroni adjusted alpha values of .0045 to control the experimentwise error rate) were performed on these aggregated scores. Specifically, the valence and arousal levels of each reason at the time of music onset were compared to their corresponding level after three minutes of music listening, to establish any significant change over time. A similar process was used to determine valence change for episodes with either negative or positive initial valence, and arousal change for episodes with either negative or positive initial arousal. Due to the low frequencies of individual social reasons for listening, and the subsequent lack of data under various initial mood conditions, only social reasons overall were included in analyses of affective change. Assumptions were met for all analyses, except for the normality of difference scores assumption for the non-music condition, due to the high frequency of zero valence change.

Results

Aim 1. To determine the frequencies of reasons for personal music listening, and explore how emotional reasons relate to individual variables.

Reason frequencies

The relative frequencies of all reasons for listening are displayed in Table 1. When considering all initial moods, listeners utilised music for Emotional reasons for 32% of music episodes, Social reasons for 10%, and Other reasons for 58%. Comparison of arcsin values revealed that Other reasons (M = 51.13, SD = 21.00) were used significantly more frequently than Emotional reasons (M = 31.64, SD = 19.28), t(326) = 9.10, p < .001, d = 0.50. The three most frequently used reasons (in order: Background, Entertainment, and Boredom) were from the Other reasons category, and together accounted for 40% of all music episodes. Within Emotional reasons, a one-way ANOVA, F(10) = 23.15, p < .001, showed that Relax was used significantly more often than any other emotional reason, while Express and Strong were used significantly less often than all others.

Table 1 also displays frequencies of reasons for music episodes in which the listener was in either a negative or positive initial mood. Reason frequencies for a negative initial mood differed noticeably from those for all music episodes, with listeners utilising Emotional reasons for 59% of music episodes, Social reasons for 6%, and Other reasons for 35%. Emotional reasons (M = 51.96, SD = 25.06) were used significantly more frequently than Other reasons (M = 33.02, SD = 25.88), t(116) = 4.11, p < .001, d = 0.38. While Background remained the most frequently used reason, the next six most frequently used reasons were all emotional. Within Emotional reasons, a one-way ANOVA, F(10) = 7.36, p < .001, showed that the top five reasons (Relax, Cope, Reflect, Improve, and Release) were used significantly more than the three lowest (Maintain, Raise, and Enjoy).

When in a positive initial mood, listeners utilised Emotional reasons for 26% of music episodes, Social reasons for 11%, and Other reasons for 63%. The order of most frequently used reasons was similar to that for music episodes overall, with the same three top reasons (Background, Entertainment, and Boredom) accounting for 46% of episodes. From the descriptive frequencies displayed in Table 1, it was determined that the top seven reasons used more often in a negative mood, as compared to a positive mood, were all emotional (in order: Release, Strong, Forget, Cope, Reflect, Express, and Improve).

Individual differences in emotional reason frequencies

No correlation was revealed between age (in years) and frequency of emotional reasons overall, with only a small negative correlation between age and Forget, r(321) = −.12, p = .034. To test the specific age hypothesis, a comparison was made between adolescents (mean age = 16.23 years, SD = 1.22) and adults (mean age = 23.19 years, SD = 6.30). No significant difference was found between the relative frequency of any emotional reasons used by adolescents (M = 31.31, SD = 18.84) and adults (M = 31.69, SD = 19.39), t(319) = 0.17, p = .869.

In overall use of emotional reasons, no significant difference was found between female (M = 32.22, SD = 19.01) and male (M = 29.34, SD = 19.80) listeners, t(319) = 1.12, p = .263. Upon closer analysis, Raise was the only specific emotional reason to show a gender difference, with males (M = 6.76, SD = 10.43) using it more frequently than females (M = 4.31, SD = 8.76), t(319) = 2.00, p = .046, d = 0.25. Emotional stability was found to be associated with less frequent use of emotional reasons, r(266) = −.13, p = .030, in particular Forget, r(266) = −.27, p < .001, Cope, r(266) = −.23, p < .001, and Strong, r(266) = −.16, p = .010.

In further analyses, several associations were observed between frequencies of emotional reasons for listening and measures of emotional health and well-being. Emotional reason frequency overall was positively correlated with levels of depression, r(193) = .23, p = .001, and anxiety, r(193) = .16, p = .024. In particular, the emotional reasons Cope and Forget correlated positively with depression, r(193) = .24, p = .001, r(193) = .35, p < .001, respectively, and anxiety, (r(193) = .15, p = .033, r(193) = .20, p = .005, and negatively with satisfaction with life, r(324) = −.21, p < .001, r(324) = −.20, p < .001, and positive affect, r(323) = −.22, p < .001, r(323) = −.15, p = .005. In addition, Forget was positively associated with negative affect, r(323) = .31, p < .001.

Aim 2. To determine the changes in emotional valence and arousal associated with reasons for personal music listening.

Valence change

Table 2 shows the initial valence and change in valence for Emotional and Other reasons for music listening. The emotional reasons Release, Cope, and Forget were the only ones which occurred with negative mean initial valence scores, while those of Enjoy, Maintain, and Raise recorded some of the highest initial valence scores.

Table 2.

Changes in valence and arousal for Emotional and Other reasons for listening.

Reason	Initial	Valence change			Initial	Arousal change
	Valence	Initial valence			Arousal	Initial arousal
	Mean	All	Negative	Positive	Mean	All	Negative	Positive
Music (total)	0.91	0.08**	0.43**	−0.04*	−0.14	0.09**	0.35**	−0.25**
Non-music (total)	0.93	0.02	0.34*	−0.05	−0.31	0.02	0.25**	−0.33**
Emotional (total)	0.40	0.05	0.33**	−0.13*	−0.38	0.10*	0.32**	−0.31**
Relax	0.60	0.02	0.42*	−0.23*	−0.68	0.03	0.39**	−0.73**
Enjoy	1.40	0.07	0.19	0.05	0.19	0.10	0.25*	−0.10
Raise	1.24	0.33**	1.08*	0.29**	0.09	0.60**	1.11**	0.22
Improve	0.14	0.19*	0.37*	−0.14	−0.23	0.11	0.46**	−0.35
Reflect	0.10	−0.10	0.24*	−0.49*	−0.81	0.08	0.27*	−0.48*
Cope	−0.50	−0.30	0.17	−0.83*	−1.02	−0.10	0.13	−0.86*
Maintain	1.28	0.08	0.36	0.00	0.25	0.13	0.42*	−0.02
Release	−0.55	−0.05	0.28	−0.80*	−0.52	−0.05	0.21	−0.71*
Forget	−0.29	−0.02	0.27	−0.59*	−0.85	0.08	0.22	−0.43
Express	0.03	−0.16	0.22	−0.40	−0.56	0.04	0.33*	−0.46
Strong	0.28	−0.04	0.50	−0.33	−0.32	−0.01	0.28	−0.68
Social (total)	1.33	0.12*	0.68*	0.04	0.36	0.17*	0.66**	−0.26**
Other (total)	1.07	0.11**	0.65**	−0.02	−0.10	0.11**	0.38**	−0.23**
Background	1.11	0.07	0.41**	−0.04	−0.14	0.11*	0.37**	−0.26**
Entertainment	1.35	0.07	0.37	−0.03	0.22	0.09	0.39**	−0.22**
Boredom	1.10	0.06	0.77*	−0.07	0.00	0.05	0.38**	−0.32**
Habit	1.07	0.19*	0.78*	0.00	0.15	0.02	0.36*	−0.29*
Concentrate	1.16	0.18*	0.73*	0.08	0.13	0.19*	0.51**	−0.01
Distraction	0.43	0.19	0.72	−0.22	−0.50	0.14	0.32*	−0.14
Appreciate	1.15	0.24*	−	0.00	−0.32	0.23*	0.24	0.16
Creative	1.39	0.09	−	0.04	−0.05	0.15	0.20	−0.27
Memories	0.76	0.03	0.31	−0.17	−0.24	−0.41	0.37	−1.32*
Understand	0.29	−0.09	0.83*	−0.57	−0.89	−0.03	0.24	−0.43
Analyse	0.91	0.34	−	0.04	−0.22	0.21	0.89*	−0.33
Identify	0.46	0.34	−	0.09	−0.56	0.15	0.37	−0.44

Note. Missing values are due to insufficient data for those cases.

Change is significant at the .05 level (2-tailed, unadjusted). **Change is significant at the .0045 level (2-tailed, adjusted).

When considering all music episodes, valence was significantly increased for Other reasons overall, t(315) = 4.15, p < .001, d = 0.23, but no such increase was observed for Emotional reasons overall. The only emotional reason to yield a significant increase in valence was Raise t(83) = 3.71, p < .001, d = 0.40, while no reasons decreased valence.

When considering music episodes with a negative initial valence, both Emotional reasons, t(105) = 4.65, p < .001, d = 0.45, and Other reasons, t(85) = 5.78, p < .001, d = 0.62, yielded a significant increase in valence. The only specific reason to increase a negative valence was Background, t(41) = 3.32, p = .002, d = 0.51, while no reasons worsened a negative mood. When considering music episodes with a positive initial valence, only the emotional reason Raise significantly improved a positive emotional state, t(74) = 3.34, p = .001, d = 0.39.

Arousal change

Also shown in Table 2 are the initial arousal levels and change in arousal for Emotional and Other reasons for listening. Mean initial arousal was negative for all Emotional reasons, except for Maintain, Enjoy, and Raise. The lowest initial arousal levels were observed for Cope, Understand, Forget, and Reflect. When considering all music episodes, arousal was significantly increased for Other reasons overall, t(315) = 3.55, p < .001, d = 0.20. Similar to valence results, Raise was the only emotional reason to yield a significant increase in arousal level, t(83) = 4.56, p < .001, d = 0.50, while no reasons decreased arousal.

When considering music episodes with negative initial arousal, Emotional reasons t(229) = 7.20, p < .001, d = 0.47, Social reasons, t(86) = 4.98, p < .001, d = 0.53, and Other reasons, t(260) = 8.85, p < .001, d = 0.55, each yielded significant increases in arousal level. A positive change in arousal level was observed for three emotional reasons: Raise, t(37) = 5.02, p < .001, d = 0.81, Improve, t(50) = 3.28, p = .002, d = 0.46, and Relax, t(123) = 4.48, p < .001, d = 0.40. Arousal was also significantly increased for the Other reasons of Concentrate, t(43) = 3.46, p < .001, d = 0.52, Entertainment, t(110) = 4.14, p < .001, d = 0.39, Boredom, t(89) = 3.55, p < .001, d = 0.37, and Background, t(176) = 5.60, p < .001, d = 0.42.

When considering music episodes with positive initial arousal, Emotional, t(161) = −3.65, p < .001, d = 0.29, Social, t(103) = -3.28, p = .001, d = 0.32, and Other reasons, t(220) = −5.10, p < .001, d = 0.34, were each associated with a significant decrease in arousal levels. An arousal decrease was also observed for the Emotional reason Relax, t(51) = −3.48, p = .001, d = 0.48, and the Other reasons of Boredom, t(85) = −3.04, p < .001, d = 0.33, Background, t(141) = −3.63, p < .001, d = 0.30, and Entertainment, t(124) = −2.90, p < .001, d = 0.26. When considering all changes in arousal, no reason for listening significantly decreased an initially low arousal level, or increased an initially high arousal level.

Discussion

The current study aimed to determine both the frequencies and affective outcomes of reasons for personal music listening. It revealed several divergences from the existing literature in regard to frequencies of emotional reasons for listening, both in general and among certain demographic groups. Furthermore, it extended previous studies to assess the real-time change in valence and arousal for individual reasons for listening.

The first aim was to determine the frequencies of reasons for personal music listening, and any associations between emotional reasons and individual variables. In regard to frequencies, the most prominent finding was the increased reporting of emotional reasons while the participant was in a negative mood. While emotional reasons for listening accounted for less than a third of overall music episodes, they were used in the majority of episodes in which the listener was in a negative mood. This was reflected in the specific reasons that were used more frequently when the initial valence was negative, as compared to positive, with the top seven of these reasons coming from the emotional category. This finding offers an alternative to the well-established view that emotional reasons are the most important and frequently used reasons for listening to music (e.g., Lonsdale & North, 2011; Schafer et al., 2013). The m-ESM results have revealed that this is dependent on initial state, with emotional reasons used more frequently only when the listener is already in a negative mood. Such an observation has been unavailable through previous methodologies, without the ability to directly associate reasons to emotional state during listening experiences. As a majority of these previous studies have relied on retrospective reports, it may be that reasons utilised while in a negative mood have been over-reported, due to the inherent emotional salience of the listening context.

Another notable finding from the frequency data was that listening to background music was the most common reason for listening across all conditions, accounting for almost one fifth of all music episodes. Together, Background (“listening for background music”), Entertainment (“for entertainment”), and Boredom (“to reduce boredom”) accounted for over two fifths of all music episode reasons (but less than half of this when listeners were in a negative mood). These most commonly reported reasons are in accordance with those observed in the ESM study performed by North et al. (2004), as well as those reported by people listening on mobile devices while travelling (Heye & Lamont, 2010). These previous studies found enjoyment, to pass the time, and habit to be among the most frequently reported functions. The frequent use of music in the background or to reduce boredom provides strong support for the finding from previous ESM research that passive reasons for listening are more common than active reasons (North et al., 2004; Sloboda et al., 2001).

These frequency results indicate that initial mood is a key factor in determining how music is used: listening is generally a passive experience, unless the listener is in a negative mood, in which case music is most often used for emotional reasons. This distinction by initial mood may partially reflect attributes unique to personal music listening. This particular style of listening is largely a solitary activity, and highly flexible, allowing listeners to deliberately select music to match their current mood and situation. Previous experience sampling research has shown that listeners are more likely to use music for emotional reasons when listening alone, and that this is associated with lower states of mood valence and arousal (Thompson & Larson, 1995). Similarly, it has been shown that mean ratings of arousal, happiness-elation, and enjoyment-pleasure are lower when listening alone (Liljeström, Juslin, & Västfjäll, 2013). The private and accessible nature of personal music listening may therefore make it an ideal resource for self-regulation of adverse affective states.

The relative frequencies of emotional reasons by demographic, personality, and emotional health variables were also investigated. Remarkably, no associations were observed between emotional reason frequencies and either listener age or gender, signalling a clear divergence from previous research. It was expected that emotional reasons would be utilised more frequently by adolescent listeners, due to the importance of these functions at this developmental stage (Laiho, 2004; Lonsdale & North, 2011). In the current study, a lack of results may have been due to the young sample, and the relatively small age difference between the adolescent and adult groups. It was also expected that emotional reasons would be utilised more frequently by female listeners, as females have been consistently shown to use music more than males to fulfil emotional needs, and to enhance or change their emotions (North et al., 2000; Sloboda et al., 2001; Wells & Hakanen, 1991). In contrast to gender and age results, clear support was found for the hypothesis that emotional use of music would be associated with lower levels of emotional stability (Chamorro-Premuzic & Furnham, 2007).

These divergences from the literature may be explained by the ecological validity advantages of the m-ESM over questionnaires or qualitative reports. Retrospective responses often reflect reconstructive memory of personally relevant events, and can be altered by beliefs of situational appropriateness or experimenter demand (Gibson et al., 2000; Sloboda et al., 2001). Longer periods of retrospective recall have been shown to lead to a distortion of events through episodic memory decay, with an under-reporting of emotional and cognitive strategies (Ptacek, Smith, Espe, & Raffety, 1994). The temporal sensitivity advantages of the m-ESM should therefore result in a more accurate and reliable assessment of reason frequencies, with minimal recall bias. This explanation may be of particular relevance to the lack of gender differences in the current m-ESM data. It has been suggested that gender differences in music listening are based on gender-role socialisation, and the stereotypical belief that females “should be” more emotional than males (Boer et al., 2012). These socially established gender norms would be more likely to be reflected in responses to retrospective questionnaires, which are influenced by external expectations and situational appropriateness. It is therefore possible that previously reported gender differences in reasons for listening are partly due to a reliance on retrospective reporting. However, as previous ESM research has also shown female listeners to use emotional reasons more often (Sloboda et al., 2001), the absence of gender differences in the current study may instead reflect a unique property of personal music listening. As this style of listening is readily available, and can be accessed in complete privacy, it may be particularly suitable for fulfilling various emotional needs, regardless of listener demographic. Thus, listeners may utilise personal music listening for reasons distinct from those for which they would generally use music, or for which they would retrospectively report using music.

The second aim was to investigate how reasons for listening related to changes in emotional valence and arousal. When considering all music episodes, or those that began with the listener in a positive mood, Raise (“To raise energy/get pumped up”) was the only specific reason to significantly improve mood. When listeners were in a negative mood, this adverse state was only improved when participants were listening for background music, and not when listening for any specific emotional reason. This distinct lack of valence change for emotional reasons was unexpected, considering that several of these reasons would be deliberately used for the purpose of altering mood. While some of the expected changes in arousal were observed, such as Relax decreasing positive arousal and Raise increasing negative arousal, several expected changes in valence were not apparent. Most notably, Improve (“To improve my mood”) failed to increase valence under any condition. The finding that background music improves mood is somewhat supported by previous research that music can still be enjoyed while not the focus of the listener’s attention (North et al., 2004). One explanation for the lack of valence change for emotional reasons could be the selection of mood-congruent music, to sustain a negative mood for the purposes of “misery-sharing” (Gibson et al., 2000; Skånland, 2013; Zillmann & Vorderer, 2000). Investigation of this possibility will require an analysis of experience-level variables, including the valence of the selected music.

Taken together, the results reveal that all significant changes in valence were observed in a positive direction, while all significant changes in arousal were towards a neutral level. This provides some support for mood management theory (Zillmann, 1988), suggesting that music is selected to produce a positive hedonic shift, and to reach excitatory homeostasis. This neutralising effect on arousal was observed for Relax, along with the three most frequently used reasons overall (Background, Entertainment, and Boredom). However, non-music episodes also elicited significant shifts in arousal towards a neutral state, which suggests the influence of an extraneous factor, rather than the presence of music (alternatively, this could be related to the violated assumption of normality for the non-music condition). Also of interest from the affective results are the apparent high correlations between valence and arousal levels for all reason categories. This indicates that the negative moods in which listeners select emotional reasons are more related to sadness and depression, rather than anger, while the positive moods experienced while listening are generally associated with high energy levels.

Through closer analysis of the data, the importance of several emotional reasons for listening becomes clear. First, it is apparent that listening to music to raise energy is uniquely effective, as it was the only emotional reason to improve valence, and one of only three to increase arousal levels. These changes were often from an already positive, high-energy state, with Raise linked to some of the highest levels of both initial valence and arousal. These unique properties, along with the finding that Raise was the only reason to be used more often by male than female listeners, warrants further analysis into the deliberate use of music to increase arousal. In contrast to Raise, it is apparent that certain emotional reasons for listening are closely associated with negative states. Specifically, Cope (“to cope with a situation”) and Forget (“to forget my problems”) were used at times when listeners were in a negative mood, and experiencing the lowest levels of arousal. This suggests that these reasons were utilised during episodes associated with the experience of sadness (rather than high-arousal states such as anger). In accordance with this, these reasons were among those most likely to be used in a negative (as compared to positive) initial mood. However, Cope and Forget failed to elicit a significant change in valence or arousal under any condition. Therefore, the reasons most frequently used during periods of sadness did not appear to alleviate these adverse states. This may further indicate a deliberate maintenance of a negative state through the selection of mood-congruent music (Gibson et al., 2000; Skånland, 2013; Zillmann & Vorderer, 2000).

In addition to being closely associated with negative initial mood, Cope and Forget were also associated with lower levels of emotional health, well-being, and emotional stability. This indicates a possible direction of influence: it appears that poor emotional health and well-being, as a predictor of negative initial mood, influences the reason selected for listening. This is similar to previous m-ESM results demonstrating that particular emotion regulation strategies used during listening are related to both negative emotional traits and states (Randall, Rickard, & Vella-Brodrick, 2014). This previous study found response-focused strategies, in particular that of suppression, to be associated with low initial valence, poor mental health, and well-being. However, while response-focused regulation strategies were found to alleviate negative states (Randall et al. 2014), no significant affective change was observed for the emotional reasons in the current study. As with suppression, both “To cope with a situation” and “To forget my problems” seem to indicate a style of listening that is initiated in response to a negative situation. This suggests that those with poor emotional health and well-being may use music more frequently in response to adverse emotional experiences. This is supported by findings that depressed listeners use emotional reasons more frequently than non-depressed listeners, as music may allow them to experience and deal with their negative affect (Wilhelm, Gillis, Schubert, & Whittle, 2013). This accentuates the flexible role of music, as a unique resource which individuals can utilise in times of suffering, to fulfil specific emotional needs (Lonsdale & North, 2011).

While a main purpose of the m-ESM design is to minimise the time between experience and reporting, this approach introduces certain limitations that must be addressed. First, the current design of music ESRs does not allow for the selection of multiple reasons for listening; instead asking the participant to select a single, primary reason. Previous research has established that music can serve multiple functions simultaneously (Greasley & Lamont, 2011; Heye & Lamont, 2010). Diary-based research has shown that on average, 1.66 strategies are employed during music episodes, while ESM research has revealed that a majority of episodes involve three or more reasons for listening (Greasley & Lamont, 2011; van Goethem & Sloboda, 2011). Therefore, the forced selection of a single reason provides only a simplified assessment of the various possible motivations of the listener. As such, the current results must be interpreted as relating to primary reasons for listening.

Related to this limitation is the forced selection of a reason category (emotional, social, or other) before selecting a specific primary reason. This branching format introduces the possibility that the most appropriate reason for listening is not always selected. For example, participants may not be immediately aware that their reason for listening is emotional or social, without viewing the specific reasons included on these lists. As such, it is possible that reasons from the “other” category are over-represented in the current set of results, as this may be seen as a default option of the three. This limitation may offer an alternate explanation for the main finding that emotional reasons were utilised more frequently only when listeners were in a negative mood, as participants may be more likely to recognise the emotional aspect of their listening experience while in this state. Conversely, several of the reasons in the non-emotional categories somewhat relate to moods, highlighting the difficulty of the forced choice of category.

These design features – selection of a single reason for listening, and selection of a reason category – were deliberately incorporated in the m-ESM. While they do introduce potential limitations, these features vastly improve answering efficiency, minimising response time and recall bias. Furthermore, the branching design allowed for selection from a total of 33 reasons, which is more than double the number provided in any previous ESM study. A recent development to the m-ESM design has addressed these limitations, with a more intuitive, user-focused interface that allows for the selection of multiple reasons for listening. Future research using the m-ESM will involve a multilevel structural equation model of personal music listening, accounting for emotional outcomes on both listener and experience levels (Randall & Rickard, in press). Analysis of such a model should clarify some of the current findings, such as the lack of valence change for emotional reasons.

The current study investigated the reasons for listening to music in everyday life, with a focus on emotional reasons for personal music listening. The findings revealed that emotional reasons for listening were the most frequently used, only when the listener was in a negative mood. The identification of prior mood as a key variable extends previous findings, and highlights the real-time emotion assessment benefits of the m-ESM. Notably, no associations were observed between the frequent use of emotional reasons for listening, and either age or gender. Another key finding was that listening to either cope with a situation or forget problems is related to both adverse emotional states and traits. This suggests that emotional reasons for listening are utilised to fulfil specific emotional needs, which are determined by a participant’s initial mood, and influenced by emotional health and well-being.

Footnotes

Ethical approval

Ethical approval for this project was given by Monash University Human Research Ethics Committee (MUHREC) [CF11/0627 – 2011000240].

Funding

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

Notes

References

Boer

Fischer

(2012). Towards a holistic model of functions of music listening across cultures: A culturally decentred qualitative approach. Psychology of Music, 40(2), 179–200.

Boer

Fischer

Tekman

H. G.

Abubakar

Njenga

Zenger

(2012). Young people’s topography of musical functions: Personal, social and cultural experiences with music across genders and six societies. International Journal of Psychology, 47(5), 355–369.

Bull

(2005). No dead air! The iPod and the culture of mobile listening. Leisure Studies, 24(4), 343–355.

Chamorro-Premuzic

Furnham

(2007). Personality and music: Can traits explain how people use music in everyday life? British Journal of Psychology, 98(2), 175–185.

Chin

T. C.

Rickard

N. S.

(2012). The music USE (MUSE) questionnaire: An instrument to measure engagement in music. Music Perception, 29(4), 429–446.

Csikszentmihalyi

LeFevre

(1989). Optimal experience in work and leisure. Journal of Personality and Social Psychology, 56(5), 815–822.

Denora

(1999). Music as a technology of the self. Poetics, 27(1), 31–56.

Diener

Emmons

R. A.

Larsen

R. J.

Griffin

(1985). The satisfaction with life scale. Journal of Personality Assessment, 49(1), 71–75.

Gibson

Aust

C. F.

Zillmann

(2000). Loneliness of adolescents and their choice and enjoyment of love-celebrating versus love-lamenting popular music. Empirical Studies of the Arts, 18(1), 43–48.

10.

Goldberg

L. R.

(1992). The development of markers for the Big-Five factor structure. Psychological Assessment, 4(1), 26–42.

11.

Greasley

A. E.

Lamont

(2011). Exploring engagement with music in everyday life using experience sampling methodology. Musicae Scientiae, 15(1), 45–71.

12.

Haake

A. B.

(2011). Individual music listening in workplace settings: An exploratory survey of offices in the UK. Musicae Scientiae, 15(1), 107–129.

13.

Hargreaves

D. J.

North

A. C.

(1999). The functions of music in everyday life: Redefining the social in music psychology. Psychology of Music, 27(1), 71–83.

14.

Hektner

J. M.

Schmidt

J. A.

Csikszentmihalyi

(2007). Experience sampling method: Measuring the quality of everyday life. Thousand Oaks, CA: Sage.

15.

Heye

Lamont

(2010). Mobile listening situations in everyday life: The use of MP3 players while travelling. Musicae Scientiae, 14(1), 95–120.

16.

Juslin

P. N.

Laukka

(2004). Expression, perception, and induction of musical emotions: A review and a questionnaire study of everyday listening. Journal of New Music Research, 33(3), 217–238.

17.

Juslin

P. N.

Liljestrom

Västfjäll

Barradas

Silva

(2008). An experience sampling study of emotional reactions to music: Listener, music, and situation. Emotion, 8(5), 668–683.

18.

Laiho

(2004). The psychological functions of music in adolescence. Nordic Journal of Music Therapy, 13(1), 47–63.

19.

Laukka

(2007). Uses of music and psychological well-being among the elderly. Journal of Happiness Studies, 8(2), 215–241.

20.

Liljeström

Juslin

P. N.

Västfjäll

(2013). Experimental evidence of the roles of music choice, social context, and listener personality in emotional reactions to music. Psychology of Music, 41(5), 579–599. doi:10.1177/0305735612440615

21.

Lonsdale

A. J.

North

A. C.

(2011). Why do we listen to music? A uses and gratifications analysis. British Journal of Psychology, 102(1), 108–134.

22.

Lovibond

S. H.

Lovibond

P. F.

(1995). Manual for the Depression Anxiety Stress Scales (2nd ed.). Sydney, Australia: Psychology Foundation.

23.

North

A. C.

Hargreaves

D. J.

Hargreaves

J. J.

(2004). Uses of music in everyday life. Music Perception, 22(1), 41–77.

24.

North

A. C.

Hargreaves

D. J.

O’Neill

S. A.

(2000). The importance of music to adolescents. British Journal of Educational Psychology, 70(2), 255–272.

25.

Ptacek

Smith

R. E.

Espe

Raffety

(1994). Limited correspondence between daily coping reports and retrospective coping recall. Psychological Assessment, 6(1), 41–49.

26.

Randall

W. M.

Rickard

N. S.

(2013). Development and trial of a mobile Experience Sampling Method (m-ESM) for personal music listening. Music Perception, 31(2), 157–170.

27.

Randall

W. M.

Rickard

N. S.

(in press). Personal music listening: A model of emotional outcomes developed through mobile experience sampling. Music Perception.

28.

Randall

W. M.

Rickard

N. S.

Vella-Brodrick

D. A.

(2014). Emotional outcomes of regulation strategies used during personal music listening: A mobile experience sampling study. Musicae Scientiae, 18(3), 275–291. doi:10.1177/1029864914536430

29.

Saarikallio

(2011). Music as emotional self-regulation throughout adulthood. Psychology of Music, 39(3), 307–327.

30.

Saarikallio

Erkkila

(2007). The role of music in adolescents’ mood regulation. Psychology of Music, 35(1), 88–109.

31.

Schafer

Sedlmeier

(2009). From the functions of music to music preference. Psychology of Music, 37(3), 279–300.

32.

Schafer

Sedlmeier

Stadtler

Huron

(2013). The psychological functions of music listening. Frontiers in Psychology, 4, 511.

33.

Schafer

Tipandjan

Sedlmeier

(2012). The functions of music and their relationship to music preference in India and Germany. International Journal of Psychology, 47(5), 370–380.

34.

Skånland

M. S.

(2013). Everyday music listening and affect regulation: The role of MP3 players. International Journal of Qualitative Studies on Health and Well-being, 8, 20595. Retrieved from http://www.ijqhw.net/index.php/qhw/article/view/20595

35.

Sloboda

J. A.

O’Neill

S. A.

Ivaldi

(2001). Functions of music in everyday life: An exploratory study using the Experience Sampling Method. Musicae Scientiae, 5(1), 9–32.

36.

Tarrant

North

A. C.

Hargreaves

D. J.

(2000). English and American adolescents’ reasons for listening to music. Psychology of Music, 28(2), 166–173.

37.

Thoma

M. V.

Ryf

Mohiyeddini

Ehlert

Nater

U. M.

(2012). Emotion regulation through listening to music in everyday situations. Cognition & Emotion, 26(3), 550–560.

38.

Thoma

M. V.

Scholz

Ehlert

Nater

U. M.

(2012). Listening to music and physiological and psychological functioning: The mediating role of emotion regulation and stress reactivity. Psychology & Health, 27(2), 227–241.

39.

Thompson

R. L.

Larson

(1995). Social context and the subjective experience of different types of rock music. Journal of Youth and Adolescence, 24(6), 731–744.

40.

van Goethem

Sloboda

. (2011). The functions of music for affect regulation. Musicae Scientiae, 15(2), 208–228.

41.

Vuoskoski

J. K.

Eerola

(2011). Measuring music-induced emotion: A comparison of emotion models, personality biases, and intensity of experiences. Musicae Scientiae, 15(2), 159–173.

42.

Watson

Clark

L. A.

Tellegen

(1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070.

43.

Wells

Hakanen

E. A.

(1991). The emotional use of popular-music by adolescents. Journalism Quarterly, 68(3), 445–454.

44.

Wilhelm

Gillis

Schubert

Whittle

E. L.

(2013). On a blue note: Depressed peoples’ reasons for listening to music. Music and Medicine, 5(2), 76–83.

45.

Zillmann

(1988). Mood management: Using entertainment to full advantage. In Donohew

Sypher

H. E.

Higgins

E. T.

(Eds.), Communication, social cognition, and affect (pp. 147–171). Hillsdale, NJ: Lawrence Erlbaum.

46.

Zillmann

Vorderer

(2000). Media entertainment: The psychology of its appeal. Mahwah, NJ: Lawrence Erlbaum.