research paper about music video

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Front Public Health

Research on the effect of different types of short music videos on viewers' psychological emotions

Associated data.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

There is now widespread agreement that different types of short music videos can cause viewers to have psychological emotions, and significant new findings have been discovered in the study of how music affects listeners' affective reactions. However, there is still controversy regarding research on the inclinations toward behavior and autonomic neurophysiological reactions of musical emotions. The psychological states of viewers of various types of short music videos are yet unknown and require further study. This study investigates how different types of short music videos affect viewers' psychological responses, placing particular emphasis on the following variables: rhythm type (stable rhythm and flow rhythm) and music short video type (narrative, live, and funny). In an experiment, viewers' psychological responses to several short music videos were investigated to determine the impact of different short music video styles and rhythms on musically induced emotions.

Introduction

Language and music both distinguish human social interaction and are essential components of communication. Music has a profound impact on all aspects of life, including work and study. It also helps individuals connect and communicate emotionally ( 1 ). Watching brief music videos allows viewers to both comprehend the makers' intentions and the emotions they hope to evoke as well as experience those emotions firsthand ( 2 , 3 ). According to Koordeman et al. ( 4 ), the process of breaking down musical emotions can be broken down into different levels of individual cognition and experience. The consequences of each level vary in terms of the emotional content and manner of production, and they include musical emotion perception, experience, and musical taste. The relationship between music and emotions has been the topic of much psychological research, according to Fuentes-Sánchez et al. ( 5 ), and relevant studies have demonstrated how specific elements of listeners' interactions with music can alter how they perceive and experience musical emotions. To identify the gaps in the theory and measurement techniques of musical emotion perception, the researchers first conducted a thorough literature review from the starting point of musical emotion, mastered the current theoretical research and application, and discovered the theoretical gaps ( 6 ). Johnson et al. ( 7 ) states that Western classical music cannot explain the emotional responses to Chinese traditional pentatonic modes, and nothing is known about how viewers' emotional responses to various pop music videos' visual components. This study focused on the visual aspects of short music videos to gauge viewers' emotional perception. Research questions and experimental hypotheses on the influence of different features on psychological emotions are proposed to discuss the modes and video types that lack a theoretical foundation. Psychological experiments were designed and conducted to test the experimental hypotheses, address research issues, draw conclusions, and improve theoretical frameworks.

The main contributions of this paper are as follows:

(1) This study investigates how different types of short music videos affect viewers' psychological responses, placing particular emphasis on the following variables: rhythm type (stable rhythm and flow rhythm) and music short video type (narrative, live, and funny).

(2) Compared to narrative and live music short videos, funny music short videos significantly increased the positive emotion in viewers. In comparison with that, the neutral emotion valence induced by the narrative and live music short video was significantly higher. Performance music short videos produced more polarized emotional valence than narrative music short videos. The emotional valence, arousal, vitality, and restlessness in a brief music video were significantly influenced by the rhythm type.

Mechanism of short music video affecting viewer's psychological emotion

Short music videos affect viewers' psychological mood.

There has been relatively little research on how short videos affect people's perception over the past few decades, despite the fact that music has been shown to influence people's perception and memory of short videos ( 8 ). Barney et al. ( 9 ) measured the emotional experience of musicians and non-musicians through short videos of musical performances and discovered that both musicians and non-musicians who only listened to music and those who watched music and short videos simultaneously reported having a stronger emotional experience than those who only watched short videos of musical performances. In addition, musicians think that symphonies and quick videos of performances enhance the emotional impact of the music ( 10 ). Different kinds of short music videos can affect viewers' emotions and trigger different cognitive or emotional responses. According to Tian and Zhao ( 11 ), listening to music while watching movie clips led to lower performance on cognitive tasks and higher mood scores. On cognitive tests, the group that included live performance and music did better. If the short music videos' content is in line with psychological feelings, the viewer will also feel different emotions and has a different experience. In their study, Egidi and Caramazza ( 12 ) discovered that the “emotional consistency” rule governs how short music videos' emotional effects affect music-induced feelings. In Egidi's study, music-based short videos were used to elicit either positive, neutral, or negative emotions in the participants. Initially, neutral mood music was combined with positive emotions in the videos after the participants reported feeling positive, while initially neutral mood music was combined with negative emotions in the videos to elicit negative emotions in the participants.

Music short video type

Since users process both the content of songs and short music videos at the same time, the associative consistency model predicts that their emotional response will be influenced by the relationship between the content of songs and lyrics and the content of short music videos. In addition, the types of music short videos, such as narrative and live performance, will influence different musical perceptions, which will then result in various emotional experiences or cognitive styles ( 13 ). As a result, pop short music videos are divided into three categories based on two criteria: whether the narrative in the short music video is related to the lyrics, and whether it is not.

Short music videos with narration: These short music videos tell one or more stories that are essentially complete;

Live short music videos: In this short music video, the photographer performs live;

Funny short music videos: Snack-sized music videos with wacky narratives, animations, or live performances.

Hypothesis of the influence of short videos of different types of music on viewers' psychological emotions

Hypothesis of the induced influence of narrative short music videos.

Gardstrom et al. ( 14 ) asserts that after being stimulated by brief music videos, users' emotional experiences will become more intense and polarized. In addition, the stimulation of brief music videos with a plot can strengthen emotional reaction while impairing cognitive task performance, whereas the stimulation of brief music videos with abstract content can make people require more cognitive processing and, as a result, weaken emotional reaction. As a result, this study puts forth hypothesis 1–3, which states that narrative short music videos are more likely to elicit strong emotional (valence and arousal) responses than performance short music videos.

Hypothesis 1: The emotional valence induced by narrative short music videos is more polarized than performance short music videos.

Hypothesis 2: The emotional arousal level induced by narrative short music videos is more polarized than performance short music videos.

Hypothesis 3: The type of short music videos has an effect on the specific types of emotions induced by short music videos.

Hypothesis of the induced influence of live short music videos

According to theoretical research on music and performing music rhythm type short video stimulus, including rhythm in music performance type short video stimulation can improve the listener's emotional experience. Ma et al. ( 15 ) believed that the kata music short video stable rhythm type has a greater partial positive titer and higher arousal level than the flowing rhythm type on the induced mood. The hypothesis of this study was that the stimulation of live short music videos would not change the differences and trends of rhythmical emotions.

Hypothesis 4: Rhythm type has an effect on the emotion induced by live short music videos. Steady rhythm has more positive valence and higher arousal levels than flow rhythm.

Hypothesis 5: Rhythmic type has an effect on specific emotions induced by live short music videos. Stable rhythm and flow rhythm induce different types of specific musical emotions.

Hypothesis of the induced influence of funny short music videos

Funny short music videos are a common type of network video, and it is important to understand what makes them so popular. Merry and Silverman ( 16 ) investigated the types of emotional reactions that humorous short music videos will elicit from the viewpoint of video-induced emotion and contrasted this short music video form with more conventional short music video types (narrative type and live performance type) induced emotion. Therefore, this study's main hypothesis is that the rhythmic and short-form music features interact to affect emotional experience.

Hypothesis 6: Funny short music videos and rhythm have an interaction on psychological emotions.

Hypothesis 7: Funny short music videos interact with rhythmic music for specific emotions.

Empirical process

Research object.

Social media and BBS were used to find 25 university students who would take part in the experiment. This study chose the applicants through an online questionnaire to take into account how the subjects' musical training and familiarity with music affected their emotions. The applicants had to list at least one line from the song or scene from the brief music video in the screening questionnaire. Applications with an overall accuracy rate of <60% will not be taken into consideration. Furthermore, all subjects were right-handed, healthy, and free from either short- or long-term hearing loss. The basic information of the research object is shown in Table 1 .

Basic information on the research object.

Independent variables

Three independent factors were examined: rhythmic style, frame type (with two levels: narrative and non-narrative), and lyric relevance (with two levels: lyric relevant and lyric irrelevant) (two levels: 1—stable rhythm and 2—flow rhythm). The two that relate to the content of music videos are picture type and lyrics relevance. To make operation and analysis easier, we integrated the influencing factors for music video content into a single variable, music short video type, which has three levels:

• 1. Narrative short music video (with a complete story).
• 2. Live short music videos (Live performance, no plot).
• 3. Funny short music videos (Parody, no plot).
• The rhythmic variable contains two levels:
• 1. Stable (mainly in 4, 2, and stable 3 time, mostly polka, march, pop music).
• 2. Flow type (mainly in 6-meter, 9-meter, or asymmetrical rhythm, moving freely and scattered).

Dependent variable

Psychometric aspects of emotion:

(1) Wake up as determined by the SAM scale. The Likert scale has nine levels, with one representing the highest level of arousal, five the neutral level, and nine the lowest level.

Titer as determined on the SAM scale. The Likert scale at level nine, with one denoting positive titers (such as joy and happiness), five denoting neutral, and nine denoting negative titers (e.g., sadness, disappointment, and anger).

(2) As evaluated by the Geneva Musical Mood Scale, poetic. The Richter scale has five levels, with one denoting very inconsistent, three denoting neutral, and five denoting very consistent ( 17 ).

Dynamic as determined by the Geneva Musical Mood Scale. The scale included five levels, with one denoting a severe disagreement, three denoting neutral, and five denoting a strong agreement.

(3) Uncomfortable, according to the Geneva Musical Mood Scale. On a Likert scale of one to five, where one is highly negative, three is neutral, and five is very positive.

Selection of short music video selection

The combination of various levels of independent variables was used to select the music videos that would be used in the study, asking four professionals with more than 4 years of music industry experience to rate the chosen music videos according to the strength of the independent variable factors they represent. Selection of short music video selection is shown in Table 2 .

Selection of short music video selection.

In addition, the subjects and music chosen for this study follow the following principles:

(1) The subjects made every attempt to become familiar with the songs to reduce the differences in familiarity levels. All subjects who were approved to participate in the experiment completed an online questionnaire survey in advance, and it was discovered that at least 60% of the music was recognizable to them.

(2) Music video integrity: Since strong emotional reactions depend on a range of circumstances and musical meanings, each brief music video, which usually lasts between three and 5 min, should be as inclusive as possible. Integrity is also demonstrated by the consistency of the elements: the vocal portion, accompaniment, and visual are all still present in the film together with all the other elements.

Experiment design

The intra-group design mandated that each participant see all six music videos. Because the entire experimental process is straightforward and interesting, the consequences of weariness and load may be ignored. The in-group design's problem with the learning effect was resolved by the Latin square design, which has six sequential permutations and an equal chance for each dependent variable processing combination to appear at each location. In this study, 25 participants were randomly assigned to six sequential pairings.

Data analysis model:

μis the population mean; short music videos type refers to the fixed effect of music short video; RHYTHM refers to the rhythmic fixed effect; ε is a random effect.

Latin square design is shown in Table 3 .

Latin square design.

Empirical results

Reliability and validity tests.

The four professionals with more than 4 years of professional music training evaluated the six songs and videos used in this experiment, and all of them passed. They strictly adhere to the corresponding independent variable level. This experiment made use of the SAM Mood Scale and the Geneva Musical Mood Scale ( 18 ). The Geneva Musical Mood Scale's internal consistency was examined using Cronbach's alpha. The used gemS-25 scale had measurement reliability scores of 0.874 for poetic, 0.929 for dynamic, and 0.709 for uneasy. In actual experiments, it is acceptable for the scale's internal consistency to rise above 0.6 if the measured value changes with the tasks. As a result, the three components of the Geneva musical mood scale will be included in the research's statistical model.

Model fit test

The single sample Kolmogorov–Smirnov test was used to test the normality of the dependent variable data. The homogeneity test of variance is shown in Table 4 .

Homogeneity test of variance.

The results showed that only poetry and vitality in the Geneva scale passed the normality test. Other psychometric data (Awaken, dynamic, and uneasy) still fail the normality test after the necessary conversion.

Psychological data analysis

The impact of the type of music rhythm and brief music video on viewers' emotions was investigated using MANOVA ( 19 ). The independent variables in this model are the genre of music short video and the beat, whereas the dependent variables are poetry and energy (within the group). As a result, this approach does not account for individual differences. The non-parametric test method of Kruskal–Wallis and Mann–Whitney U was used in this work to conduct a one-way ANOVA for the arousal level, potency, and anxiety of emotions ( 20 ). In contrast to live performance short music videos, it is believed that narrative short music videos will elicit more emotions directly tied to emotional reaction. Narrative short music videos are thought to trigger more emotions associated with cognitive processing (like poetry). The expectation is that narrative short music videos will stir up more feelings than live short music videos (such as vitality). Hypothesis testing of narrative short music videos is shown in Table 5 .

Hypothesis testing of narrative short music videos.

The findings demonstrated that listening to humorous short music videos significantly increased listeners' positive valence emotions (happiness) compared to listening to narrative and live short music videos. Comparing the subjects' scores for the three brief music videos' potency and arousal levels revealed that while their overall potency was favorable, the scores for the videos' arousal levels were neutral. The potency and arousal levels, however, did not significantly differ between the live short music videos and the narrative short music videos. Funny short music videos are titillating, have a happy, joyful mood, and site-type music myopia bands, and give the viewers more of an associated dynamic mood. Narrative short music videos do induce more related to the poetic mood, and the plot contains emotional content that can cause psychological emotional resonance.

It is believed that while flowing rhythm can induce low arousal levels, which are more closely related to negative emotions, emotional, and poetic emotions, stable rhythm can induce high arousal levels, energetic, and positive emotions. The psychological and emotional impact of live short music videos on viewers is shown in Table 6 .

Psychological and emotional impact of live short music videos on viewers.

The results showed that while awakening levels for the two rhythm types tended to be neutral, the awaken level of listeners who preferred the steady rhythm type was significantly higher than that of listeners who preferred the flow rhythm type. According to valence, steady rhythm significantly increases good feelings while a flowing rhythm significantly increases negative emotion. Fixed rhythm thus considerably increases viewers' emotional arousal level and promotes the growth of more favorable feelings. Stable rhythms induce cheerful, upbeat, and energizing feelings, whereas flow rhythms are more likely to elicit sad and dismal ones.

The results of the statistical test show that the interaction between funny short music video type and rhythm type has significant effects on the measured values of poetic emotion ( F = 4.281, P < 0.05) and vitality emotion ( F = 5.074, P < 0.05).

To further analyze the interaction effect between poetic and dynamic, this study conducted a simple effect analysis on three types of short music videos, respectively. The results are analyzed as shown in Table 7 .

Simple effect analysis of short music video.

Almost nothing in the amusing short music videos had any rhythmic impact on the participants' poetic feelings. While the participants were watching live short music videos, their sense of poetry revealed very substantial changes with distinct rhythm types: When subjects were viewing narrative short music videos, they felt more poetic when the beat was more steady; when they were watching live short music videos, they felt less poetic when the rhythm was more constant.

The individuals' sense of dynamic is essentially unaffected by the rhythm type when they watch the live short music videos. Short music videos significantly increased dynamic emotion when the individuals were watching narrative short music videos as opposed to humorous short music videos. This resulted from an improvement in rhythm stability. To put it another way, narratives might accentuate the potential for intense feelings to rise along with greater rhythmic steadiness.

Correlation analysis of dependent variables

To further understand the psychological measures of music-video-induced emotion, this study conducted a correlation analysis of dependent variables. The results with strong correlation are listed in Table 8 .

Correlation analysis of dependent variables.

The outcomes demonstrated that, in contrast to music-induced emotions, unease played a dominant role in those emotions. Emotion valence and short music video type showed a significant positive correlation. A close relationship exists between the brief video of a sense of vitality music and a particular musical mood.

Hypothesis testing results

The verification of all hypotheses is summarized in this study, and the hypothesis testing results are shown in Table 9 .

Hypothesis testing results.

Discussion on the result

The results of this study are summarized as follows:

(1) Influence of short music video type

Different types of music short videos can elicit emotional reactions of varying intensity. In terms of emotional valence, narrative short music video ratings are typically more divisive than live short music video ratings. Storylines do not require as much fine processing, so people are more likely to process information along the edges, which intensifies their emotional reactions. Short performance music videos, on the contrary, tend to be more abstract and may call for more delicate cognitive processing. The short live music video induces a lower level of poetic emotion than the short narrative music video, which is more sensitive to poetic emotion.

(2) Funny music short video-induced emotions

Funny short music films cause various types and levels of emotions than standard short music videos do. Compared to the narrative and live short music videos, the highly entertaining content made the subjects laugh and they tended to score favorably on the valence scale. According to psychological, physiological, and experimental observation data, when people like the humorous music short video used in this study, they experience strong feelings of happiness and excitement.

(3) Music short video type is influenced by rhythm type

The results of the hypothesis testing indicate that various types of music-specific short videos are responsive to various emotions. Poetry-related rhythm types respond to narrative and live music short videos for poetic emotion, but not to funny music short videos. While flowing rhythm produced a more poetic mood than steady rhythm in a short live music video, steady rhythm produced a more poetic mood in a short narrative music video. While live music videos are almost never affected by rhythm, the first two categories are higher and will be affected by rhythm for dynamic emotions in narrative and funny short music videos. The atmosphere, the performance, and the movements of the singers in short live music videos based on live performance and audience interaction energized the subjects more than the music itself did (in this study).

According to the findings of this study, the type of short music video significantly affects the valence of emotion, as does the interaction between the type of short music video and musical characteristics. Compared to narrative and live music short videos, funny music short videos significantly increased the positive emotion in viewers. In comparison with that, the neutral emotion valence induced by the narrative and live music short video was significantly higher. Performance music short videos produced more polarized emotional valence than narrative music short videos. The emotional valence, arousal, vitality, and restlessness in a brief music video were significantly influenced by the rhythm type. This study encountered some issues while investigating the impact of different types of music videos on musically induced emotions. The influence of personality, age, and gender on musical emotion perception has been discussed in studies on individual differences in musical emotion perception, but previous findings have not produced significant effects in this study due to time and energy constraints. Future research can begin from two angles: first, experiment on people of different ages and backgrounds; and second, conduct group experiments to see whether the emotions elicited by pop music and classical music with the same musical characteristics were influenced by personal differences and musical background.

Data availability statement

Ethics statement.

Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

Author contributions

LM completed relevant research work on the manuscript and agreed to publish it in this journal.

Conflict of interest

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

To read this content please select one of the options below:

Please note you do not have access to teaching notes, music videos on youtube: exploring participatory culture on social media.

Symbolic Interactionist Takes on Music

ISBN : 978-1-78635-048-0 , eISBN : 978-1-78635-047-3

Publication date: 1 October 2016

The body of scholarship on YouTube is an expanding area of scholarly inquiry. Existent research indicates that music videos are one of the most salient features of YouTube. Interactionist research about popular music has provided important insights through interviews with fans and audience members; however, this work has yet to examine audience engagement with music videos on YouTube. Using Qualitative Media Analysis, I illustrate how the researcher of popular music can work with user comments collected from YouTube. Thematic understandings largely consistent with nostalgia that emerged from an analysis of user-generated comments in response to selected music videos on YouTube are explored. I conclude by suggesting some directions for future research.

• Social media
• Qualitative media analysis

Acknowledgements

Acknowledgments.

Thanks to Joe Kotarba for his tireless efforts to advance the interactionist study of music; and thanks to Isabel Scheuneman Scott for her helpful comments on an earlier draft of this paper. Additionally, I am grateful to the participants of the 2014 Couch-Stone Symposium; and to T.K. Johnson for helping me to think through some of the ideas presented in this paper.

Schneider, C.J. (2016), "Music Videos on YouTube: Exploring Participatory Culture on Social Media", Symbolic Interactionist Takes on Music ( Studies in Symbolic Interaction, Vol. 47 ), Emerald Group Publishing Limited, Leeds, pp. 97-117. https://doi.org/10.1108/S0163-239620160000047016

Emerald Group Publishing Limited

Copyright © 2016 Emerald Group Publishing Limited

All feedback is valuable

Please share your general feedback

Report an issue or find answers to frequently asked questions

Contact Customer Support

Assessing complementarities between live performances and YouTube video streaming

• Open access
• Published: 01 June 2023
• Volume 65 , pages 2953–2978, ( 2023 )

Cite this article

You have full access to this open access article

• Juan D. Montoro-Pons   ORCID: orcid.org/0000-0002-5180-7360 1 ,
• María Caballer-Tarazona   ORCID: orcid.org/0000-0001-9242-0464 1 &
• Manuel Cuadrado-García   ORCID: orcid.org/0000-0002-0145-515X 2  

1694 Accesses

Explore all metrics

Digitization and increased accessibility to recorded music have made revenue-generating activities increasingly tied to live performances. In this context, identifying the full impact of concerts (namely capturing the value of activities that emerge as a consequence of them) is of primary interest to assess the sustainability of the different music ecosystems. This paper analyzes spillover effects from playing live to YouTube video streaming. A sample of 190 artists performing in two international music festivals in years 2016 to 2019 has been selected, and the temporal patterns of online video searches for each one have been collected. Using a regression discontinuity design, results show a discrete jump of the YouTube search index for the average performer in the sample after playing live. Furthermore, there is evidence of a gender-specific effect: female performers experience a greater increase in YouTube searches. Though exploratory, this gender bias is consistent with potential theoretical explanations to be explored. Overall, findings provide causal evidence of the effect of live performances on a related but different market (i.e., recorded music), which underlines how technological disruptions may enable alternative revenue sources for musicians.

Similar content being viewed by others

It takes two, baby! Feature artist collaborations and streaming demand for music

Economics, Marketing and Performances of US Classical Music: Journeyin’ Together to de Promise Land

Attaining #fame: Female Cover Musician’s Self-Fashioning and Socio-musical Interactions with Live Stream Audiences in Chennai and Beyond

Avoid common mistakes on your manuscript.

1 Introduction

Digitization and the subsequent increased accessibility to content have challenged traditional revenue sources in the cultural sectors while new ones have emerged. In the case of music, this disruption has had a structural impact on the organization of the industry and the way in which audiences, musicians and intermediaries relate to each other (Hesmondhalgh and Meier 2018 ). Indeed, the last twenty years have witnessed a rapid transformation in music consumption and distribution where the decline of physical sales has been matched by an increasing relevance of live music attendance, and the streaming of digital contents has emerged as a business that has partially made up for the decline in recorded music sales (Montoro-Pons et al. 2021 ). Overall, technological innovation and change have transformed the landscape actors within the sector face.

The widespread use of new digital environments has facilitated interactions that open new and innovative ways of value creation and capture. Although the outcome of these may not be relevant in terms of their magnitude, they feedback and reinforce other revenue streams and, as such, should be considered as part of the wider music ecosystems (Hesmondhalgh 2021 ). The literature on the economics of the cultural industries has analyzed the impact emerging business models (especially in the music sectors) have on related products and/or services by creating complementarities or inducing substitutive effects. These can be broadly classified as within-market, mainly related to the effect on established business models as new ones are introduced, or cross-market, as spillovers across different (though related) products and/or services. The latter implies that cultural participation in one market can foster participation in a related, yet different, market. In this paper, we posit that live performances have a net positive effect on audiovisual (i.e., recorded) music consumption. These are spillover effects that emerge as a consequence of the exposure of consumers’ to new market information (through direct participation or indirectly by contagion) which increases the demand for a related market.

The focus of this research is to empirically examine cross-market effects from live music into audiovisual content viewing. While most of the literature focuses on the path from recorded music to live performances, we analyze the inverse link between both markets—i.e., the live-induced increased consumption of recorded music. In particular we analyze how playing live in a festival increases YouTube video searches, that we use as a proxy for the streaming of video content. YouTube, as the most used online music video content channel and a relevant source of information, plays a predominant promotional role for artists. In addition, with the fall of profits in the music industry, musicians draw on this channel as a revenue-generating source—from music-videos playbacks—as well as for developing new income streams (Edmond 2014 ).

This paper contributes to the understanding of cross-market effects in the creative industries, where interactions emerge between the physical and digital realms. It does so by reassessing the (more traditional or pre-filesharing) causal link between live performances and recorded music. Results show that playing live has a positive effect in a related market, that of streaming recorded content. To identify demand external effects on related markets with a clear causal direction is in itself non-trivial. To the best of our knowledge, there is no previous attempt to analyze the indirect effect of playing live to video streaming. Moreover, the variety of robustness checks we use provide strong support of the hypothesis of spillover effects from live to recorded music use.

The paper is organized as follows. We start by reviewing the literature on cross-market effects in the music sectors. Next, we discuss the empirical method and the dataset we use to test the main research hypothesis, i.e., the positive spillover effect from live performances on recorded music. Then, results are presented along with some robustness checks. Finally, we close with a detailed discussion of the results and some conclusions.

2 Theoretical background

The impact between different business models in the music industry has been extensively assessed in the economics and management literature, with a focus on cross-demand effects emerging from the complementary or substitutive nature of marketed products and/or services. Contributions could be loosely classified according to whether they analyze within-market or cross-market effects. The former, i.e., effects are restricted to one market, deal mostly, albeit not exclusively, with the impact of new consumption ways of recorded music within the recorded music industry.

2.1 Within-market effects in recorded music

Considering within-market effects in recorded music, Hendricks and Sorensen ( 2009 ) study the effect of product discovery in consumer demand. The authors analyze and find evidence supporting backward spillovers, that is, a positive effect of the release of a new album on an artist’s back-catalog. This effect stems from the enhanced diffusion of information between market participants associated with new albums, specially if these turn out to be a hit.

The emergence of legal online channels for digital music distribution has been found to have a negative effect on physical sales (Koh et al. 2014 ), as they reinforce the substitution of physical music by (legal) digital downloads and the trend toward unbundling—selling of individual tracks instead of full albums—(Elberse 2010 ; Papies and van Heerde 2017 ; Koh et al. 2019 ).

Streaming, as the main distribution channel of recorded music, has also attracted the interest of the academia. One early topic was that of the impact of streaming music services on other channels of recorded music distribution, i.e., the displacement of sales or cannibalization hypothesis. Nguyen et al. ( 2014 ) find that free streaming has no impact on CD sales. Wlömert and Papies ( 2016 ) use a panel of consumers to measure the impact of free (ads-based) and paid streaming services on music expenditures. They find that cannibalization does indeed occur, but its net effect is dependent on the type of service (whether it is free or fee-based) and consumer (active/inactive).

More recently, Aguiar and Waldfogel ( 2018 ) arrive to somehow conflicting results. From a market aggregate approach, they find that streaming displaces both permanent music downloads and music piracy. However, using a song-approach results are reversed: streaming increases sales and piracy. The latter, the authors claim, might be explained either to streaming actually increasing demand or to the presence of unobserved heterogeneity (which could invalidate the result).

Furthermore, the analysis of streaming platforms as new intermediaries in music (what has been labeled as the process of platformization of music, see Prey et al. 2022 ) stresses their curatorial power through playlists. These, in turn, contribute to the unbundling (and rebundling) of music (Bonini and Gandini 2019 ) with a potential deepening of substitution effects.

2.2 Cross-market effects

When it comes to cross-market effects, authors have focused on the connection between live and recorded music markets. Krueger ( 2005 ) sets the agenda by stressing the decoupling between live and recorded music: the inflation in live tickets, the author claims, is a consequence of live music (in a post-filesharing world) being no longer a means of promoting recorded sales. In other words, a weakened complementary relation between live and recorded music weakens the magnitude of cross-market effects.

These findings open a new avenue of research in which the search for evidence goes in the opposite direction: from recorded to live music. In this new context recorded music would be an additional promotion means of live performances. Montoro-Pons and Cuadrado-García ( 2011 ) report evidence of an asymmetric cross-effect with the consumption of recorded music increasing the likelihood of live music attendance but not the other way around. In general, this line of research underscores the connection of media-based consumption with the increased demand for live performances. Mortimer et al. ( 2012 ) find a positive effect of file-sharing on live music (an effect that is more pronounced for smaller artists who see a relatively larger rise in awareness). Moreover, Nguyen et al. ( 2014 ) show that free music streaming affects positively live attendance.

Papies and van Heerde ( 2017 ) introduce a model to test diverse cross-format effects conditional on music/artists attributes and other market-related traits. While their model predicts a positive relation between concert revenue and record revenue and vice versa, these effects are mediated by different control variables. Overall they find a stronger effect of recorded music consumption on concert revenue than the inverse. Moreover their results turn out to be negative when certain mediators (such as piracy or unbundling) are considered, which suggests that the evidence supporting the connection between live and recorded music is, at best, ambiguous.

Christensen ( 2022 ) uses a natural experiment-like setting to evaluate the impact institutional disruptions have on live music demand. By drawing on the temporary removal in 2009 of Warner catalog on YouTube, the author finds a negative effect on live revenue for Warner acts. This suggests a complementarity in a performer’s ticket prices and streaming penetration.

So far, the literature reviewed underscores the impact of digital channels on live concerts. However, and as new business models settle, it may be reasonable to reassess the impact of live concerts on recorded music. By following this line, one acknowledges a transient decoupling between both markets, as new businesses and consumer practices develop. Here, the recent literature provides scant evidence.

Using aggregate data, Maasø ( 2018 ) analyzes the impact of a festival on music streaming. The authors leverage the data in WiMP Music (a subscription-based music streaming service that later became Tidal ) to analyze a time series of streams before, during and after the 2010–2013 editions of the Øya festival in Norway. Findings support a 43% increase in streams of artists playing in the festival. However, the descriptive nature of the methodology falls short of determining whether this finding could be attributed to causality. Specifically, the fact that the streaming platform curates playlists associated with the festival lineup could cast doubts on the direction of the observed effect.

Drawing on individual-level data from the music website Last.fm Ternovski and Yasseri ( 2020 ) aim at identifying the direct and indirect impact of live attendance on streaming content. The authors extract those live events between 2013 and 2014 documented on Last.fm website and attendance as reported by users. Furthermore, they leverage data on friends of users who did not attend the live events to test the existence of peer or contagion (indirect) effects. A regression discontinuity design is applied to the time series of songs listened to by attenders three weeks before and after the event. Results show an increase of (roughly) 1 song per day per attendee which is consistent with the existence of a cross-market effect. As for the non-attenders, the effect is on average smaller, only found for the most popular artists and its magnitude depends on the number of Last.fm friends who actually attended.

Recently, Sim et al. ( 2022 ) analyze the negative impact of COVID-19 on music streaming. The authors conclude that restrictions in mobility implemented worldwide led to a substantial decrease in complementary activities (commuting or driving to cite two) to streaming music. While this finding is robustly validated, the hypothesis put forward in this paper—the causal link between live performances have on recorded music—could provide an additional explanation. As live performances were canceled due to lockdowns, a potential complementarity connection between live and streaming music was blocked.

Given that the balance of the music industry has shifted from the recorded music to live performances, it remains a challenge to identify ways in which the latter affects the former. This paper aims at further exploring this issue and testing the hypothesis of a positive impact of live performances on recorded music. To do so we benefit from an empirical approach that draws on the (online) user-generated data before and after a live event. A dataset composed of time series of a video search index across a sample of artists is used to test the existence of spillover effects from live to recorded music. Results agree with a positive impact of playing live with video search activity which is, on the other hand, consistent with an increase in video streaming. Our findings are robust and have a local and simple causal interpretation.

This paper aligns with that of Maasø ( 2018 ) in that aggregate data is used. Furthermore, and given the data source, estimates are consistent with joint evidence of direct and contagion effects—that is, searches performed by both attenders and non-attenders. However, the methodology is that used by Ternovski and Yasseri ( 2020 ) which, under specific conditions, allows to infer causal relationships from observational data. Nevertheless, we depart from them in that we consider a social media platform with a larger user base: YouTube ranks second (to Facebook) in number of users. Footnote 1

3 Methods and dataset

3.1 regression discontinuity designs, live performances and youtube video searches.

To quantify the (positive) spillover effects that live performances have on video streaming one must draw on causal inference methods. While randomized experiments provide the gold standard to measure the average effect of an intervention (playing live) on a response variable (video streaming), they are infeasible in many applications in the social sciences. If this is the case, researchers rely on observational data which, in turn, preclude the estimation of the kind of causal effect we are interested in unless some strict conditions are met or some additional information (such as the case of valid instruments) is provided (see Winship and Morgan 1999 ).

Regression discontinuity (RD) designs offer a quasi-experimental setup for making causal inferences using observational data. There is a growing literature using RD as one of the strongest non-randomized designs to estimate causal effects (see the survey by Villamizar-Villegas et al. 2021 ). Indeed, RD has been considered a valid tool to identify treatment effects with observational data sharing some properties of randomized trials (Angrist and Pischke 2009 ; Lee and Lemieux 2010 ; Imbens and Kalyanaraman 2012 ; Choi and Lee 2017 ).

We apply an RD design in the context of a sample of performers playing live to detect the influence, if any, on YouTube users’ behavior. In this case, the framework is that of an RD design in time, i.e., the running variable is time to an event, and we aim to estimate the causal impact of live performances on video streams. Note that we will not deal directly with video streams but use instead a YouTube searches index as a proxy.

The design works as follows: music video searches— \(Y_{i,t}\) , the outcome variable—are observed for performer i over a period of time. Having played live is the treatment status: at each time point performers have either played live ( \(D_{i,t}=1\) ) or not ( \(D_{i,t}=0\) ). The running variable \(S_{i,t}\) —time in weeks to the performance—determines whether unit i has undergone treatment, i.e., it splits units into treated and untreated at point t . Note the score \(S_{i,t}\) is normalized to point to the intervention at the cutoff \(S_{i,t}=t_i-c\) , being \(t_i\) time (week) at which outcome is observed and c the week at which performance occurs. Therefore it takes negative values ( \(S_{i,t}<0\) ) before a band plays live and positive ones ( \(S_{i,t}>0\) ) afterward. Finally, the value of the treatment \(D_{i,t}\) is univocally affected by the running variable \(S_{i,t}\) , time to the event.

Define \(\tau \) , the average treatment effect, as the parameter of interest. Then, we propose the following specification

being \(f(\cdot )\) being a polynomial functional which is specified such that the regression function differs on both sides ( L and R ) of the cutoff:

While the simplest way to implement RD is to fit the outcome on the score, we use artist-related covariates \(X_{i,t}\) to improve efficiency. Then expression ( 1 ) turns into:

Note that a local regression for expression ( 2 ) is proposed: only observations within a neighborhood (i.e., a bandwidth) of the cutoff are used to estimate \(\tau \) . Furthermore, a kernel is employed, so observations closer to the cutoff receive a greater weight.

All the analysis rests on the assumption that YouTube searches change abruptly due to performing live. Thus, the identifying assumption underlying the discontinuity approach is that in the absence of the event, YouTube searches related to bands playing would have changed smoothly. Empirical evidence will be consistent with the existence of spillover effect if the outcome is observed to discretely change at some threshold value of the score. Namely, is there a discrete jump in video searches once a band has performed live?

Furthermore, the appropriateness of an RD approach draws on the fact that its main assumptions hold. First, there is comparability of units around the cutoff. This might sound circular (Nick Cave is Nick Cave) but note that performers’ features evolve over time: released records, awareness, popularity and fan base... which may change how they are perceived by the public. Analyzing video searches on performers in a window around the cutoff (when the festival takes place), allows us to assume artists’ features remain unchanged and comparability applies. Second, there is no manipulation of the running variable. This is obvious as individuals do not affect dates of festivals which are set in advance well before the arrangement of the lineup takes place. Third, continuity of the relationship between the outcome and the running variable is assumed. It is reasonable to posit that once we remove the treatment (i.e., the live performance) there should not be a discontinuity in the relationship between outcome (YouTube searches) and running variable (time to performance).

Note that the local nature of estimates limits their external validity. However, this sits well with the analysis of the music market, where short life cycle and oversupply make the expected effects to be transitory. In other words increased video searches occur in a short window after the intervention takes place, i.e., searches are not permanently enhanced. However, this does not discard other permanent effects for the performer (e.g., performances could lead to an increase of the fan base).

Finally, RD designs where the score or running variable is time do have particularities that set them apart from the standard RD design. Note that, contrary to standard RD designs where the running variable splits subjects intro treated and non-treated, no randomization is at play as eventually all units (performers) become treated. Furthermore, sorting of the units around the cutoff has to be discarded. In our case one could consider the event occurs as if randomly as individuals cannot manipulate the date at which the event takes place—see Ternovski and Yasseri ( 2020 ) for a detailed discussion. The literature offers many examples of designs in which time is a running variable (Anderson 2014 ; Bahrs and Schumann 2020 ; Chenyihsu and Whalley 2012 ; Gallego et al. 2013 ). Besides, Hausman and Rapson ( 2018 ) offer a constructive critic and checklist on this type of designs. Additionally, we have resorted to enhanced robustness checks.

3.2 The dataset

The dataset herein analyzed collects information on YouTube searches for performers in the 2016, 2017, 2018 and 2019 editions of two reputed music festivals in Spain: Primavera Sound and Festival Internacional de Benic à ssim ( FIB ). Choosing bands performing at two well established events simplified (and removed arbitrariness in) the decision on sampling units. Myriad of bands perform live in Spain every year and their heterogeneity is difficult to capture. By choosing bands performing at leading festivals, we assume they have achieved a certain reputation or have a reasonable expectation to become so. Furthermore, our empirical strategy draws on bands that generate enough Internet activity, something that is easier to achieve for bands associated with top ranking festivals. Indeed, both events were listed in the top ten music festivals in Spain using data available from the website of the Spanish association for music promoters. Footnote 2 In addition, they offer a varied lineup proposal allowing us to capture the diversity of genres and styles of popular music and a reasonable mix of well-established and newer acts.

Prior to describing the dataset, we provide some details about the steps involved in the data collection process. First, we start by identifying the lineup of Primavera Sound and FIB in the selected years through web scraping. Ideally, for each band in the lineup a weekly index of YouTube searches is to be retrieved using Google Trends Footnote 3 (GT). GT’s main functionality is that of providing a time series index of the volume of queries of any given search term in a specific geography. In our case, we used video searches in Spain. The index is produced from an unbiased sample of video search data over a week that is further translated into a normalized 0–100 scale. In this case the week with the maximum volume of queries is set as 100 and all other weeks are given relative weights. Data gathering was automated through the use of a Python application programming interface for GT. Footnote 4

Second, the retrieval of information draws on the use of a search term which, in some instances, can be too vague or general. In order to disambiguate we resorted to GT’s suggestions , a functionality that allows to select queries that unambiguously refer to a specific search term (in our case a performer’s name). Suggestions’ output is a (set of) key(s) or identifier(s). Each one links the different specific meanings related to a search term. Footnote 5 In other words, each outcome comes with a set of identifiers that allow to explicitly refine the retrieved search. After inspection of the output only those identifiers that refer to artists/bands/performers are retained. Based on these, the sample was thus restricted to 246 bands for which a valid suggestion is generated.

Third, while theoretically the index can be retrieved at different frequencies (months, weeks, days and hours) in practical terms high-frequency data are only adequate for queries that generate a large amount of searches. On the contrary, search terms that do not generate enough searches in specific points in time are automatically assigned an index value of zero, which creates a downward-bias problem. To minimize this risk and, on the other hand, to be able to identify the time evolution of the index with a reasonable detail, we opt to retrieve the index for weekly data. Again, choosing a more precise time period, i.e., daily data, would mean many observations not generating enough searches, hence being assigned a value equal to zero. Once we retrieved the preliminary dataset we dropped those performers with more than 50% of zeroes for the index values in a window around the performance date of 8 weeks. This further reduced the sample to 190 performers.

Fourth, for each performer, roughly five years worth of search data are collected: each time series spans a period that starts 253 weeks before the festival takes place and ends 6 weeks after it. However note that the local nature of the RD design estimates implies that only those observations around the cutoff date (i.e., the live performance) will be used. Nevertheless, we make use of observations further apart from the cutoff for robustness checks.

Finally, the RD design’s estimation procedure will make use of additional (band-level) covariates to enhance the efficiency of the estimates (Calonico et al. 2019 ). Specifically, the number of years since first album was released (and its square) and genre of the band (a dummy which takes on value 1 if the band is classified as pop or rock in the website https://www.pitchfork.com ) are included. Moreover, a trend line is added to control for time patterns. Table  1 summarizes some of the features of the sample. It includes number of bands, average year of first release and most frequent genre-distribution. Summaries are provided for the full sample as well as for the individual festivals.

4.1 Selecting the running variable

Before introducing the results, we need to clarify some design choices regarding the running variable. The structure of the dataset is a panel following bands/performers over time. It contains each performer’s YouTube weekly search index (the outcome), a set of covariates, and a transformation of the time variable (the running variable). Note that events (performances) take place in a specific point in time (a day) that differs from the scale of the time-variable we use (weekly aggregate data on searches). This implies some design decisions with regard to the running variable.

Performances take place within events, i.e., festivals. For each festival edition, the start and end dates are known. In this regard, we need to make some clarifications regarding the specific time stamp of the video search index. First, the weekly search index collects information of the week where the starting point is always Sunday. For instance, the index at date 2019-09-01 refers to YouTube searches in the week starting Sunday September 1 and ending Saturday September 7 (year 2019). This means the index summarizes forward-looking searches at the specific date it points.

Second, for all performers we collect the start and end dates of the festival edition they play at. These dates are then used to build the running variable. However, the use of weekly data introduces an ambiguity to the choice of cutoff date. Choosing the end of the festival as cutoff (always on Sunday) means the index at the cutoff retrieves video searches on the week following the performance. Alternatively, choosing the start of the festival (on Wednesday or Thursday) the cutoff would point to somewhere in between the actual start and the week after the festival. To actually retrieve information on the week a performance takes place we need to move the pointer to the Sunday before the Wednesday/Thursday the festival starts. By shifting it slightly backward, treated observations are always located to the right of the cutoff. In practice we use all the three cutoffs to check the robustness of the estimates.

As a graphical aid to identify the existence of an effect of playing live on video searches, Fig.  1 shows the discontinuity plots using the three running variables. Data are binned across the running variable and means for the observations within each bin are computed (dots in the plots). In this case, the number of bins was data-driven, Footnote 6 and the resulting smoothed local regression line is shown. A discrete jump in the cutoff date becomes apparent in the three plots, which becomes more evident when using week of the festival as the score. The next subsection provides estimates for this observed effect.

Regression discontinuity plots for different adjustments of the running variable. Cutoff point is: start of the festival (top); end of the festival (middle); week of the festival (bottom)

4.2 Estimating the effect

The R package rdrobust (Calonico et al. 2021 ) has been used to estimate the effect through a nonparametric local polynomial approach. In this respect, there are two choices to make: the kernel (weights) used in the estimation and the bandwidth that controls the neighborhood around the cutoff that is to be used in the estimation. As for the weights, a triangular kernel has been employed, although results are robust to alternative kernels. With regard to the bandwidth, we resort to a data-driven approach (Calonico et al. 2016 ) by choosing a bandwidth that minimizes the mean square error (MSE) of the RD point estimator. We allow bandwidth to differ on both sides of the cutoff given the asymmetric length of the time series. However, results remain qualitatively unchanged when the same bandwidth is used on both sides. In addition, results were robust to changes in the pre-event window of the time series. Footnote 7

Furthermore, we control for non-independent residuals due to the panel structure of the data by clustering the observations by performer. In addition, the efficiency of the estimator can be increased by introducing covariates. Specifically we control for: (i) time (in years) since a band/performer released its first recording (and its square); (iii) genre of the performer through a binary variable to indicate whether a performer is classified as pop/rock in pitchfork.com; and (iii) a trend to account for the dynamics of the data. Note none of these covariates are expected to differ at the cutoff; hence, they are balanced and further (restrictive) assumptions are not needed.

Table  2 summarizes the main results. In addition to conventional 95% confidence interval estimators—based on the MSE-optimal point estimator—it provides bias-corrected confidence intervals. Footnote 8 Note that inference based on the conventional (and optimal) point estimator is not desirable as it assumes the local polynomial provides an exact approximation to the true underlying regression function. Robust (bias-corrected) estimates lead to improved coverage in finite samples and result in valid inferences when optimal MSE-bandwidth is used.

Overall, results provide evidence consistent with the effect of playing live on the index of video searches, which experiences an increase ranging from 8.12 to 10.73 points. While the magnitude of the effect is affected by the choice of the cutoff, robust bias-corrected intervals suggest a positive treatment effect in all cases. Note these results are also found when pooled data are used (without clustering of observations) and no covariates are included in the model. Graphically, Fig.  2 plots the RD estimates in the three different scenarios and provides a visual account of the positive jump (i.e., the discontinuity) in video searches at the time performances take place. This, in turn, delivers visual evidence of spillover effects emerging from live performances to the videostreaming of recorded music.

4.3 Conditional effects: genre and gender of performer

Next, we analyze the treatment effect conditioning on specific covariates of interest. In particular we estimate the effect across: (i) genre of performer and (ii) gender of the performer. The goal of this exercise is to identify (if any) potential differences in audiences response to performers when looking at two classification criteria.

Genres can be seen as institutional devices that organize supply and demand in cultural markets. They help position artists and, in so doing, provide signals to potential audiences and help set expectations to market participants (Lena and Peterson 2008 ). Interestingly, genres have been found predictors of commercial success (Askin and Mauskapf 2017 ), which makes them a good candidate to be used as a source of differential spillover effects. Without imposing any a priori assumption on how genre mediates spillovers, and given the restricted variability in music genres in the sample we use a binary classification were we subset the sample into performers classified as either Rock or Pop and the remainder. Roughly 66% of the performers fall into the pop/rock class.

Estimates of the discontinuity effects using a local polynomial regression

The approach to gender is more contentious, and results are to be considered of a more exploratory nature. Again, we expect it to influence supply and demand in the music sectors. Considering demand, the literature on cultural consumption has found relative support for different headline participation rates across genders. Furthermore differences have been found when accounting for high-brow/low-brow participation. For instance, women have been found to have lower participation rates for both live and prerecorded music (Montoro-Pons and Cuadrado-García 2011 ) and to show a tendency for being omnivorous—i.e., showing a higher propensity to participate in both high-brow and low-brow manifestations (Favaro and Frateschi 2007 ).

On the other hand, and to the best of our knowledge, no study on the role of gender on spillovers in cultural supply is to be found in the literature. Nevertheless, the scant presence of female performers in the music industry has been brought to the center of the debate in recent years. A study in the UK found that 70% of performers of lineups of festivals in 2017 were male or all-male bands (Sherlock and Bradshaw 2017 ). In a similar vein Mitchum and García-Olano ( 2018 ) find that in 2018 70% of the artists across the largest US festivals were male (down from 75% in 2017). As a response, an initiative from within the industry aiming at achieving gender-balance is gaining momentum with a few relevant players (such as Primavera Sound starting in its 2019 edition) already sticking to a 50:50 gender split. Footnote 9 Do these gender differences translate into asymmetric effects? To analyze the role of gender we condition the sample on the gender of the performer or (if a band) the genre of the individual who fronts the band.

Results for both exercises are shown in Tables 3 and  4 . With regard to genre, estimated effects are significant and found to be greater for pop/rock performers who experience a 14 points increase in the search index compared to 8.61 by other genres . This finding is consistent with the hypothesis of a differential impact of genres, with pop/rock having a broader appeal to the general public. In addition, female performers (or bands fronted by females) generate more video streaming activity (increase in 14.6 points) than non-female ones (11.3 points). Both results suggest some interesting lines of work. Findings indicate a clear match between pop/rock and female performers and YouTube video streaming. In the case of genre, we find support to the hypothesis of the broader appeal of pop to audiences. As for female acts, the fact they could generate more online activity than male performers opens an interesting research avenue worth of exploring.

4.4 Conditional effects: popularity of the performer

One could question whether the observed effects are asymmetric across different levels of popularity of performers. Theoretically, the expected conditional effect is ambiguous. As noted in the literature review, evidence has been found of spillovers from file-sharing to live demand being stronger for less successful artists, as they could benefit most from the raised awareness (Mortimer et al. 2012 ). Alternatively, one might argue that most popular performers should create more buzz and therefore generate more online activity generated after a performance.

To provide evidence on this research question, we recover the effect of live performances on video searches conditional on the popularity of the performer. In this regard, we must first define a measure of success or popularity. Note that success could be measured on different dimensions. We have come up with a feasible and simple alternative: using the number of results a Google search for a performer returns. Footnote 10 The rationale is simple: the more popular an artist is, the more search results is should produce. Footnote 11

Using the distribution of search results one can classify the sample across different quantiles. In particular we used three different classifications: (1) top 25% performers vs. bottom 75%; (2) top 50% vs. bottom 50%; and (3) top 33% vs. bottom 66%.

Table  5 shows the findings from this exercise. Results do not provide strong evidence of asymmetric effects. Indeed neither sample split has produced remarkable differences regarding the magnitude of the estimated spillovers. While marginally larger point estimates were found for more popular acts in the 25–75 and 33–66 splits, these results were reversed when partitioning the sample using the median. Anyhow, as it can be seen, differences in magnitude are rather thin.

5 Validation of the results

The literature on RD designs provides alternative strategies to validate empirical results (Thoemmes et al. 2017 ). In our case, different approaches have been undertaken.

First, a placebo cutoff strategy has been followed. Different artificial cutoffs have been selected from roughly one year (52 weeks) to one month (4 weeks) before the actual performance takes place. Note that choosing arbitrary cutoffs should produce non-significant estimates and therefore support the estimated effects in Table  2 . Conversely, if estimated effects for this artificial interventions were significant, they would challenge the reliability of the results.

As for the rationale of the specific artificial cutoffs chosen, note that end of spring-beginning of summer marks the start of the tour season for many performers and that evidence of a seasonal pattern in live music consumption has been consistently found in the literature (Montoro-Pons and Cuadrado-García 2016 ). Could this drive a general interest in music regardless of the specific performers we choose? If so, one could find a peak in the public interest/awareness in music to be coupled with music video searches happening around the start of summer—roughly 52 weeks before the live performances considered—regardless of whether actual live performances are taking place. As for the other placebo cutoffs—26, 12 and 4 weeks before the event—they might be capturing announcement effects, promotion efforts by the festivals and/or anticipation to the actual event by consumers.

Regression discontinuity plots with artificial cutoffs at roughly) one year (top left), six months (top right), three months (bottom left) and one month (bottom right) before actual cutoff

Table  6 shows the results for different placebo cutoffs when using week of the festival as running variable, while Fig.  3 displays the RD plots. Setting the cutoff one year to one month before the actual intervention takes place we cannot conclude with significant effects using the standard 5% significance level. In all cases, the effect is indistinguishable from zero as the estimated confidence intervals (both conventional and robust) cross the zero line. Furthermore, the small magnitude of the estimated effects adds to the evidence supporting the findings in the previous section.

An additional placebo test was run. In this case, we want to rule out alternative explanations why a performer should see an increase in YouTube searches. Here, we could find performer-specific characteristics, such as new album or video releases, or other external factors such as seasonal confounding that makes audiences more prone to music in general.

To do so we have collected data on the same sample of performers for the same time periods but changing the geography from which searches were run: in this exercise we use YouTube searches in the US. If the estimated effects are causal, then results of this exercise should be non-significant. Table  7 shows the resulting effects for the alternative cutoffs. In it, no effect is significantly different from zero which reassures the existence of a causal effect from playing live.

Second, we have analyzed the sensitivity of the solutions to bandwidth choice. Results are shown in table  8 using week of the festival as the running variable. Besides the MSE-optimal bandwidth we have estimated a coverage error (CER) optimal bandwidth (Calonico et al. 2019 ). For each one, symmetric—equal width to the left and right of the cutoff—and non-symmetric bandwidths have been used. Furthermore, twice the optimal bandwidths were used. As it can be seen, all results are qualitatively similar and produce estimated effects that are robust to bandwidth specification.

Finally, to validate the results we run some tests to check the density of the running variable and used predetermined covariates as outcomes, i.e., placebo outcomes. All supported the robustness of the findings and are not included for the sake of brevity and given these tests (used in the literature) are less meaningful in this specific case. The nature of the running variable (time) implies no possibility of sorting around the cutoff; hence, its density should be balanced around the threshold. As for the continuous covariates (years since first release and a trend variable), unsurprisingly they did not show a discontinuity at cutoff.

6 Discussion of the results

Business model innovations have deeply changed value creation in the music industry, altering traditional revenue streams and facilitating new ones. Such is the case of offline/online cross-market effects. This paper has tested the hypothesis of spillover effects going from live music to prerecorded music (video streaming through an online platform). To do so, an index of YouTube video searches is analyzed for a sample of performers at different editions of two established music festivals in Spain.

Results support the cross-market effect hypothesis: after playing live the index of video searches increases by roughly 9–10 points with the robust confidence intervals being in most cases well above this amount (see Table  2 ). Findings are, by and large, consistent with the hypothesis of positive spillover effects from live to recorded music. This, we believe, is a non-trivial contribution in itself as the evidence so far about this type of effects is scant and/or inconclusive. We add to recent evidence identifying the causality link from live to recorded music—which sets this work apart from Maasø ( 2018 ). Besides, we use recent data on video streams from one of the largest social media platforms and analyze performer traits that could condition video streams, namely genre and gender.

Overall, findings support the connection between the two markets (live and recorded), a link that can be transformed into a revenue source for performers. This stems from the fact that findings refer to YouTube videos, which artists do monetize directly—as they are, at least partially, right holders—but also potentially in an indirect way—through an increased awareness that can help raise artists’ profiles, expand their fan base and increase other revenue-generating consumer-related activities such as sound streaming.

From a methodological standpoint, results were reached through an RD design, which can be seen as a quasi-experimental method that, provided some conditions are met, produces valid causal inferences. In this research, findings provide evidence of a causal effect from live to prerecorded music. The fact that, after plotting the dataset, a discontinuity in the outcome emerges around the time the festivals take place (i.e., artists perform live) is suggestive of the adequacy of the RD design and provides a powerful graphical tool to illustrate the results. However, given in an RD design in time all units eventually become treated, a thorough set of robustness tests have been run which further have supported the results.

6.1 The economic value of (increased) YouTube searches

One could question whether the economic value brought about by increased YouTube searches is of relevance. In this regard, we must note that the outcome in this design is an index (not the number) of YouTube searches. It measures searches relative to its maximum value (where it achieves a value of 100) over a period of time. Therefore, for the sample of performers, we are using a relative measure of searches and not their actual number (something that GT does not disclose). Having said so, an increase in the search index means a jump in actual searches (whatever this value may be). Then, assuming most searches are goal oriented, that is, one searches a video to play it back, the number of actual video streams must also jump. The question at stake is that of the economic value of searches.

First, one should note there is a direct revenue stream that emerges from video playbacks through ad and subscription revenue sharing. This is what we could identify as the short-term effect. In this respect, Soha and McDowell ( 2016 ) state that “YouTube splits ad revenue with the rights holder [...] somewhere between 40% and 50%”. Unfortunately, it is not simple to come up with a figure for the cost per 1000 views a YouTube video generates: it depends among other things on country of origin of the view and type of video. Anyhow as with any other revenue source from recorded music, that amount is split between music label, songwriters and publishers.

Furthermore, one must note that we are dealing with searches (and not actual playbacks) that can potentially generate multiple streams over time. For instance if a user after running a search includes a video in a playlist or replays it through their playback history. In this case the actual effect could last longer (and its economic value be greater) than the short-term impact we measured through search activity. In other words, while we have found a transient effect on search activity after a performance, the streaming effect could extend for a longer time period.

Second, besides the economic value associated with video streams, one should consider additional longer-term effects. We have already mentioned an increased fan base, an effect that builds up from live performances and which is not restricted to actual audiences but expands beyond these (through a contagion effect) through the media and social networks enabling discovery and increased awareness making bands better known to the public and reducing barriers they face to reach audiences. Here, we must point out the unconstrained nature of the data (geography includes all YouTube searches in Spain), implying we are looking at the amplified effect that a performance has on potential audience including attendees (the actual audience) who are positively affected through a direct exposure to the event and non-attendees (or indirect audiences) who are affected through exposure through traditional media (TV, newspapers and the so) plus social networks and eWOM. As Lee and Kim ( 2022 ) note, success in recorded music involves a mix of an artist’s talent, the influence of record labels or platforms and the influence of all stakeholders, especially consumers. This refers to the process of content production, distribution and consumption, where consumers influence through user-generated content such as e-WOM.

Additionally, one must note that for those performers signed to a label, a boost in YouTube playback activity is linked to increasing copyright claims (for performers but, especially, for record labels). This, in turn, raises the market value of the creators (band/performer) as an asset in the portfolio of bands of the labels that sign them, which, in turn, could sustain or boost a career with the record label.

All in all, it is reasonable to assume increased searches set in motion both short-term effects—video streams that go beyond the time frame in which the increase in searches is observed—and longer-term effects—namely audience enlargement through discovery, increased awareness and eWOM, or increased band value to record label and/or platforms.

6.2 The gender effect

Furthermore, results conditional on genre and gender of the performer were obtained. As for genre, results are in line with estimations by Askin and Mauskapf ( 2017 ), who find that pop songs reached higher position and stayed longer on charts. Translated into our framework, this broader appeal of pop carries a greater spillover effect.

More interestingly the gender effect opens a potentially fruitful line of research. Indeed, our findings are consistent with a gender bias gap, worthy of further research. However, given the nature of the data set at hand, these results are to be considered of an exploratory nature and as such rather preliminary.

Three alternative explanations are consistent with the observed evidence. On the one hand, it could be a case of discovery and visibility at play where live events help increase the public awareness of women musicians more than that of their male counterparts. The fact that women are systematically underrepresented in the typical festival lineup—a situation which is not much different from those found in other labor settings (Lutter 2015 )—could make audiences learning about female acts more impactful. Lack of visibility and recognition of women in music constrains female musicians’ opportunities to showcase their talent and reach wider audiences. Once the right conditions are set, e.g., playing to audiences of reputed events, awareness is triggered. In addition, institutional attempts at leveling the field could also add to the impact by bringing women performers to the front of lineups.

Anecdotal evidence suggests discovery to be associated with female musicians in other music industry environments. One instance is that of gender differences at the Grammy Awards. Overall, from 2013 to 2021 the total number of Grammy male nominees was 86.6%, while that of female nominees was 13.4%. Furthermore, focusing on different award categories the percentage of male nominees is always greater than that of female nominees except for best new artist , where shares are more balanced- \(-\) 54.55% male nominations compared to 45.45% female nominations (Smith et al. 2020 ).

Alternatively, this underrepresentation could imply that (on average) better known female artists are included in lineups bearing a greater impact on audiences. To address this biased effect, one should control for some metric of the popularity of performers. Looking at the number of results after running a Google search provides some interesting insights. Figure  4 shows the distribution of success (in a log scale) for female and male performers. Overall, it appears that the range of popularity for male performers is greater than that for female ones. We also see that success quantiles (as reference lines in the plots) indicate that male performers are more popular across the whole distribution. On the other hand, popularity is less concentrated for females than for males.

We further investigate this by looking at the percentage of female/male performers across the distribution of popularity. Figure  5 shows female(male) performers in each quartile(tercile) of the popularity distribution as a share of the total number of female(male) performers. It complements Fig. 4 in that it shows the greater concentration in the top quartile(tercile) for female artists. Now, it becomes apparent that the share of top popular female artists is greater than that of male artists.

Note that we have not find strong evidence supporting a greater effect for more popular artists. If this gender bias is related to success, it should enter the model as an interaction of popularity and gender (i.e., being female). However, note this would reduce the sample for this specific group making the estimates rather noisy, which advises against it.

Distribution of popularity across gender. Both plots include reference lines for the first quartile (bottom), the median (center), and the third quartile (top)

Finally, a third potential explanation could be related to the nature of the streamed materials. The literature has acknowledged differences in gender in relation to performers, specifically in music videos (Conrad et al. 2009 ; Seidman 1992 ) as they often reproduce gender stereotypes—which tend to be genre-related (see Aubrey and Frisby 2011 ; Weitzer and Kubrin 2009 ). Specifically, female artists are more sexually objectified compared to male artists (Wallis 2011 ). The literature seems to agree that the content offered in music videos is different depending on the artist’s gender. Thus, it could be conjectured that the streaming of music videos is also biased by these differences.

Percentage of performers in each quartile (left) and tercile (right) across gender

6.3 Limitations

Finally, with regard to the limitations of this study, mostly related to the empirical strategy, we would like to mention three. First, and with regard to the sample—i.e., performers playing at two established music festivals—we must acknowledge that results might not be unambiguously generalized. The type of cross-market effects we have measured need some scale and the creation of buzz, exposure and contagion through media (traditional and otherwise). This, in turn, limits the validity of the results to those contexts similar to the one analyzed in this paper. However, in this respect one must note there is a significant variability in the sample of performers analyzed (small acts, middle class performers and big stars) that becomes apparent when looking at the diverse lineups of the editions of the festivals considered.

Second, data on the outcome (i.e., the search index) are obtained as an unbiased random sample of video searches. This means the actual values we obtain change every time a retrieval request is made. These changes need not be important if artists generate enough Internet activity. However, one should acknowledge that the outcome variable is subject to sampling error that could affect the estimates most likely by driving them down.

Third, related to the previous point, by choosing weekly data we minimized the problem of zeroes for performers who generate low search activity. However, it creates an inherent ambiguity in the choice of the running variable (week to event) as the events always span over two values of the running variable. We tried to overcome this by using different definitions of the running variable, which provided robust results.

Euromonitor International reports a monthly user base of over 20 million in Spain in 2020. Retrieved June 13, 2022 from https://www.euromonitor.com .

See https://www.apmusicales.com/ranking-festivales-2018/ . Accessed on July 5th, 2022.

See website https://trends.google.es .

Available at https://github.com/GeneralMills/pytrends .

For instance, a query for the term suede returns, among others, the suggestions “Suede: Topic”, “Suede: Shoe store in Rome, Italy” and “Suede: Band”. We are interested in the latter.

Determined by mimicking variance evenly-spaced method using spacing estimators.

Estimates remain qualitatively unchanged when restricting the window before the live performance to 8 weeks.

These are robust to the choice of bandwidths. See Calonico et al. ( 2014 ).

See the Keychange manifesto at https://www.keychange.eu/ .

It should be noted that any measure of success has a time-dependent component: if commercial success evolves over time, then any attempt to capture popularity should incorporate this fact. However, this (note that as performances take place over different years) along with the multidimensional nature of success complicates the task of finding a comprehensive metric. We consider our solution provides a reasonable approximation to success.

Searches were conducted on January 26 2023.

Aguiar L, Waldfogel J (2018) As streaming reaches flood stage, does it stimulate or depress music sales? Int J Ind Organ 57:278–307

Article   Google Scholar  

Anderson ML (2014) Subways, strikes, and slowdowns: the impacts of public transit on traffic congestion. Am Econ Rev 104(9):2763–2796

Angrist JD, Pischke J-S (2009) Mostly harmless econometrics: an empiricist’s companion. Princeton University Press, Princeton

Book   Google Scholar  

Askin N, Mauskapf M (2017) What makes popular culture popular? product features and optimal differentiation in music. Am Sociol Rev 82(5):910–944

Aubrey JS, Frisby CM (2011) Sexual objectification in music videos: a content analysis comparing gender and genre. Mass Commun Soc 14(4):475–501

Bahrs M, Schumann M (2020) Unlucky to be young? The long-term effects of school starting age on smoking behavior and health. J Popul Econ 33(2):555–600

Bonini T, Gandini A (2019) “First week is editorial, second week is algorithmic”: platform gatekeepers and the platformization of music curation. Soc Media Soc 5(4)

Calonico S, Cattaneo MD, Farrell MH, Titiunik R (2019) Regression discontinuity designs using covariates. Rev Econ Stat 101(3):442–451

Calonico S, Cattaneo MD, Farrell MH, Titiunik R (2021) rdrobust: Robust data-driven statistical inference in regression-discontinuity designs [Computer software manual]. Retrieved from https://CRAN.R-project.org/package=rdrobust (R package version 1.0.4)

Calonico S, Cattaneo MD, Titiunik R (2014) Robust nonparametric confidence intervals for regression-discontinuity designs. Econometrica 82(6):2295–2326

Calonico S, Cattaneo MD, Titiunik R (2016) Optimal data-driven regression discontinuity plots. J Am Stat Assoc 110(512):1753–1769

Chenyihsu Y, Whalley A (2012) Green infrastructure: the effects of urban rail transit on air quality. Am Econ J Econ Pol 4(1):58–97

Choi JY, Lee MJ (2017) Regression discontinuity: review with extensions. Stat Pap 58(4):1217–1246

Christensen F (2022) Streaming stimulates the live concert industry: evidence from YouTube. Int J Ind Organ 85:102873

Conrad K, Dixon TL, Zhang Y (2009) Controversial rap themes, gender portrayals and skin tone distortion: a content analysis of rap music videos. J Broadcast Electron Media 53(1):134–156

Edmond M (2014) Here we go again: music videos after YouTube. Television New Media 15(4):305–320

Elberse A (2010) Bye-bye bundles: the unbundling of music in digital channels. J Mark 74(3):107–123

Favaro D, Frateschi C (2007) A discrete choice model of consumption of cultural goods: the case of music. J Cult Econ 31(3):205–234

Gallego F, Montero JP, Salas C (2013) The effect of transport policies on car use: evidence from Latin American cities. J Public Econ 107:47–62

Hausman C, Rapson DS (2018) Regression discontinuity in time: considerations for empirical applications. Annu Rev Resour Econ 10(1):533–552

Hendricks K, Sorensen A (2009) Information and the skewness of music sales. J Polit Econ 117(2):324–369

Hesmondhalgh D (2021) Is music streaming bad for musicians? Problems of evidence and argument. New Media Soc 23(12):3593–3615

Hesmondhalgh D, Meier LM (2018) What the digitalisation of music tells us about capitalism, culture and the power of the information technology sector. Inf Commun Soc 21(11):1555–1570

Imbens G, Kalyanaraman K (2012) Optimal bandwidth choice for the regression discontinuity estimator. Rev Econ Stud 79(3):933–959

Koh B, Hann I-H, Raghunathan S (2019) Digitization of music: consumer adoption amidst piracy, unbundling, and rebundling. MIS Q 43(1):23–45

Koh B, Murthi B, Raghunathan S (2014) Shifting demand: online music piracy, physical music sales, and digital music sales. J Organ Comput Electron Commer 24(4):366–387

Krueger AB (2005) The economics of real superstars: the market for rock concerts in the material world. J Law Econ 23(1):1–30

Google Scholar  

Lee DS, Lemieux T (2010) Regression discontinuity designs in economics. J Econ Lit 20(1):281–355

Lee M, Kim H-J (2022) Exploring determinants of digital music success in South Korea. Electron Commerce Res, 1–22

Lena JC, Peterson RA (2008) Classification as culture: types and trajectories of music genres. Am Sociol Rev 73(5):697–718

Lutter M (2015) Do women suffer from network closure? The moderating effect of social capital on gender inequality in a project-based labor market, 1929 to 2010. Am Sociol Rev 80(2):329–358

Maasø A (2018) Music streaming, festivals, and the eventization of music. Pop Music Soc 41(2):154–175. https://doi.org/10.1080/03007766.2016.1231001

Mitchum R, García-Olano D (2018) Tracking the gender balance of this year’s music festival lineups. https://pitchfork.com/features/festival-report/tracking-the-gender-balance-of-this-years-music-festival-lineups/ . Accessed 14 June 2021

Montoro-Pons JD, Caballer-Tarazona M, Cuadrado-García M (2021) From pirates to subscribers: 20 years of music consumption research. Int J Consum Stud 45(4):690–718

Montoro-Pons JD, Cuadrado-García M (2011) Live and prerecorded popular music consumption. J Cult Econ 35(1):19–48

Montoro-Pons JD, Cuadrado-García M (2016) Unveiling latent demand in the cultural industries: an application to live music participation. Int J Arts Manag, 5–24

Mortimer JH, Nosko C, Sorensen A (2012) Supply responses to digital distribution: recorded music and live performances. Inf Econ Policy 24(1):3–14

Nguyen GD, Dejean S, Moreau F (2014) On the complementarity between online and offline music consumption: the case of free streaming. J Cult Econ 38(4):315–330

Papies D, van Heerde HJ (2017) The dynamic interplay between recorded music and live concerts: rhe role of piracy, unbundling, and artist characteristics. J Mark 81(4):67–87

Prey R, Esteve Del Valle M, Zwerwer L (2022) Platform pop: disentangling Spotify’s intermediary role in the music industry. Inf Commun Soc 25(1):74–92. https://doi.org/10.1080/1369118X.2020.1761859

Seidman SA (1992) Profile: an investigation of sex-role stereotyping in music videos. J Broadcast Electron Media 36(2):209–216

Sherlock P, Bradshaw P (2017) Festivals dominated by male acts, study shows, as Glastonbury begins. https://www.bbc.com/news/uk-england-40273193 . Accessed 14 June 2021

Sim J, Cho D, Hwang Y, Telang R (2022) Frontiers: virus shook the streaming star: estimating the covid-19 impact on music consumption. Mark Sci 41(1):19–32

Smith SL, Pieper K, Clark H, Case A, Choueiti M (2020) Inclusion in the recording studio? Ratio 25:16–8

Soha M, McDowell ZJ (2016) Monetizing a meme: Youtube, content id, and the harlem shake. Social Media + Society 2(1):2056305115623801

Ternovski J, Yasseri T (2020) Social complex contagion in music listenership: a natural experiment with 1.3 million participants. Soc Netw 61(October 2019):144–152

Thoemmes F, Liao W, Jin Z (2017) The analysis of the regression-discontinuity design in R. J Educ Behav Stat 42(3):341–360. https://doi.org/10.3102/1076998616680587

Villamizar-Villegas M, Pinzon-Puerto FA, Ruiz-Sanchez MA (2021) A comprehensive history of regression discontinuity designs: an empirical survey of the last 60 years. J Econ Surv, 1–49

Wallis C (2011) Performing gender: a content analysis of gender display in music videos. Sex Roles 64(3–4):160–172

Weitzer R, Kubrin CE (2009) Misogyny in rap music: a content analysis of prevalence and meanings. Men Masculinities 12(1):3–29

Winship C, Morgan SL (1999) The estimation of causal effects from observational data. Ann Rev Sociol 25:659–706

Wlömert N, Papies D (2016) On-demand streaming services and music industry revenues-insights from Spotify’s market entry. Int J Res Mark 33(2):314–327

Download references

Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.

Author information

Authors and affiliations.

Department of Applied Economics, Universitat de València, València, Spain

Juan D. Montoro-Pons & María Caballer-Tarazona

Department of Marketing, Universitat de València, València, Spain

Manuel Cuadrado-García

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Juan D. Montoro-Pons .

Ethics declarations

Competing interests.

The authors have no competing interests to disclose.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Montoro-Pons, J.D., Caballer-Tarazona, M. & Cuadrado-García, M. Assessing complementarities between live performances and YouTube video streaming. Empir Econ 65 , 2953–2978 (2023). https://doi.org/10.1007/s00181-023-02444-4

Download citation

Received : 19 July 2022

Accepted : 22 May 2023

Published : 01 June 2023

Issue Date : December 2023

DOI : https://doi.org/10.1007/s00181-023-02444-4

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Spillover effects
• Recorded music
• Digitization
• Regression discontinuity design

JEL Classification

• Find a journal
• Publish with us
• Track your research

Dialogue: The Interdisciplinary Journal of Popular Culture and Pedagogy

A framework for using popular music videos to teach media literacy.

Jordan M. McClain Drexel University Philadelphia, Pennsylvania, USA [email protected]

This article discusses the use of popular music videos as a tool for teaching media literacy. First, the article addresses the importance of music videos as popular culture, what other music video research has examined, and what features make music videos a good fit for in-class work investigating media and popular culture. Then the article details a single-class activity for introducing and teaching media literacy through the use of music videos. To achieve this objective, the article also proposes a set of original music video-specific discussion questions. Finally, a particular music video is considered to illustrate possible results of this activity and the broader issues that may arise from class discussion.

Communication, Media, Media Studies, Popular Culture, Pedagogy, New Media, Digital Media, Media Literacy, Media Education, Music Videos

Although popular music videos have long been criticized for their superficiality, fast edits, and sensational content, features like these help make the videos an excellent teaching tool, effective for getting students’ attention and exploring broad issues. Many educators may be skeptical about or may have never thought about the benefits of using music videos in the classroom—thus the shortage of research on this approach. Cayari wrote about students creating music videos in order to learn music and technology skills.  Maskell discussed the use of music videos for teaching English, saying the content has “huge potential for use across the entire English curriculum” (54). There is still, however, much to uncover about the myriad possible uses of music videos as a pedagogical instrument.

With a focus on popular music videos, this essay discusses their importance, describes an activity using them to teach media literacy skills, offers some new music video-specific ideas for introductory media literacy exercises, and shares example results of the activity. This information may appeal to a wide range of educators, especially media and popular culture scholars teaching undergraduate college courses such as Media and Society, Media Literacy, or Introduction to Popular Culture.

Although the pedagogical value of music videos remains formally under-recognized, many have thoroughly established why music videos are an important and potent way to learn about life around the globe. “Music television deserves serious attention from students of popular culture” (Goodwin and Grossberg ix), proclaimed the introduction of Sound and Vision: The Music Video Reader, the influential collection edited by Frith, Goodwin, and Grossberg. Supporting this call to study music videos, Austerlitz saw them as a “fascinating oddity” (1) and a “compelling marker of cultural history” (1). He concluded that the music video’s “triumphs render it a subject worthy of deeper study and attention” (1). In summarizing the state of music video research and demonstrating why they are more than just entertainment, Straw wrote, “music videos are increasingly seen as elements within complex assemblages of image and sound that circulate the world and are recombined within a variety of diasporic media, from satellite television networks through DVD and Internet video clip sites” (3176).

Consideration of certain music video research trends indicates their diverse potential. One major trend adopts a media effects perspective and examines how music videos influence the ways audiences think and behave, especially younger groups like adolescents, teens, or college students. Studies have looked at music video effects in terms of sex, such as how kids imitate the content (Ey and Cupit), how they sext (Van Ouytsel, Ponnet, and Walrave), and what their attitudes are toward sex (Aubrey, Hopper, and Mbure; Beentjes and Konig; Kistler and Lee; Zhang, Miller, and Harrison). Others have researched music videos’ effects on perceptions of rape (Burgess and Burpo; Sprankle, End, and Bretz). There is also much work on the influence of music videos on how people think about gender-specific ideas related to misogyny (van Oosten, Peter, and Valkenburg) or bodily self-perception (Mischner et al.).

Overlapping with work that emphasizes effects, there is a trend of research interested in representational patterns in music videos. Gender often emerges as a main focal point, such as Wallis’s content analysis of differences in gender displays. Many have also tied race to genre, with rap being a dominant line of inquiry (Balaji; Conrad, Dixon, and  Zhang; Zhang, Dixon, and Conrad). Overall, work on representation has spanned topics like sexual objectification (Aubrey and Frisby; Frisby and Aubrey), sexuality (Turner), and violence (Aikat; Smith and Boyson; Thaller and Messing).

Such trends show the utility of music videos in media research, popular culture studies, and beyond. In addition, music videos are characterized by a combination of features that make them an ideal fit for in-class activities about media and popular culture:

• They are conventionally short, compared to a full movie or television episode.
• They are often familiar, which benefits group discussion because many students bring background knowledge.
• They are common online, which makes it simple for instructors to find multiple good examples.
• They are easy to access, such as the free official content available on video-sharing sites like YouTube or hosting services like Vevo.
• They are often controversial, working as a compelling catalyst for critical discussion and thus able to help students identify important issues, then articulate their views on social or political matters.
• They are commonly imitated on the Web, as evidenced by remakes, parodies, satires, and mash-ups that have become a common way for lovers and haters—including amateurs, professionals, and people in between—to express themselves online. 1
• They are popular culture, as a collective form and as individual artifacts, which gives them instant student appeal and significance as a teaching tool.  

Activity: Popular Music Videos and Media Literacy

The following activity is a productive way to use music videos to introduce and teach media literacy. This exercise is intended to occur in class and requires the instructor’s use of an Internet-connected device that can play music videos viewable by the whole class at once (e.g., via projector or on a large monitor). Objectives include these:

• The exercise will (A) strategically use music videos as a teaching tool, (B) demonstrate the importance of critical thinking about music videos, and (C) demonstrate the importance of critical thinking about popular culture.
• Students will (A) strengthen media literacy skills and (B) increase comprehension of popular music videos as a significant form of entertainment media.  

Preparation: Prior to class, carefully select a popular music video accessible online and useful as a teaching tool. Billboard charts and YouTube’s “Popular on YouTube” section are helpful starting points. The instructor should select something that will resonate with students; this can be based on recency or the interests and personalities of the class. I suggest watching the video many times before class. It is also essential to research the video’s production background and popular reception. Immediately before class begins, it is smart to prepare the music video for easy start-up and test all necessary technology—video connection, audio levels, video start function, video end point.

Execution: Once class begins, start the activity by announcing its order (i.e., discuss media literacy, watch music video, analyze video alone and then together) and expected outcomes (i.e., enhance media literacy comprehension and skills).

Part 1: Introduce Media Literacy and Music Video-Specific Follow-Up Questions

First, I explain media literacy and the following five key questions of media literacy, using visual aids like PowerPoint slides and the Center for Media Literacy’s website, medialit.org:

• Authorship: “Who created this message?”
• Format: “What creative techniques are used to attract my attention?”
• Audience: “How might different people understand this message differently than me?”
• Content: “What values, lifestyles and points of view are represented in, or omitted from, this message?”
• Purpose: “Why is this message being sent?”

As justified in the rationale above, we then briefly discuss why music videos are media content worthy of critical thought.

Next, to successfully analyze popular music videos and expand on the preexisting five key questions of media literacy, I propose the following set of original follow-up questions that are music video-specific—four follow-ups for each of the main questions—to help prompt critical thought and advance media literacy about popular music videos:

• Who is explicitly identified as a creator?
• Who created the song?
• Who created the music video?
• What are some major components of the music video that people created?
• What techniques are used in the music?
• What techniques are used in the music video?
• How does this music video seem influenced by popular culture?
• How has this music video seemingly influenced popular culture?
• Who do you think are some target audiences for this music video?
• What components of the music video indicate its target audience?
• What parts of the music video seem open to interpretation?
• What parts of the music video seem controversial? To whom?
• How does the music video convey this?
• How do you think this relates to the music video’s creators?
• How do you think this relates to the music video’s target audience?
• What may have caused these representations and omissions?
• Why was this music created?
• Why was the music video created?
• Why was the music video created for this format? (I.e., cable television, the Web, DVD, etc.)
• Who would benefit from the music video’s popularity?  

Part 2: Watch a Music Video

After focusing on media literacy questions, introduce the music video by identifying the song and performer. I find it useful to informally survey how many students know the song or artist and how many like the song or artist. It is crucial to establish the significance of studying this artifact. For instance, instructors should cite facts about awards the artist or song has won, sales information like albums or singles sold, rankings from Billboard/Nielsen chart data, concert grosses, YouTube views, and social media metrics (e.g., how many likes or followers an artist has online). It is best also to show students visuals like a Twitter feed or Billboard.com article to support those claims. This will help students recognize the significance of putting popular culture under the microscope—this is not just a song but a social phenomenon that deserves to be studied, and the class is learning a system for accomplishing that.

Here it is helpful to notify students that after watching the video once, they will need to answer and discuss the five media literacy questions and music video-specific follow-ups. Thus, as they watch, students should think about answers to the questions, which they may wish to quickly review before watching the video at this point.

Part 3: Practice Media Literacy Skills by Discussing the Music Video

Solo: After watching the video, students should individually write answers to each media literacy question and the follow-ups. When dealing with time constraints for this in-class activity, I advise students to focus on answers that come easiest, instead of straining to complete all questions (i.e., quality over quantity). This is a good time to encourage optional Internet use for those with enabled devices. Answers are possible with only a pencil and paper, but Web-based research will probably strengthen responses.

Small groups: After the solo work, students form pairs or triads and share their findings with each other. They should consider what they learned from peers to expand their answer list and prepare for a full-class discussion.

As a class: After the small group work, reconvene as a class and watch the video for a second and final time. This provides a chance to see more, helps solidify what students learned so far, and refreshes memories for the following discussion.

I then lead a Q&A through each of the five key media literacy questions and follow-ups. Instructors should seek many answers to each question, solicit like and unlike observations across the group, and play devil’s advocate to help students form their opinions.

Activity Results

This activity results in valuable dialogues, which will vary based on the video(s) examined. One highly recommended music video to choose for this activity is Katy Perry’s 2013 hit, “Roar” (Lipshutz; Perry, “Katy Perry – Roar”) 2 . Using this video would give the instructor a chance to talk about Perry’s many Grammy nominations, MTV Awards, Nickelodeon Kids’ Choice Awards, and Guinness World Records. The instructor could also discuss her remarkable billion-plus views that place this song in the top ten most-viewed YouTube and Vevo videos (Jang; Lane; “Vevo Top Videos”) and made Perry “the first artist to ever have two videos with over 1 Billion [ sic ] views” (“Katy Perry – Vevo”; “Roar10xCertified”). Students respond well to these kinds of arguments for a video’s significance and facts like Perry’s status as the most-followed Twitter user—with over 75 million followers, she ranks above people like Justin Bieber and President Obama (Perry, “Tweets”; “Twitter Top 100”).

Discussing Perry’s “Roar” video would likely cause students to answer the media literacy questions and follow-ups in ways that lead to fascinating conversations about the major media literacy concepts. “Authorship” would relate to the song being co-written by a team of professional hit makers including Max Martin, Dr. Luke, and Bonnie McKee (Hampp; Seabrook). “Format” would connect to sexualization, familiar pop song ingredients, and the use of visual effects. “Audience” would lead to concerns about young fans, PETA’s objections to the video’s use of animals (Boardman; Palmer), or the video’s twist ending. “Content” would tie to portrayals of selfies, makeup use, and heterosexuality or sexual orientation. “Purpose” would relate to product sales, promotional culture, the modern music industry, free YouTube content, conspicuous use of Nokia merchandise, and celebrity branding.

This kind of popular music video analysis, based on the five key media literacy questions and follow-ups, enables discussion of many broad issues. In particular, this includes:

• How race, class, age, and ability are represented in music videos.
• How gender, sex, sexuality, and sexism are treated in music videos.
• How beauty norms are reflected in music videos; how this impacts body image, self-esteem, or eating disorders outside music videos.
• How celebrities appear in music videos; how musicians are positioned as celebrities in music videos.
• What music videos tell us about censorship, evolving moral standards, political correctness, and cultural taboos.
• How product placement shapes music videos.
• How genre affects music videos.
• How new and digital media impact music videos.

By using this activity, I have found that students thoroughly enjoy practicing and developing critical thinking skills through the study of everyday media and popular culture. The classroom becomes a space where fun and learning can logically and productively intersect. Students become more consistently engaged with class topics and discussions, searching for such intersection. Their media literacy skills improve—instantly and long-term—through the type of practice and collaborative critique that this exercise facilitates. As a result, students are more sensitive, informed, and skilled critical consumers of entertainment media.

This essay expands on general media literacy principles and produces original music video-specific questions, enabling systematic use of music videos as effective resources for teaching media literacy and critical thinking about media and popular culture. The five key media literacy questions are a valuable framework for studying popular music videos and exploring the broader issues they raise. Without the media literacy framework, this exercise might allow only surface-level scrutiny. Using the media literacy foundation strengthens, deepens, and formalizes this learning process, enhancing student comprehension, analysis, and evaluation of popular music videos as important media content.

The in-class activity described in this essay is ideal for undergraduate courses, but can be adapted by prefacing the work with level-appropriate lectures about media and popular culture for a variety of potential student audiences, such as tweens, pre-college teens, or graduate students. One alternative to the in-class activity is to remake it as a written test, which would benefit from a rubric used to grade answers. For example, instructors may choose to teach the five key media literacy questions first, then, on the same or a different day, show a music video and require students to answer the five questions and music video-specific follow-ups as a test of knowledge and skills. Other possibilities include a student presentation (individuals or groups pick a modern video, argue for its significance, analyze its content using the music video-specific follow-ups, and consider the implications); a reflection paper (students address the extent to which media literacy about music videos will impact how they think about such entertainment); or a self-produced video essay (students use the media literacy questions and music video-specific follow-ups as prompts for a prepared, recorded oral critique of a popular music video; bonus points to those who share their video essay on YouTube).

Popular music videos have many educational uses, which span disciplines. These videos are excellent instruments, effective for getting students’ attention, and helpful for teaching about many complex and meaningful concepts. Educators should therefore embrace and experiment with music videos as a powerful teaching tool.

1. By way of illustration, consider the many humorous takeoffs on The Black Eyed Peas song, “My Humps,” which inspired popular online videos by alt-rock celebrity Alanis Morissette, gender-role-defying electronic musician Peaches, and pre-teen remix video YouTube-star MattyBRaps.

2. Here are some other recommended popular music videos that work well for this activity: Michael Jackson, “Thriller”; Madonna, “Erotica”; Shania Twain, “Man! I Feel Like a Woman!”; One Direction, “What Makes You Beautiful”; Robin Thicke, “Blurred Lines”; Pharrell Williams, “Happy”; Taylor Swift, “Shake it Off”; Drake, “Hotline Bling.”

Works Cited

Aikat, Debashis. “Streaming Violent Genres Online: Visual Images in Music Videos on BET.com, Country.com, MTV.com, and VH1.com.” Popular Music and Society 27.2 (2004): 221-240. Web. 16 Sept. 2015.

Aubrey, Jennifer Stevens, and Cynthia M. Frisby. “Sexual Objectification in Music Videos: A Content Analysis Comparing Gender and Genre.” Mass Communication and Society 14.4 (2011): 475-501. Web. 16 Sept. 2015.

Aubrey, Jennifer Stevens, K. Megan Hopper, and Wanjiru G. Mbure. “Check That Body! The Effects of Sexually Objectifying Music Videos on College Men’s Sexual Beliefs.” Journal of Broadcasting & Electronic Media 55.3 (2011): 360-79. Web. 16 Sept. 2015.

Austerlitz, Saul. Money for Nothing: A History of the Music Video, from the Beatles to the White Stripes . New York: Continuum, 2007. Print.

Balaji, Murali. “Owning Black Masculinity: The Intersection of Cultural Commodification and Self-Construction in Rap Music Videos.” Communication, Culture & Critique 2.1 (2009): 21-38. Web. 16 Sept. 2015.

Beentjes, Johannes W. J., and Ruben P. Konig. “Does Exposure to Music Videos Predict Adolescents’ Sexual Attitudes?” European Scientific Journal 9.14 (2013): 1-20. Web. 16 Sept. 2015.

Boardman, Madeline. “PETA: Katy Perry’s ‘Roar” Music Video is Cruel to Animals.” HuffingtonPost.com . The Huffington Post, 15 Sept. 2013. Web. 24 Sept. 2015.

Burgess, Melinda C. R., and Sandra Burpo. “The Effect of Music Videos on College Students’ Perceptions of Rape.” College Student Journal 46.4 (2012): 748-763. Web. 16 Sept. 2015.

Cayari, Christopher. “Using Informal Education Through Music Video Creation.” General Music Today 27.3 (2014): 17-22. Web. 16 Sept. 2015.

Center for Media Literacy. “Five Key Questions Form Foundation for Media Inquiry: Keywords and Guiding Questions Help Build Habits of Critical Thinking.” MediaLit.org. Center for Media Literacy, n.d.: Web. 24 Sept. 2015.

Conrad, Kate, Travis L. Dixon, and Yuanyuan Zhang. “Controversial Rap Themes, Gender Portrayals and Skin Tone Distortion: A Content Analysis of Rap Music Videos.” Journal of Broadcasting & Electronic Media 53.1 (2009): 134-56. Web. 16 Sept. 2015.

Ey, Lesley-Anne, and C. Glenn Cupit. “Primary School Children’s Imitation of Sexualised Music Videos and Artists.” Children Australia 38.3 (2013): 115-123. Web. 16 Sept. 2015.

Frisby, Cynthia M., and Jennifer Stevens Aubrey. “Race and Genre in the Use of Sexual Objectification in Female Artists’ Music Videos.” Howard Journal of Communications 23.1 (2012): 66-87. Web. 16 Sept. 2015.

Goodwin, Andrew, and Lawrence Grossberg. Introduction. Sound and Vision: The Music Video Reader . Ed. Simon Frith, Andrew Goodwin, and Lawrence Grossberg. New York: Routledge, 1993. ix-xi. Print.

Hampp, Andrew. “Katy Perry, ‘Roar’: Track Review.” Billboard.com . Billboard, 12 Aug. 2013. Web. 24 Sept. 2015.

Jang, Meena. “YouTube’s 10th Anniversary: Watch the Top 10 Most Viewed Videos to Date.” Billboard.com . Billboard, 14 Feb. 2015. Web. 24 Sept. 2015.

“Katy Perry – Vevo Certified Artist.” Vevo.com . Vevo, 2015. Web. 24 Sept. 2015.

Kistler, Michelle E., and Moon J. Lee. “Does Exposure to Sexual Hip-Hop Music Videos Influence the Sexual Attitudes of College Students?” Mass Communication and Society 13.1 (2009): 67-86. Web. 16 Sept. 2015.

Lane, Laura. “These Are the Most-Watched YouTube Videos Ever – Have You Seen Them All?” People.com. Time Inc., 30 Apr. 2015. Web. 24 Sept. 2015.

Lipshutz, Jason. “Katy Perry’s ‘Roar’ Music Video: Watch the Singer’s Jungle Adventure.” Billboard.com . Billboard, 5 Sept. 2013. Web. 24 Sept. 2015.

Maskell, Hayden. “Using Music Videos.” English in Aotearoa 74 (2011): 54-57. Print.

Mischner, Isabelle H. S., Hein T. Van Schie, Daniël H. J. Wigboldus, Rick B. Van Baaren, and Rutger C. M. E. Engels. “Thinking Big: The Effect of Sexually Objectifying Music Videos on Bodily Self-Perception in Young Women.” Body Image 10.1 (2013): 26-34. Web. 16 Sept. 2015.

Palmer, Chris. “Katy Roars, Elephant Whimpers.” Peta.org. PETA, 10 Oct. 2013. Web. 24 Sept. 2015.

Perry, Katy (katyperry). “Tweets.” Twitter account. Twitter.com. Twitter, n.d. Web. 24 Sept. 2015.

Perry, Katy. “Katy Perry – Roar (Official).” Video file. KatyPerryVEVO. YouTube.com. YouTube, 5 Sept. 2013. Web. 24 Sept. 2015.

“Roar10xCertified.” KatyPerry.com . Capitol Records, 6 July 2015. Web. 24 Sept. 2015.

Seabrook, John. “The Doctor Is In: A Technique for Producing No. 1 Songs.” NewYorker.com. Conde Nast, 14 Oct. 2013. Web. 24 Sept. 2015.

Smith, Stacy L., and Aaron R. Boyson. “Violence in Music Videos: Examining the Prevalence and Context of Physical Aggression.” Journal of Communication 52.1 (2002): 61-83. Web. 16 Sept. 2015.

Sprankle, Eric L., Christian M. End, and Miranda N. Bretz. “Sexually Degrading Music Videos and Lyrics: Their Effects on Males’ Aggression and Endorsement of Rape Myths and Sexual Stereotypes.” Journal of Media Psychology 24.1 (2012): 31-39. Web. 16 Sept. 2015.

Straw, Will. “Music videos.” The International Encyclopedia of Communication. Ed. W. Donsbach. 2008. Print.

Thaller, Jonel, and Jill Theresa Messing. “(Mis)Perceptions Around Intimate Partner Violence in the Music Video and Lyrics for ‘Love the Way You Lie’.” Feminist Media Studies 14.4 (2014): 623-39. Web. 16 Sept. 2015.

Turner, Jacob S. “Sex and the Spectacle of Music Videos: An Examination of the Portrayal of Race and Sexuality in Music Videos.” Sex Roles 64.3-4 (2011): 173-91. Web. 16 Sept. 2015.

“Twitter top 100 most followers.” Twittercounter.com . Twitter, 2015. Web. 24 Sept. 2015.

Van Oosten, Johanna M. F., Jochen Peter, and Patti M. Valkenburg. “The Influence of Sexual Music Videos on Adolescents’ Misogynistic Beliefs: The Role of Video Content, Gender, and Affective Engagement.” Communication Research 42.7 (2015): 986-1008. Web. 16 Sept. 2015.

Van Ouytsel, Joris, Koen Ponnet, and Michel Walrave. “The Associations Between Adolescents’ Consumption of Pornography and Music Videos and Their Sexting Behavior.” Cyberpsychology, Behavior, and Social Networking 17.12 (2014): 772-78. Web. 16 Sept. 2015.

“Vevo Top Videos Most Viewed All Time.” Vevo.com. 2015. Web. 24 Sept. 2015.

Wallis, Cara. “Performing Gender: A Content Analysis of Gender Display in Music Videos.” Sex Roles 64.3-4 (2011): 160-72. Web. 16 Sept. 2015.

Zhang, Yuanyuan, Laura E. Miller, and Kristen Harrison. “The Relationship Between Exposure to Sexual Music Videos and Young Adults’ Sexual Attitudes.” Journal of Broadcasting & Electronic Media 52.3 (2008): 368-86. Web. 16 Sept. 2015.

Zhang, Yuanyuan, Travis L. Dixon, and Kate Conrad. “Rap Music Videos and African American Women’s Body Image: The Moderating Role of Ethnic Identity.” Journal of Communication 59.2 (2009): 262-78. Web. 16 Sept. 2015.

Author Bio:

Dr. Jordan M. McClain is Assistant Teaching Professor of Communication at Drexel University in Philadelphia, PA. He enjoys researching and teaching about framing in music journalism, celebrity, the intersection of television and music culture, and consumer culture. For the Mid-Atlantic Popular & American Culture Association (MAPACA) he serves on the executive board,  as Music area co-chair,  and as Journalism and News Media area chair. For the Popular Culture Association/American Culture Association (PCA/ACA), he chairs the Professional Development area.

Social media:

Academia.edu: https://drexel.academia.edu/JordanMcClain LinkedIn:  https://www.linkedin.com/in/jordan-m-mcclain-72304163 Twitter: https://twitter.com/j_mcclain

Reference Citation:

McClain, Jordan M. “ A Framework for Using Popular Music Videos to Teach Media Literacy.” Dialogue: The Interdisciplinary Journal of Popular Culture and Pedagogy  3.1 (2016). Web and Print.  

McClain, J. M. (2016).   A framework for using popular music videos to teach media literacy.  Dialogue: The Interdisciplinary Journal of Popular Culture and Pedagogy. 3 (1).  http://journaldialogue.org/issues/a-framework-for-using-popular-music-videos-to-teach-media-literacy/ 

Tags: Communication , Digital Media , Media , Media Education , Media Literacy , media studies , Music Videos , New Media , pedagogy , Popular Culture

• Volume 11, Issue 1 — On Media Literacy, Power, and Representation
• Volume 10 Issue 3 — 10 Years of Dialogue: Highlights from 2014–2022
• Volume 10, Issue 2 — Pop Culture Page-Turners: Unveiling the Must-Read Books for Summer Enthusiasts and Pedagogues
• Volume 10, Issue 1 — Cultivating the Futures of Popular Culture and Pedagogy
• Volume 9, Issue 4 — Provoking Awareness and Practical Applications in Popular Culture and Pedagogy: Syllabi, Games, and Teaching in Higher Education
• Volume 9, Issue 3 — Critique and “Controversy” in Pedagogy and Pop Culture
• Volume 9, Issue 1 & 2 — Teaching and Learning with the Grateful Dead
• Volume 8, Issue 3 — Traversing Borders, Transgressing Boundaries in Popular Culture and Pedagogy
• Volume 8, Issue 2 — Infusing Pedogogy with Empathy, Social Action and Value through Popular Culture
• Volume 8, Issue 1 — Evolving Awareness in Popular Culture and Pedagogy
• Volume 7, Issue 3 — (Un)conscious Representation: Race, Gender, Ideology
• Volume 7, Issue 2 — Engaged Popular Culture and Pedagogy: Awareness, Understanding and Social Justice
• Special Series: Pedagogy During COVID-19
• Volume 7, Issue 1 — Bodies in Motion: Challenging Imagery, Tradition, and Teaching
• Volume 6, Issue 3 — Otherness , Survival and Hope: Pedagogies in Popular Media
• Volume 6, Issue 2 — Criminals as Heroes: Problems and Pedagogy in Popular Culture
• Volume 6, Issue 1 — Reinterpretation: Situating Culture from Pedagogy to Politics
• Volume 5, Issue 3 — Behind the Scenes: Uncovering Violence, Gender, and Powerful Pedagogy
• Volume 5, Issue 2 — Reading into Creativity: New Approaches in Concept and Practice
• Volume 5, Issue 1 — Visualizing Popular Culture: From Theater to the Graphic Novel
• Volume 4, Issue 1 — Intersections: Belief, Pedagogy, and Politics
• Volume 3, Issue 2 — Adapting Our Approaches: (In)Formal Learning, Stereotypes, and Traumas
• Volume 3, Issue 1 — Popular Culture Pedagogy: Theory and Application in Academia
• Volume 2, Issue 1 — Traversing Realities: Genres, Histories, and Politics in Popular Culture
• Volume 1, Issue 1 — Classics in Contemporary Culture

Dialogue is indexed by Google Scholar

Call for submissions, connect with dialogue.

Examining Transmedia Storytelling in BTS’s Music Videos and Short Films

28 Pages Posted: 15 Sep 2020 Last revised: 17 Sep 2020

Aleena Faisal

University of Ontario Institute of Technology

Date Written: April 3, 2020

Trans-media storytelling is a concept where the whole narrative can only be understood through media extensions that relate to each other and expand the narrative. Despite the handful of studies done using numerous texts and fictional worlds, none have looked at music videos with this concept. BTS is a Korean pop group whose rise in popularity is given partial credit due to the trans-media narrative of their videography. This research paper did a content analysis to analyze BTS’s music videos and short films to determine how the group uses trans-media storytelling in their narrative and if music video genre plays an important role in building a cohesive, ever-expanding universe. Narrative elements and visual symbols are actively used to connect BTS’s music videos together directly to the narrative and as distant peripherals to the main story, and the addition of each video creates more opportunity for fans to engage with the text and add to the narrative. Although BTS is a unique group to follow through with a consistent narrative expanding over the course of their music video history, they have created more opportunity for artists and music video producers to follow suit and use trans-media storytelling for their own content.

Keywords: Trans-Media Storytelling, Narrative, Music Video, Genre, Content Analysis, BTS, K-Pop, Symbolism, Visual Rhetoric, Music, Entertainment, Fandom

Suggested Citation: Suggested Citation

Aleena Faisal (Contact Author)

University of ontario institute of technology ( email ), do you have a job opening that you would like to promote on ssrn, paper statistics, related ejournals, writing technologies ejournal.

Subscribe to this free journal for more curated articles on this topic

Communication & Technology eJournal

Subscribe to this fee journal for more curated articles on this topic

Mass Communication & Popular Culture eJournal

Visual communication ejournal, film, media & tv studies ejournal, visual anthropology, media studies & performance ejournal.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Review Article
• Open access
• Published: 22 June 2021

Mental health and music engagement: review, framework, and guidelines for future studies

• Daniel E. Gustavson   ORCID: orcid.org/0000-0002-1470-4928 1 , 2 ,
• Peyton L. Coleman   ORCID: orcid.org/0000-0001-5388-6886 3 ,
• John R. Iversen 4 ,
• Hermine H. Maes 5 , 6 , 7 ,
• Reyna L. Gordon 2 , 3 , 8 , 9 &
• Miriam D. Lense 2 , 8 , 9  

Translational Psychiatry volume  11 , Article number:  370 ( 2021 ) Cite this article

43k Accesses

24 Citations

79 Altmetric

Metrics details

• Medical genetics
• Psychiatric disorders

Is engaging with music good for your mental health? This question has long been the topic of empirical clinical and nonclinical investigations, with studies indicating positive associations between music engagement and quality of life, reduced depression or anxiety symptoms, and less frequent substance use. However, many earlier investigations were limited by small populations and methodological limitations, and it has also been suggested that aspects of music engagement may even be associated with worse mental health outcomes. The purpose of this scoping review is first to summarize the existing state of music engagement and mental health studies, identifying their strengths and weaknesses. We focus on broad domains of mental health diagnoses including internalizing psychopathology (e.g., depression and anxiety symptoms and diagnoses), externalizing psychopathology (e.g., substance use), and thought disorders (e.g., schizophrenia). Second, we propose a theoretical model to inform future work that describes the importance of simultaneously considering music-mental health associations at the levels of (1) correlated genetic and/or environmental influences vs. (bi)directional associations, (2) interactions with genetic risk factors, (3) treatment efficacy, and (4) mediation through brain structure and function. Finally, we describe how recent advances in large-scale data collection, including genetic, neuroimaging, and electronic health record studies, allow for a more rigorous examination of these associations that can also elucidate their neurobiological substrates.

Similar content being viewed by others

Biological principles for music and mental health

A comprehensive investigation into the genetic relationship between music engagement and mental health

The effects of playing music on mental health outcomes

Introduction.

Music engagement, including passive listening and active music-making (singing, instrument playing), impacts socio-emotional development across the lifespan (e.g., socialization, personal/cultural identity, mood regulation, etc.), and is tightly linked with many cognitive and personality traits [ 1 , 2 , 3 ]. A growing literature also demonstrates beneficial associations between music engagement and quality of life, well-being, prosocial behavior, social connectedness, and emotional competence [ 4 , 5 , 6 , 7 , 8 ]. Despite these advances linking engagement with music to many wellness characteristics, we have a limited understanding of how music engagement directly and indirectly contributes to mental health, including at the trait-level (e.g., depression and anxiety symptoms, substance use behaviors), clinical diagnoses (e.g., associations with major depressive disorder (MDD) or substance use disorder (SUD) diagnoses), or as a treatment. Our goals in this scoping review are to (1) describe the state of music engagement research regarding its associations with mental health outcomes, (2) introduce a theoretical framework for future studies that highlight the contribution of genetic and environmental influences (and their interplay) that may give rise to these associations, and (3) illustrate some approaches that will help us more clearly elucidate the genetic/environmental and neural underpinnings of these associations.

Scope of the article

People interact with music in a wide variety of ways, with the concept of “musicality” broadly including music engagement, music perception and production abilities, and music training [ 9 ]. Table 1 illustrates the breadth of music phenotypes and example assessment measures. Research into music and mental health typically focuses on measures of music engagement, including passive (e.g., listening to music for pleasure or as a part of an intervention) and active music engagement (e.g., playing an instrument or singing; group music-making), both of which can be assessed using a variety of objective and subjective measures. We focus primarily on music engagement in the current paper but acknowledge it will also be important to examine how mental health traits relate to other aspects of musicality as well (e.g., perception and production abilities).

Our scoping review and theoretical framework incorporate existing theoretical and mechanistic explanations for how music engagement relates to mental health. From a psychological perspective, studies have proposed that music engagement can be used as a tool for encouraging self-expression, developing emotion regulation and coping skills, and building community [ 10 , 11 ]. From a physiological perspective, music engagement modulates arousal levels including impacts on heart rate, electrodermal activity, and cortisol [ 12 , 13 ]. These effects may be driven in part by physical aspects of music (e.g., tempo) or rhythmic movements involved in making or listening to music, which impact central nervous system functioning (e.g., leading to changes in autonomic activity) [ 14 ], as well as by personality and contextual factors (e.g., shared social experiences) [ 15 ]. Musical experiences also impact neurochemical processes involved in reward processing [ 10 , 13 , 14 , 16 , 17 , 18 ], which are also implicated in mental health disorders (e.g., substance use; depression). Thus, an overarching framework for studying music-mental health associations should integrate the psychological, physiological, and neurochemical aspects of these potential associations. We propose expanding this scope further through consideration of genetic and environmental risk factors, which may give rise to (and/or interact with) other factors to impact health and well-being.

Regarding mental health, it is important to recognize the hierarchical structure of psychopathology [ 19 , 20 ]. Common psychological disorders share many features and cluster into internalizing (e.g., MDD, generalized anxiety disorder (GAD), posttraumatic stress disorder (PTSD)), externalizing (e.g., SUDs, conduct disorder), and thought disorders (e.g., bipolar disorder, schizophrenia), with common variance shared even across these domains [ 20 ]. These higher-order constructs tend to explain much of the comorbidity among individual disorders, and have helped researchers characterize associations between psychopathology, cognition, and personality [ 21 , 22 , 23 ]. We use this hierarchical structure to organize our review. We first summarize the emerging literature on associations between music engagement and generalized well-being that provides promising evidence for associations between music engagement and mental health. Next, we summarize associations between music engagement and internalizing traits, externalizing traits/behaviors, and thought disorders, respectively. Within these sections, we critically consider the strengths and shortcomings of existing studies and how the latter may limit the conclusions drawn from this work.

Our review considers both correlational and experimental studies (typically, intervention studies; see Fig. 1 for examples of study designs). We include not only studies that examine symptoms or diagnoses based on diagnostic interviews, but also those that assess quantitative variation (e.g., trait anxiety) in clinical and nonclinical populations. This is partly because individuals with clinical diagnoses may represent the extreme end of a spectrum of similar, sub-clinical, problems in the population, a view supported by evidence that genetic influences on diagnosed psychiatric disorders or DSM symptom counts are similar to those for trait-level symptoms in the general population [ 24 , 25 ]. Music engagement may be related to this full continuum of mental health, including correlations with trait-level symptoms in nonclinical populations and alleviation of symptoms from clinical disorders. For example, work linking music engagement to subjective well-being speaks to potential avenues for mental health interventions in the population at large.

Within experimental studies, music interventions can include passive musical activities (e.g., song listening, music and meditation, lyric discussion, creating playlists) or active musical activities (e.g., creative methods, such as songwriting or improvisation and/or re-creative methods, such as song parody).

The goal of this scoping review was to integrate across related, but often disconnected, literatures in order to propose a comprehensive theoretical framework for advancing our understanding of music-mental health associations. For this reason, we did not conduct a fully systematic search or quality appraisal of documents. Rather, we first searched PubMed and Google Scholar for review articles and meta-analyses using broad search terms (e.g., “review” and “music” and [“anxiety” or “depression” or “substance use”]). Then, when drafting each section, we searched for additional papers that have been published more recently and/or were examples of higher-quality research in each domain. When giving examples, we emphasize the most recent and most well-powered empirical studies. We also conducted some targeted literature searches where reviews were not available (e.g., “music” and [“impulsivity” or “ADHD”]) using the same databases. Our subsequent framework is intended to contextualize diagnostic, symptom, and mechanistic findings more broadly within the scope of the genetic and environmental risk factors on psychopathology that give rise to these associations and (potentially) impact the efficacy of treatment efforts. As such, the framework incorporates evidence from review articles and meta-analyses from various literatures (e.g., music interventions for anxiety [ 26 ], depression [ 27 ]) in combination with experimental evidence of biological underpinnings of music engagement and the perspective provided by newly available methods for population-health approaches (i.e., complex trait genetics, gene–environment interactions).

Music engagement and well-being

A growing body of studies report associations between music engagement and general indices of mental health, including increased well-being or emotional competence, lending support for the possibility that music engagement may also be associated with better specific mental health outcomes. In over 8000 Swedish twins, hours of music practice and self-reported music achievement were associated with better emotional competence [ 5 ]. Similarly, a meta-ethnography of 46 qualitative studies revealed that participation in music activities supported well-being through management of emotions, facilitation of self-development, providing respite from problems, and facilitating social connections [ 28 ]. In a sample of 1000 Australian adults, individuals who engaged with music, such as singing or dancing with others or attending concerts reported greater well-being vs. those who engaged in these experiences alone or did not engage. Other types of music engagement, such as playing an instrument or composing music were not associated with well-being in this sample [ 4 ]. Earlier in life, social music experiences (including song familiarity and synchronous movement to music) are associated with a variety of prosocial behaviors in infants and children [ 6 ], as well as positive affect [ 7 ]. Thus, this work provides some initial evidence that music engagement is associated with better general mental health outcomes in children and adults with some heterogeneity in findings depending on the specific type of music engagement.

Music engagement and internalizing problems

MDD, GAD, and PTSD are the most frequently clustered aspects of internalizing psychopathology [ 19 , 24 , 29 , 30 ]. Experimental studies provide evidence for the feasibility of music intervention efforts and their therapeutic benefits but are not yet rigorous enough to draw strong conclusions. The most severe limitations are small samples, the lack of appropriate control groups, few interventions with multiple sessions, and publications omitting necessary information regarding the intervention (e.g., intervention fidelity, inclusion/exclusion criteria, education status of intervention leader) [ 31 , 32 , 33 ]. Correlational studies, by contrast, suggest musicians are at greater risk for internalizing problems, but that they use music engagement as a tool to help manage these problems [ 34 , 35 ].

Experimental studies

Randomized controlled trials have revealed that music interventions (including both music therapies administered by board-certified music therapists and other music interventions) are associated with reduced depression, anxiety, and PTSD symptoms [ 26 , 27 , 33 , 36 ]. A review of 28 studies reported that 26 revealed significantly reduced depression levels in music intervention groups compared to control groups, including the 9 studies which included active non-music intervention control groups (e.g., reading sessions, “conductive-behavior” psychotherapy, antidepressant drugs) [ 27 ]. A similar meta-analysis of 19 studies demonstrated that music listening is effective at decreasing self-reported anxiety in healthy individuals [ 26 ]. A review of music-based treatment studies related to PTSD revealed similar conclusions [ 36 ], though there were only four relevant studies. More recent studies confirm these findings [ 37 , 38 , 39 ], such as one randomized controlled trial that demonstrated reduced depression symptoms in older adults following musical improvisation exercises compared to an active control group (gentle gymnastic activities) [ 39 ].

This work is promising given that some studies have observed effects even when compared to traditional behavior therapies [ 40 , 41 ]. However, there are relatively few studies directly comparing music interventions to traditional therapies. Some music interventions incorporate components of other therapeutic methods in their programs including dialectic or cognitive behavior therapies [ 42 ], but few directly compare how the inclusion of music augments traditional behavioral therapy. Still other non-music therapies incorporate music into their practice (e.g., background music in mindfulness therapies) [ 43 , 44 ], but the specific contribution of music in these approaches is unclear. Thus, there is a great need for further systematic research relating music to traditional therapies to understand which components of music interventions act on the same mechanisms as traditional therapies (e.g., developing coping mechanisms and building community) and which bolster or synchronize with other approaches (e.g., by adding structure, reinforcement, predictability, and social context to traditional approaches).

Aside from comparison with other therapeutic approaches, an earlier review of 98 papers from psychiatric in-patient studies concluded that promising effects of music therapy were limited by small sample sizes and methodological shortcomings including lack of reporting of adverse events, exclusion criteria, possible confounders, and characteristics of patients lost to follow-up [ 33 ]. Other problems included inadequate reporting of information on the source population (e.g., selection of patients and proportion agreeing to take part in the study), the lack of masking of interviewers during post-test, and concealment of randomization. Nevertheless, there was some evidence that therapies with active music participation, structured sessions, and multiple sessions (i.e., four or more) improved mood, with all studies incorporating these characteristics reporting significant positive effects. However, most studies have focused on passive interventions, such as music listening [ 26 , 27 ]. Active interventions (e.g., singing, improvising) have not been directly compared with passive interventions [ 27 ], so more work is needed to clarify whether therapeutic effects are indeed stronger with more engaging and active interventions.

Correlational studies

Correlational studies have focused on the use of music in emotional self-regulation. Specifically, individuals high in neuroticism appear to use music to help regulate their emotions [ 34 , 35 ], with beneficial effects of music engagement on emotion regulation and well-being driven by cognitive reappraisal [ 45 ]. Music listening may also moderate the association between neuroticism and depression in adolescents [ 46 ], consistent with a protective effect.

A series of recent studies have used validated self-reported instruments that directly assess how individuals use music activities as an emotion regulation strategy [ 47 , 48 , 49 , 50 ]. In adults, the use of music listening for anger regulation and anxiety regulation was positively associated with subjective well-being, psychological well-being, and social well-being [ 50 ]. In studies of adolescents and undergraduates, the use of music listening for entertainment was associated with fewer depression and anxiety symptoms [ 51 ]. “Healthy” music engagement in adolescents (i.e., using music for relaxation and connection with others) was also positively associated with happiness and school satisfaction [ 49 ]. However, the use of music listening for emotional discharge was also associated with greater depression, anxiety, and stress symptoms [ 51 ], and “unhealthy” music engagement (e.g., ‘hiding’ in music to block others out) was associated with lower well-being, happiness, school satisfaction, and greater depression and rumination [ 49 ]. Other work has highlighted the role of valence in these associations, with individuals who listen to happier music when they are in a bad mood reporting stronger ability for music to influence their mood than those who listen to sad music while in a negative mood [ 52 , 53 ].

This work highlights the importance of considering individuals’ motivations for engaging with music in examining associations with well-being and mental health, and are consistent with the idea that individuals already experiencing depression, anxiety, and stress use music as a therapeutic tool to manage their emotions, with some strategies being more effective than others. Of course, these correlational effects may not necessarily reflect causal associations, but could be due to bidirectional influences, as suggested by claims that musicians may be at higher risk for internalizing problems [ 54 , 55 , 56 ]. It is also necessary to consider demographic and socioeconomic factors in these associations [ 57 ], for example, because arts engagement may be more strongly associated with self-esteem in those with higher education [ 58 ].

It is also necessary to clarify if musicians (professional and/or nonprofessional) represent an already high-risk group for internalizing problems. In one large study conducted in Norway ( N  = 6372), professional musicians were higher in neuroticism than the general population [ 56 ]. Another study of musician cases ( N  = 9803) vs. controls ( N  = 49,015) identified in a US-based research database through text-mining of medical records found that musicians are at greater risk of MDD (Odds ratio [OR] = 1.21), anxiety disorders (OR = 1.25), and PTSD (OR = 1.13) [ 55 ]. However, other studies demonstrate null associations between musician status and depression symptoms [ 5 ] or mixed associations [ 59 ]. In N  = 10,776 Swedish twins, for example, professional and amateur musicians had more self-reported burnout symptoms [ 54 ]. However, neither playing music in the past, amateur musicianship, nor professional musicianship was significantly associated with depression or anxiety disorder diagnoses.

Even if musicians are at higher risk, such findings can still be consistent with music-making being beneficial and therapeutic (e.g., depression medication use is elevated in individuals with depressive symptoms because it is a treatment). Clinical samples may be useful in disentangling these associations (i.e., examining if those who engage with music more frequently have reduced symptoms), and wider deployment of measures that capture emotion regulation strategies and motivations for engaging with music will help shed light on whether high-risk individuals engage with music in qualitatively different ways than others [ 51 , 57 ]. Later, we describe how also considering the role of genetic and environmental risk factors in these associations (e.g., if individuals at high genetic and/or environmental risk self-select into music environments because they are therapeutic) can help to clarify these questions.

Music engagement and externalizing problems

The externalizing domain comprises SUDs, and also includes impulsivity, conduct disorder, and attention-deficit hyperactivity disorder (ADHD), especially in adolescents [ 20 , 24 , 60 , 61 ]. Similar to the conclusions for internalizing traits, experimental studies show promising evidence that music engagement interventions may reduce substance use, ADHD, and other externalizing symptoms, but conclusions are limited by methodological limitations. Correlational evidence is sparce, but there is less reason to suspect musicians are at higher risk for externalizing problems.

Intervention studies have demonstrated music engagement is helpful in patients with SUDs, including reducing withdrawal symptoms and stress, allowing individuals to experience emotions without craving substance use, and making substance abuse treatment sessions more enjoyable and motivating [ 62 , 63 , 64 ] (for a systematic review, see [ 65 ]). Similar to the experimental studies of internalizing traits, however, these studies would also benefit from larger samples, better controls, and higher-quality reporting standards.

Music intervention studies for ADHD are of similar quality. Such interventions have been shown to reduce inattention [ 66 ], decrease negative mood [ 67 ], and increase reading comprehension for those with ADHD [ 68 ]. However, there is a great amount of variability among children with ADHD, as some may find music distracting while others may focus better in the presence of music [ 69 ].

Little research has been conducted to evaluate music engagement interventions for impulsivity or conduct disorder problems, and findings are mixed. For example, a music therapy study of 251 children showed that beneficial effects on communication skills (after participating in a free improvisation intervention) was significant, though only for the subset of children above age 13 [ 70 ]. Another study suggested the promising effects of music therapy on social skills and problem behaviors in 89 students selected based on social/emotional problem behaviors, but did not have a control group [ 71 ]. Other smaller studies ( N  < 20 each) show inconsistent results on disruptive behaviors and aggression [ 72 , 73 ].

Correlational studies on externalizing traits are few and far between. A number of studies examined how listening habits for different genres of music relate to more or less substance use [ 74 , 75 , 76 , 77 ]. However, these studies do not strongly illuminate associations between music engagement and substance use because musical genres are driven by cultural and socioeconomic factors that vary over the lifespan. In the previously cited large study of American electronic medical records [ 55 ] where musicianship was associated with more internalizing diagnoses, associations were nonsignificant for “tobacco use disorder” (OR = 0.93), “alcoholism” (OR = 1.01), “alcohol-related disorders” (OR = 1.00), or “substance addiction and disorders” (OR = 1.00). In fact, in sex-stratified analyses, female musicians were at significantly decreased risk for tobacco use disorder (OR = 0.85) [ 55 ]. Thus, there is less evidence musicians are at greater risk for externalizing problems than in other areas.

Regarding other aspects of externalizing, some studies demonstrate children with ADHD have poor rhythm skills, opening a possibility that working on rhythm skills may impact ADHD [ 78 , 79 ]. For example, music might serve as a helpful scaffold (e.g., for attention) due to its regular, predictable rhythmic beat. It will be important to examine whether these associations with music rhythm are also observed for measures of music engagement, especially in larger population studies. Finally, musicians were reported to have lower impulsiveness than prior population samples, but were not compared directly to non-musicians [ 80 , 81 ].

Music engagement and thought disorders

Thought disorders typically encompass schizophrenia and bipolar disorder [ 20 ]. Trait-level measures include schizotypal symptoms and depression symptoms. Much like internalizing, music interventions appear to provide some benefits to individuals with clinical diagnoses, but musicians may be at higher risk for thought disorders. Limitations of both experimental and correlational studies are similar to those for internalizing and externalizing.

Music intervention studies have been conducted with individuals with schizophrenia and bipolar disorder. A recent meta-analysis of 18 music therapy studies for schizophrenia (and similar disorders) [ 82 ] demonstrated that music therapy plus standard care (compared to standard care alone) demonstrated improved general mental health, fewer negative symptoms of schizophrenia, and improved social functioning. No effects were observed for general functioning or positive symptoms of schizophrenia. Critiques echoed those described above. Most notably, although almost all studies had low risk of biases due to attrition, unclear risk of bias was evident in the vast majority of studies (>75%) for selection bias, performance bias, detection bias, and reporting bias. These concerns highlight the need for these studies to report more information about their study selection, blinding procedure, and outcomes.

More recent papers suggest similar benefits of music therapies in patients with psychosis [ 83 ] and thought disorders [ 84 ], with similar limitations (e.g., one study did not include a control group). Finally, although a 2021 review did not uncover more recent articles related to bipolar disorder, they argued that existing work suggests music therapy has the potential both to treat bipolar disorder symptoms and alleviate subthreshold symptoms in early stages of the disorder [ 85 ].

Much like internalizing, findings from the few existing studies suggest that musicians may be at higher risk for thought disorders. In the large sample of Swedish twins described earlier [ 54 ], playing an instrument was associated with more schizotypal symptoms across multiple comparisons (professional musicians vs. non-players; amateur musicians vs. non-players; still plays an instrument vs. never played). However, no associations were observed for schizophrenia or bipolar disorder diagnoses across any set of comparison groups. Another study demonstrated that individuals with higher genetic risk for schizophrenia or bipolar disorder were more likely to be a member of a creative society (i.e., actor or dancer, musician, visual artist, or writer) or work in a profession in these fields [ 86 ]. Furthermore, musician status was associated with “bipolar disorder” (OR = 1.18) and “schizophrenia and other psychotic disorders” (OR = 1.18) in US electronic health records (EHRs) [ 55 ].

Interim summary

There is promising evidence that music engagement is associated with better mental health outcomes. Music engagement is positively associated with quality of life, well-being, social connectedness, and emotional competence. However, some individuals who engage with music may be at higher risk for mental health problems, especially internalizing and thought disorders. More research is needed to disentangle these contrasting results, including clarifying how “healthy” music engagement (e.g., for relaxation or social connection) leads to greater well-being or successful emotion regulation, and testing whether some individuals are more likely to use music as a tool to regulate emotions (e.g., those with high neuroticism) [ 34 , 35 ]. Similarly, it will be important to clarify whether the fact that musicians may be an at-risk group is an extension of working in an artistic field in general (which may feature lower pay or lack of job security) and/or if similar associations are observed with continuous music engagement phenotypes (e.g., hours of practice). As we elaborate on later, genetically informative datasets can help clarify these complex associations, for example by tested whether musicians are at higher genetic risk for mental health problems but their music engagement mitigates these risks.

Music intervention studies are feasible and potentially effective at treating symptoms in individuals with clinical diagnoses, including depression, anxiety, and SUDs. However, it will be essential to expand these studies to include larger samples, random sampling, and active control groups that compare the benefits of music interventions to traditional therapies and address possible confounds. These limitations make it hard to quantify how specific factors influence the effectiveness of interventions, such as length/depth of music training, age of sample, confounding variables (e.g., socioeconomic status), and type of intervention (e.g., individual vs. group sessions, song playing vs. songwriting, receptive vs. active methods). Similarly, the tremendous breadth of music engagement activities and measures makes it difficult to identify the specific aspects of music engagement that convey the most benefits to health and well-being [ 87 ]. It is therefore necessary to improve reporting quality of studies so researchers can better identify these potential moderators or confounds using systematic approaches (e.g., meta-analyses).

Various mechanisms have been proposed to explain the therapeutic effects of music on mental health, including psychological (e.g., building communities, developing coping strategies) [ 10 , 11 ] and specific neurobiological drivers (e.g., oxytocin, cortisol, autonomic nervous system activity) [ 12 , 13 , 14 ]. However, it will be vital to conduct more systematic research comparing the effects of music interventions to existing therapeutic methods and other types of creative activities (e.g., art [ 88 ]) to quantify which effects and mechanisms are specific to music engagement. Music interventions also do not have to be an alternative to other treatments, but may instead support key elements of traditional interventions, such as being engaging, enjoyable, providing social context, and increasing structure and predictability [ 89 ]. Indeed, some music therapists incorporate principals from existing psychotherapeutic models [ 42 , 90 ] and, conversely, newer therapeutic models (e.g., mindfulness) incorporate music into their practice [ 43 , 44 ]. It is not yet possible to disentangle which aspects of music interventions best synergize with or strengthen standard psychotherapeutic practices (which are also heterogeneous), but this will be possible with better reporting standards and quality experimental design.

To encapsulate and extend these ideas, we next propose a theoretical framework that delineates key aspects of how music engagement may relate to mental health, which is intended to be useful for guiding future investigations in a more systematic way.

Theoretical framework for future studies

Associations between music engagement and mental health may take multiple forms, driven by several different types of genetic predispositions and environmental effects that give rise to, and interact with, proposed psychological and neurobiological mechanisms described earlier. Figure 2 displays our theoretical model in which potential beneficial associations with music are delineated into testable hypotheses. Four key paths characterize specific ways in which music engagement may relate to (and influence) mental health traits, and thus represent key research questions to be addressed in future studies.

Progression of mental health problems is based on a diathesis-stress model, where genetic predispositions and environmental exposures result in later problems (which can be remedied through treatment). Potential associations with music engagement include (Path 1; blue arrows) correlated genetic/environmental influences and/or causal associations between music engagement and trait-level mental health outcomes; (Path 2; red arrows) interactions between music engagement and risk factors to predict later trait-level or clinical level symptoms; and (Path 3; gold arrow) direct effects of music engagement on reducing symptoms or improving treatment efficacy. Path 4 (orange arrows) illustrates the importance of understanding how these potential protective associations are driven by neuroanatomy and function. MDD major depressive disorder, GAD generalized anxiety disorder, PTSD posttraumatic stress disorder, SUD substance use disorder(s).

Path 1: Music engagement relates to mental health through correlated genetic and environmental risk factors and/or causation

The diathesis-stress model of psychiatric disease posits that individuals carry different genetic liabilities for any given disorder [ 91 , 92 , 93 ], with disorder onset depending on the amount of negative vs. protective environmental life events and exposures the individual experiences. Although at first glance music engagement appears to be an environmental exposure, it is actually far from it. Twin studies have demonstrated that both music experiences and music ability measures are moderately heritable and genetically correlated with cognitive abilities like non-verbal intelligence [ 94 , 95 , 96 , 97 ]. Music engagement may be influenced by its own set of environmental influences, potentially including socioeconomic factors and availability of instruments. Thus, music engagement can be viewed as a combination of genetic and environmental predispositions and availability of opportunities for engagement [ 98 ] that are necessary to consider when evaluating associations with mental health [ 54 ].

When examining music-mental health associations, it is thus important to evaluate if associations are in part explained by correlated genetic or environmental influences (see Fig. 3 for schematic and explanation for interpreting genetic/environmental correlations). On one hand, individuals genetically predisposed to engage with music may be at lower risk of experiencing internalizing or externalizing problems. Indeed, music engagement and ability appear associated with cognitive abilities through genetic correlations [ 3 , 99 ], which may apply to music-mental health associations as well. On the other, individuals at high genetic risk for neuroticism or psychopathology may be more likely to engage with music because it is therapeutic, suggesting a genetic correlation in the opposite direction (i.e., increased genetic risk for musicians). To understand and better contextualize the potential therapeutic effects of music engagement, it is necessary to quantify these potential genetic associations, while simultaneously evaluating whether these associations are explained by correlated environmental influences.

Variance in any given trait is explained by a combination of genetic influences (i.e., heritability) and environmental influences. For complex traits (e.g., MDD or depression symptoms), cognitive abilities (e.g., intelligence), and personality traits (e.g., impulsivity), many hundreds or thousands of independent genetic effects are combined together in the total heritability estimate. Similarly, environmental influences typically represent a multitude of factors, from individual life events to specific exposures (e.g., chemicals, etc.). The presence of a genetic or environmental correlation between traits indicates that some set of these influences have an impact on multiple traits. A Displayed using a Venn diagram. Identifying the strength of genetic vs. environmental correlations can be useful in testing theoretical models and pave the way for more complex genetic investigations. Beyond this, gene identification efforts (e.g., genome-wide association studies) and additional analyses of the resulting data can be used to classify whether these associations represent specific genetic influences that affect both traits equally (i.e., genetic pleiotropy ( B )) or whether a genetic influence impacts only one trait which in turn causes changes in the other (i.e., mediated genetic pleiotropy ( C )). Environmental influences can also act pleiotropically or in a mediated-pleiotropy manner, but only genetic influences are displayed for simplicity.

Beyond correlated genetic and environmental influences, music engagement and mental health problems may be associated with one another through direct influences (including causal impacts). This is in line with earlier suggestions that music activities (e.g., after-school programs, music practice) engage adolescents, removing opportunities for drug-seeking behaviors [ 100 ], increasing their social connections to peers [ 101 ], and decreasing loneliness [ 41 ]. Reverse causation is also possible, for example, if experiencing mental health problems causes some individuals to seek out music engagement as a treatment. Longitudinal and genetically informative studies can help differentiate correlated risk factors (i.e., genetic/environmental correlations) from causal effects of music engagement (Fig. 2 , blue arrows) [ 102 ].

Path 2: Engagement with music reduces the impact of genetic risk

Second, genetic and environmental influences may interact with each other to influence a phenotype. For example, individual differences in music achievement are more pronounced in those who engage in practice or had musically enriched childhood environments [ 97 , 98 ]. Thus, music exposures may not influence mental health traits directly but could impact the strength of the association between genetic risk factors and the emergence of trait-level symptoms and/or clinical diagnoses. Such associations might manifest as decreased heritability of trait-level symptoms in musicians vs. non-musicians (upper red arrow in Fig. 2 ). Alternatively, if individuals high in neuroticism use music to help regulate their emotions [ 34 , 35 ], those who are not exposed to music environments might show stronger associations between neuroticism and later depressive symptoms or diagnoses than those engaged with music (lower red arrow in Fig. 2 ). Elucidating these possibilities will help disentangle the complex associations between music and mental health and could be used to identify which individuals would benefit most from a music intervention (especially preventative interventions). Later, we describe some specific study designs that can test hypotheses regarding this gene-environment interplay.

Path 3: Music engagement improves the efficacy of treatment (or acts as a treatment)

For individuals who experience severe problems (e.g., MDD, SUDs), engaging with music may reduce symptoms or improve treatment outcomes. This is the primary goal of most music intervention studies [ 27 , 33 ] (Fig. 2 , gold arrow). However, and this is one of the central messages of this model, it is important to consider interventions in the context of the paths discussed above. For example, if music engagement is genetically correlated with increased risk for internalizing or externalizing problems (Path 1) and/or if individuals at high genetic risk for mental health problems have already been using music engagement to develop strategies to deal with subthreshold symptoms (Path 2), then may be more likely to choose music interventions over other alternatives and find them more successful. Indeed, the beneficial aspects of music training on cognitive abilities appear to be drastically reduced in samples that were randomly sampled [ 103 ]. Therefore, along with other necessary reporting standards discussed above [ 32 , 33 ], it will be useful for studies to report participants’ prior music experience and consider these exposures in evaluating the efficacy of interventions.

Path 4: Music engagement influences brain structure and function

Exploring associations between music engagement and brain structure and function will be necessary to elucidate the mechanisms driving the three paths outlined above. Indeed, there are strong links between music listening and reward centers of the brain [ 104 , 105 ] including the nucleus accumbens [ 106 , 107 ] and ventral tegmental areas [ 108 ] that are implicated in the reward system for all drugs of abuse [ 109 , 110 , 111 , 112 ] and may relate to internalizing problems [ 113 , 114 , 115 ]. Moreover, activity in the caudate may simultaneously influence rhythmic sensorimotor synchronization, monetary reward processing, and prosocial behavior [ 116 ]. Furthermore, music listening may help individuals control the effect of emotional stimuli on autonomic and physiological responses (e.g., in the hypothalamus) and has been shown to induce the endorphinergic response blocked by naloxone, an opioid antagonist [ 18 , 117 ].

This work focusing on music listening and reward processing has not been extended to music making (i.e., active music engagement), though some differences in brain structure and plasticity between musicians and non-musicians have been observed for white matter (e.g., greater fractional anisotropy in corpus callosum and superior longitudinal fasciculus) [ 118 , 119 , 120 , 121 ]. In addition, longitudinal studies have revealed that instrument players show more rapid cortical thickness maturation in prefrontal and parietal areas implicated in emotion and impulse control compared to non-musician children/adolescents [ 122 ]. Importantly, because the existing evidence is primarily correlational, these cross-sectional and longitudinal structural differences between musicians and non-musicians could be explained by genetic correlations, effects of music training, or both, making them potentially relevant to multiple paths in our model (Fig. 2 ). Examining neural correlates of music engagement in more detail will shed light on these possibilities and advance our understanding of the correlates and consequences of music engagement, and the mechanisms that drive the associations discussed above.

New approaches to studying music and mental health

Using our theoretical model as a guide, we next highlight key avenues of research that will help disentangle these music-mental health associations using state-of-the-art approaches. They include the use of (1) genetic designs, (2) neuroimaging methods, and (3) large biobanks of EHRs.

Genetic designs

Genetic designs provide a window into the biological underpinnings of music engagement [ 123 ]. Understanding the contribution of genetic risk factors is crucial to test causal or mechanistic models regarding potential associations with mental health. At the most basic level, twin and family studies can estimate genetic correlations among music ability or engagement measures and mental health traits or diagnoses. Genetic associations can be examined while simultaneously quantifying environmental correlations, as well as evaluating (bidirectional) causal associations, by testing competing models or averaging across different candidate models [ 102 , 124 ], informing Path 1.

By leveraging samples with genomic, music engagement, and mental health data, investigators can also examine whether individuals at higher genetic risk for psychopathology (e.g., for MDD) show stronger associations between music engagement measures and their mental health outcomes (Path 2). As a theoretical example, individuals with low genetic risk for MDD are unlikely to have many depressive symptoms regardless of their music engagement, so the association between depressive symptoms and music engagement may be weak if focusing on these individuals. However, individuals at high genetic risk for MDD who engage with music may have fewer symptoms than their non-musician peers (i.e., a stronger negative correlation). This is in line with recent work revealing the heritability of depression is doubled in trauma exposed compared to non-trauma exposed individuals [ 125 ].

Gene–environment interaction studies using polygenic scores (i.e., summed indices of genetic risk based on genome-wide association studies; GWAS) are becoming more common [ 126 , 127 ]. There are already multiple large GWAS of internalizing and externalizing traits [ 128 , 129 , 130 ], and the first large-scale GWAS of a music measure indicates that music rhythm is also highly polygenic [ 131 ]. Importantly, is not necessary to have all traits measured in the same sample to examine cross-trait relationships. Studies with only music engagement and genetic data, for example, can still examine how polygenic scores for depression predict music engagement, or interact with music engagement measures to predict other study outcomes. Figure 4 displays an example of a GWAS and how it can be used to compute and apply a polygenic score to test cross-trait predictions.

A GWAS are conducted by examining whether individual genetic loci (i.e., single-nucleotide polymorphisms, or SNPs, depicted with G, A, C, and T labels within a sample (or meta-analysis) differentiate cases from controls. The example is based on a dichotomous mental health trait (e.g., major depressive disorder diagnosis), but GWAS can be applied to other dichotomous and continuous phenotypes, such as trait anxiety, musician status, or hours of music practice. Importantly, rather than examining associations on a gene-by-gene basis, GWAS identify relevant genetic loci using SNPs from across the entire genome (typically depicted using a Manhattan plot, such as that displayed at the bottom of A ). B After a GWAS has been conducted on a given trait, researchers can use the output to generate a polygenic score (sometimes called a polygenic risk score) in any new sample with genetic data by summing the GWAS effect sizes for each SNP allele present in a participant’s genome. An individual with a z  = 2.0 would have many risk SNPs for that trait, whereas an individual with z = −2 would have much fewer risk SNPs. C Once a polygenic score is generated for all participants, it can be applied like any other variable in the new sample. In this example, researchers could examine whether musicians are at higher (or lower) genetic risk for a specific disorder. Other more complex analyses are also possible, such as examining how polygenic scores interact with existing predictors (e.g., trauma exposure) or polygenic scores for other traits to influence a phenotype or predict an intervention outcome. Created with BioRender.com.

Finally, longitudinal twin and family studies continue to be a promising resource for understanding the etiology and developmental time-course of the correlates of mental health problems. Such designs can be used to examine whether associations between music and mental health are magnified based on other exposures or psychological constructs (gene-by-environment interactions) [ 132 ], and whether parents engaged with music are more likely to pass down environments that are protective or hazardous for later mental health (gene-environment correlations) in addition to passing on their genes. These studies also provide opportunities to examine whether these associations change across key developmental periods. The publicly available Adolescent Brain Cognitive Development study, for example, is tracking over 10,000 participants (including twin and sibling pairs) throughout adolescence, with measures of music engagement and exhaustive measures of mental health, cognition, and personality, as well as neuroimaging and genotyping [ 133 , 134 ]. Although most large samples with genomic data still lack measures of music engagement, key musical phenotypes could be added to existing study protocols (or to similar studies under development) with relatively low participant burden [ 135 ]. Musical questionnaires and/or tasks may be much more engaging and enjoyable than other tasks, improving volunteers’ research participation experience.

Neuroimaging

Another way to orient the design of experiments is through the exploration of neural mechanisms by which music might have an impact on mental health. This is an enormous, growing, and sometimes fraught literature, but there is naturally a great potential to link our understanding of neural underpinnings of music listening and engagement with the literature on neural bases of mental health. These advances can inform the mechanisms driving successful interventions and inform who may benefit the most from such interventions. We focus on two areas among many: (1) the activation of reward circuitry by music and (2) the impact music has on dynamic patterns of neural activity, both of which are likely vectors for the interaction of music and mental health and provide examples of potential interactions.

Music and reward

The strong effect of music on our emotions has been clearly grounded in its robust activation of reward circuitry in the brain, and motivational and hedonic effects of music listening have been shown to be specifically modulated by dopamine [ 16 , 105 , 136 ]. The prevalence of reward and dopaminergic dysfunction in mental illness makes this a rich area for future studies. For example, emotional responses to music might be used as a substitute for reward circuit deficiencies in depression, and it is intriguing to consider if music listening or music engagement could potentiate such function [ 137 , 138 ].

Music and brain network dynamics

The search for neuronally based biomarkers of aspects of mental illness has been a central thrust within the field [ 139 ], holding promise for the understanding of heterogeneity within disorders and identification of common mechanistic pathways [ 140 ]. A thorough review is beyond the scope of this paper, but several points of contact can be highlighted that might suggest neuro-mechanistic mediators of musical effects on mental health. For example, neurofeedback-directed upregulation of activity in emotion circuitry has been proposed as a therapy for MDD [ 141 ]. Given the emotional effects of music, there is potential for using musical stimuli as an adjuvant, or as a more actively patient-controlled output target for neurofeedback. Growing interest in measures of the dynamic complexity of brain activity in health and disease as measured by magnetic resonance imaging or magneto/electroencephalography (M/EEG) [ 142 ] provides a second point of contact, with abnormalities in dynamic complexity suggested as indicative of mental illness [ 143 ], while music engagement has been suggested to reflect and perhaps affect dynamic complexity [ 144 , 145 ].

The caveats identified in this review apply equally to such neuro-mechanistic studies [ 146 ]. High-quality experimental design (involving appropriate controls and randomized design) has been repeatedly shown to be critical to providing reliable evidence for non-music outcomes of music engagement [ 103 ]. For such studies to have maximal impact, analysis of M/EEG activity not at the scalp level, but at the source level, has been shown to improve the power of biomarkers, and their mechanistic interpretability [ 147 , 148 ]. Moreover, as with genetic influences that typically influence a trait through a multitude of small individual effects [ 149 ], the neural underpinnings of music-mental health associations may be highly multivariate. In the longer term, leveraging large-scale studies and large-scale data standardization and aggregation hold the promise of gleaning deeper cross-domain insights, for which current experimentalists can prepare by adopting standards for the documentation, annotation, and storage of data [ 150 ].

Biobanks and electronic health records

Finally, the use of EHR databases can be useful in quantifying associations between music engagement and mental health in large samples. EHR databases can include hundreds of thousands of records and allow for examination with International Statistical Classification of Diseases and Related Health Problems codes, including MDD, SUD, and schizophrenia diagnoses. This would allow for powerful estimates of music-mental health associations, and exploration of music engagement with other health outcomes.

The principal roadblock to this type of research is that extensive music phenotypes are not readily available in EHRs. However, there are multiple ways to bypass this limitation. First, medical records can be scraped using text-mining tools to identify cases of musician-related terms (e.g., “musician”, “guitarist”, “violinist”). For example, the phenome-wide association study described earlier [ 55 ] compared musician cases and controls identified in a large EHR database through text-mining of medical records and validated with extensive manual review charts. This study was highly powered to detect associations with internalizing and thought disorders (but showed null or protective effects for musicians for SUDs). Many EHR databases also include genomic data, allowing for integration with genetic models even in the absence of music data (e.g., exploring whether individuals with strong genetic predispositions for musical ability are at elevated or reduced risk for specific health diagnosis).

EHRs could also be used as recruitment tools, allowing researchers to collect additional data for relevant music engagement variables and compare with existing mental health diagnoses without having to conduct their own diagnostic interviews. These systems are not only relevant to individual differences research but could also be used to identify patients for possible enrollment in intervention studies. Furthermore, if recruitment for individual differences or intervention studies is done in patient waiting rooms of specific clinics, researchers can target specific populations of interest, have participants complete some relevant questionnaires while they wait, and be granted access to medical record data without having to conduct medical interviews themselves.

Concluding remarks

Music engagement, a uniquely human trait which has a powerful impact on our everyday experience, is deeply tied with our social and cultural identities as well as our personality and cognition. The relevance of music engagement to mental health, and its potential use as a therapeutic tool, has been studied for decades, but this research had not yet cohered into a clear picture. Our scoping review and framework integrated across a breadth of smaller literatures (including extant reviews and meta-analyses) relating music engagement to mental health traits and treatment effects, though it was potentially limited due to the lack of systematic literature search or formal quality appraisal of individual studies. Taken together, the current body of literature suggests that music engagement may provide an outlet for individuals who are experiencing internalizing, externalizing, or thought disorder problems, potentially supporting emotion regulation through multiple neurobiological pathways (e.g., reward center activity). Conducting more rigorous experimental intervention studies, improving reporting standards, and harnessing large-scale population-wide data in combination with new genetic analytic methods will help us achieve a better understanding of how music engagement relates to these mental health traits. We have presented a framework that illustrates why it will be vital to consider genetic and environmental risk factors when examining these associations, leading to new avenues for understanding the mechanisms by which music engagement and existing risk factors interact to support mental health and well-being.

Mankel K, Bidelman GM. Inherent auditory skills rather than formal music training shape the neural encoding of speech. Proc Natl Acad Sci. 2018;115:13129–34.

Article   CAS   PubMed   PubMed Central   Google Scholar  

Swaminathan S, Schellenberg EG. Musical competence is predicted by music training, cognitive abilities, and personality. Sci Rep. 2018;8:9223.

Article   PubMed   PubMed Central   CAS   Google Scholar  

Mosing MA, Pedersen NL, Madison G, Ullen F. Genetic pleiotropy explains associations between musical auditory discrimination and intelligence. PLos One. 2014;9:e113874.

Weinberg MK, Joseph D. If you’re happy and you know it: Music engagement and subjective wellbeing. Psychol Music. 2017;45:257–67.

Article   Google Scholar  

Theorell TP, Lennartsson AK, Mosing MA, Ullen F. Musical activity and emotional competence - a twin study. Front Psychol. 2014;5:774.

Article   PubMed   PubMed Central   Google Scholar  

Cirelli LK, Trehub SE, Trainor LJ. Rhythm and melody as social signals for infants. Ann N Y Acad Sci. 2018;1423:66–72.

Zentner M, Eerola T. Rhythmic engagement with music in infancy. Proc Natl Acad Sci. 2010;107:5768–73.

Lense MD, Beck S, Liu C, Pfeiffer R, Diaz N, Lynch M, et al. Parents, peers, and musical play: Integrated parent-child music class program supports community participation and well-being for families of children with and without Autism Spectrum Disorder. Front Psychol. 2020;11:11.

Honing H. On the biological basis of musicality. Ann N Y Acad Sci. 2018;1423:51–6.

Maratos AS, Gold C, Wang X, Crawford MJ. Music therapy for depression. Cochrane Database Syst Rev. 2008;1:CD004517.

Ansdell G, Meehan J. “Some Light at the End of the Tunnel”: exploring Users’ evidence for the effectiveness of music therapy in adult mental health settings. Music Med. 2010;2:29–40.

Khalfa S, Bella SD, Roy M, Peretz I, Lupien SJ. Effects of relaxing music on salivary cortisol level after psychological stress. Ann N Y Acad Sci. 2003;999:374–6.

Article   PubMed   Google Scholar  

McKinney CH, Antoni MH, Kumar M, Tims FC, McCabe PM. Effects of guided imagery and music (GIM) therapy on mood and cortisol in healthy adults. Health Psychol. 1997;16:390–400.

Article   CAS   PubMed   Google Scholar  

Chanda ML, Levitin DJ. The neurochemistry of music. Trends Cogn Sci. 2013;17:179–93.

Olff M, Koch SB, Nawijn L, Frijling JL, Van Zuiden M, Veltman DJ. Social support, oxytocin, and PTSD. Eur J Psychotraumatol. 2014;5:26513.

Ferreri L, Mas-Herrero E, Zatorre RJ, Ripollés P, Gomez-Andres A, Alicart H, et al. Dopamine modulates the reward experiences elicited by music. Proc Natl Acad Sci. 2019;116:3793–8.

Evers S, Suhr B. Changes of the neurotransmitter serotonin but not of hormones during short time music perception. Eur Arch Psychiatry Clin Neurosci. 2000;250:144–7.

Blum K, Simpatico T, Febo M, Rodriquez C, Dushaj K, Li M, et al. Hypothesizing music intervention enhances brain functional connectivity involving dopaminergic recruitment: Common neuro-correlates to abusable drugs. Mol Neurobiol. 2017;54:3753–58.

Kotov R, et al. The hierarchical Taxonomy of psychopathology (HiTOP): a dimensional alternative to traditional nosologies. J Abnorm Psychol. 2017. https://doi.org/10.1037/abn0000258 .

Caspi A, Houts RM, Belsky DW, Goldman-Mellor SJ, Harrington H, Israel S, et al. The p factor: one general psychopathology factor in the structure of psychiatric disorders? Clin Psychological Sci. 2014;2:119–37.

Whiteside SP, Lynam DR. Understanding the role of impulsivity and externalizing psychopathology in alcohol abuse: Application of the UPPS impulsive behavior scale. Exp Clin Psychopharmacol. 2003;11:210–7.

Tackett JL, Lahey BB, van Hulle C, Waldman I, Krueger RF, Rathouz PJ. Common genetic influences on negative emotionality and a general psychopathology factor in childhood and adolescence. J Abnorm Psychol. 2013;122:1142–53.

Young SE, Friedman NP, Miyake A, Willcutt EG, Corley RP, Haberstick BC, et al. Behavioral disinhibition: liability for externalizing spectrum disorders and its genetic and environmental relation to response inhibition across adolescence. J Abnorm Psychol. 2009;118:117–30.

Gustavson DE, Franz CE, Panizzon MS, Lyons MJ, Kremen WS. Internalizing and externalizing psychopathology in middle age: Genetic and environmental architecture and stability of symptoms over 15 to 20 years. Psychological Med. 2019;50:1–9.

Google Scholar  

Martin J, Taylor MJ, Lichtenstein P. Assessing the evidence for shared genetic risks across psychiatric disorders and traits. Psychological Med. 2018;48:1759–74.

Panteleeva Y, Ceschi G, Glowinski D, Courvoisier DS, Grandjean D. Music for anxiety? Meta-analysis of anxiety reduction in non-clinical samples. Psychol Music. 2017;46:473–87.

Leubner D, Hinterberger T. Reviewing the effectiveness of music interventions in treating depression. Front Psychol. 2017;8:1109.

Perkins R, Mason-Bertrand A, Fancourt D, Baxter L, Williamon A. How participatory music engagement supports mental well-being: a meta-ethnography. Qualitative Health Res. 2020. https://doi.org/10.1177/1049732320944142 .

Lilienfeld SO. Comorbidity between and within childhood externalizing and internalizing disorders: reflections and directions. J. Abnorm Child Psychol. 2003;31:285–91.

Kendler KS, Myers J. The boundaries of the internalizing and externalizing genetic spectra in men and women. Psychological Med. 2014;44:647–55.

Article   CAS   Google Scholar  

Robb SL, Burns DS, Carpenter JS. Reporting guidelines for music-based interventions. J Health Psychol. 2011;16:342–52.

Robb SL, Hanson-Abromeit D, May L, Hernandez-Ruiz E, Allison M, Beloat A, et al. Reporting quality of music intervention research in healthcare: a systematic review. Complement Ther Med. 2018;38:24–41.

Carr C, Odell-Miller H, Priebe S. A systematic review of music therapy practice and outcomes with acute adult psychiatric in-patients. PLos One. 2013;8:e70252.

Miranda D, Blais-Rochette C. Neuroticism and emotion regulation through music listening: a meta-analysis. Musica Sci. 2018. https://doi.org/10.1177/1029864918806341 .

Miranda D. The emotional bond between neuroticism and music. Psychomusicology: Music, Mind, Brain. 2019. https://doi.org/10.1037/pmu0000250 .

Landis-Shack N, Heinz AJ, Bonn-Miller MO. Music therapy for posttraumatic stress in adults: a theoretical review. Psychomusicology. 2017;27:334–342.

Schäfer K, Saarikallio S, Eerola T. Music may reduce loneliness and act as social surrogate for a friend: evidence from an experimental listening study. Music Sci. 2020;3. https://doi.org/10.1177/2059204320935709 .

Braun Janzen T, Al Shirawi MI, Rotzinger S, Kennedy SH, Bartel L. A pilot study investigating the effect of music-based intervention on depression and anhedonia. Front Psychol. 2019;10:1038.

Biasutti M, Mangiacotti A. Music training improves depressed mood symptoms in elderly people: a randomized controlled trial. Int J Aging Hum Dev. 2019. https://doi.org/10.1177/0091415019893988 .

Castillo-Pérez S, Gómez-Pérez V, Velasco MC, Pérez-Campos E, Mayoral M-A. Effects of music therapy on depression compared with psychotherapy. Arts Psychother. 2010;37:387–390.

Hendricks CB, Robinson B, Bradley LJ, Davis K. Using music techniques to treat adolescent depression. J Humanist Counseling. 1999;38:39–46.

Chwalek CM, McKinney CH. The use of dialectical behavior therapy (DBT) in music therapy: a sequential explanatory study. J. Music Ther. 2015;52:282–318.

Tang YY, Yang L, Leve LD, Harold GT. Improving executive function and its neurobiological mechanisms through a mindfulness-based intervention: advances within the field of developmental neuroscience. Child Dev. Perspect. 2012;6:361–66.

PubMed   PubMed Central   Google Scholar  

Didonna F. Mindfulness-based interventions in an inpatient setting. In: Clinical handbook of mindfulness. Springer, New York, NY; 2009. p. 447–62.

Chin T, Rickard NS. Beyond positive and negative trait affect: flourishing through music engagement. Psychol Well-Being. 2014;4:25.

Miranda D, Claes M. Personality traits, music preferences and depression in adolescence. Int J Adolescence Youth. 2008;14:277–98.

Fancourt D, Garnett C, Spiro N, West R, Mullensiefen D. How do artistic creative activities regulate our emotions? Validation of the Emotion Regulation Strategies for Artistic Creative Activities Scale (ERS-ACA). PLos One. 2019;14:e0211362.

Saarikallio S. Development and validation of the brief music in mood regulation scale (B-MMR). Music Percept. 2012;30:97–105.

Saarikallio S, Gold C, McFerran K. Development and validation of the healthy-unhealthy music scale. Child Adolesc Ment Health. 2015;20:210–17.

Groarke JM, Hogan MJ. Development and psychometric evaluation of the adaptive functions of music listening scale. Front Psychol. 2018;9:516.

Thomson CJ, Reece JE, Di Benedetto M. The relationship between music-related mood regulation and psychopathology in young people. Musica Sci. 2014;18:150–65.

Shifriss R, Bodner E, Palgi Y. When you’re down and troubled: views on the regulatory power of music. Psychol Music. 2014;43:793–807.

Garrido S, Schubert E. Moody melodies: do they cheer us up? A study of the effect of sad music on mood. Psychol Music. 2013;43:244–261.

Wesseldijk LW, Ullen F, Mosing MA. The effects of playing music on mental health outcomes. Sci Rep. 2019;9:12606.

Niarchou M, Lin G, Lense MD, Gordon RL, Davis LK. The medical signature of musicians: a phenome-wide association study using an electronic health record database. medRxiv. 2020;10:51. https://doi.org/10.1101/2020.08.14.20175109 .

Vaag J, Sund ER, Bjerkeset O. Five-factor personality profiles among Norwegian musicians compared to the general workforce. Musica Sci. 2017;22:434–445.

Fancourt D, Garnett C, Müllensiefen D. The relationship between demographics, behavioral and experiential engagement factors, and the use of artistic creative activities to regulate emotions. Psychol Aesthet Creat Arts. 2020.(advance online publication)

Mak HW, Fancourt D. Longitudinal associations between ability in arts activities, behavioural difficulties and self-esteem: Analyses from the 1970 British Cohort Study. Sci Rep. 2019;9:14236.

West SG, Taylor AB, Wu W. Model fit and model selection in structural equation modeling. In: Hoyle RH, editor. Handbook of structural equation modeling. New York, NY: The Guilford Press; 2012. p. 209–31.

Weiss B, Susser K, Catron T. Common and specific features of childhood psychopathology. J Abnorm Psychol. 1998;107:118–27.

Krueger RF, McGue M, Iacono WG. The higher-order structure of common DSM mental disorders: Internalization, externalization, and their connections to personality. Personal Individ Differences. 2001;30:1245–159.

Morse S, et al. Audio therapy significantly attenuates aberrant mood in residential patient addiction treatment: putative activation of dopaminergic pathways in the meso-limbic reward circuitry of humans. J Addict Res Ther. 2011;3:2.

Baker FA, Gleadhill LM, Dingle GA. Music therapy and emotional exploration: exposing substance abuse clients to the experiences of non-drug-induced emotions. Arts Psychother. 2007;34:321–330.

Dingle GA, Gleadhill L, Baker FA. Can music therapy engage patients in group cognitive behaviour therapy for substance abuse treatment? Drug Alcohol Rev. 2008;27:190–6.

Hohmann L, Bradt J, Stegemann T, Koelsch S. Effects of music therapy and music-based interventions in the treatment of substance use disorders: A systematic review. PLos One. 2017;12:e0187363.

Swope PM. Effects of learning the drums on inattention, vigilance, and sustained attention in adolescents with ADHD. Spalding University; 2018.

Zimmermann MB, Diers K, Strunz L, Scherbaum N, Mette C. Listening to Mozart improves current mood in adult ADHD: a randomized controlled pilot study. Front Psychol. 2019;10:1104.

Madjar N, Gazoli R, Manor I, Shoval G. Contrasting effects of music on reading comprehension in preadolescents with and without ADHD. Psychiatry Res. 2020;291:113207.

Pelham WE, Waschbusch DA, Hoza B, Gnagy EM, Greiner AR, Sams SE, et al. Music and video as distractors for boys with ADHD in the classroom: comparison with controls, individual differences, and medication effects. J Abnorm Child Psychol. 2011;39:1085–98.

Porter S, McConnell T, McLaughlin K, Lynn F, Cardwell C, Braiden HJ, et al. Music therapy for children and adolescents with behavioural and emotional problems: a randomised controlled trial. J Child Psychol Psychiatry. 2017;58:586–94.

Chong HJ, Kim SJ. Education-oriented music therapy as an after-school program for students with emotional and behavioral problems. Arts Psychother. 2010;37:190–96.

Rickson DJ, Watkins WG. Music therapy to promote prosocial behaviors in aggressive adolescent boys—a pilot study. J. Music Ther. 2003;40:283–301.

Montello L, Coons EE. Effects of active versus passive group music therapy on preadolescents with emotional, learning, and behavioral disorders. J Music Ther. 1999;35:49–67.

Mulder J, Ter Bogt TF, Raaijmakers QA, Gabhainn SN, Monshouwer K, Vollebergh WA. The soundtrack of substance use: Music preference and adolescent smoking and drinking. Subst Use Misuse. 2009;44:514–31.

Mulder J, Ter Bogt TF, Raaijmakers QA, Nic Gabhainn S, Monshouwer K, Vollebergh WA. Is it the music? Peer substance use as a mediator of the link between music preferences and adolescent substance use. J Adolescence. 2010;33:387–94.

Chen MJ, Miller BA, Grube JW, Waiters ED. Music, substance use, and aggression. J Stud Alcohol. 2006;67:373–81.

ter Bogt TF, Gabhainn SN, Simons-Morton BG, Ferreira M, Hublet A, Godeau E, et al. Dance is the new metal: adolescent music preferences and substance use across Europe. Subst Use Misuse. 2012;47:130–42.

Slater JL, Tate MC. Timing deficits in ADHD: insights from the neuroscience of musical rhythm. Front Comput Neurosci. 2018;12:51.

Carrer LR. Music and sound in time processing of children with ADHD. Front Psychiatry. 2015;6:127.

Miksza P. Relationships among impulsivity, achievement goal motivation, and the music practice of high school wind players. Bull Council Res Music Educ. 2009;180:9–27.

Miksza P. Relationships among achievement goal motivation, impulsivity, and the music practice of collegiate brass and woodwind players. Psychol Music. 2010;39:50–67.

Geretsegger M, Mössler KA, Bieleninik Ł, Chen XJ, Heldal TO, Gold C. Music therapy for people with schizophrenia and schizophrenia-like disorders. Cochrane Database Syst Rev. 2017;5:CD004025.

PubMed   Google Scholar  

Volpe U, Gianoglio C, Autiero L, Marino ML, Facchini D, Mucci A, et al. Acute effects of music therapy in subjects with psychosis during inpatient treatment. Psychiatry. 2018;81:218–27.

Pavlov A, Kameg K, Cline TW, Chiapetta L, Stark S, Mitchell AM. Music therapy as a nonpharmacological intervention for anxiety in patients with a thought disorder. Issues Ment Health Nurs. 2017;38:285–8.

Haugwitz, B. (2021). Music therapy in the early detection and indicated prevention in persons at risk of bipolar disorders: state of knowledge and potential. Br J Music Ther. 2021. https://doi.org/10.1177/1359457521997386 .

Power RA, Steinberg S, Bjornsdottir G, Rietveld CA, Abdellaoui A, Nivard MM, et al. Polygenic risk scores for schizophrenia and bipolar disorder predict creativity. Nat Neurosci. 2015;18:953–5.

Fancourt D, Ockelford A, Belai A. The psychoneuroimmunological effects of music: a systematic review and a new model. Brain Behav. Immun. 2014;36:15–26.

Baker FA, Metcalf O, Varker T, O’Donnell M. A systematic review of the efficacy of creative arts therapies in the treatment of adults with PTSD. Psychol Trauma. 2018;10:643–51.

Lense MD, Camarata S. PRESS-Play: musical engagement as a motivating platform for social interaction and social play in young children with ASD. Music Sci. 2020. https://doi.org/10.1177/2059204320933080 .

Walker EF, Diforio D. Schizophrenia: a neural diathesis-stress model. Psychological Rev. 1997;104:667–85.

Zuckerman M, Riskind JH. Vulnerability to psychopathology: a biosocial model. J Cogn Psychother. 2000;14:407–8.

Trucco EM, Madan B, Villar M. The impact of genes on adolescent substance use: a developmental perspective. Curr Addict Rep. 2019;6:522–531.

Butkovic A, Ullen F, Mosing MA. Personality related traits as predictors of music practice: underlying environmental and genetic influences. Personal Individ Differences. 2015;74:133–8.

Coon H, Carey G. Genetic and environmental determinants of musical ability in twins. Behav Genet. 1989;19:183–93.

Ullén F, Mosing MA, Holm L, Eriksson H, Madison G. Psychometric properties and heritability of a new online test for musicality, the Swedish Musical Discrimination Test. Personal Individ Differences. 2014;63:87–93.

Hambrick DZ, Tucker-Drob EM. The genetics of music accomplishment: evidence for gene-environment correlation and interaction. Psychonomic Bull Rev. 2015;22:112–20.

Wesseldijk LW, Mosing MA, Ullen F. Gene-environment interaction in expertise: the importance of childhood environment for musical achievement. Dev Psychol. 2019;55:1473–9.

Ullen F, Mosing MA, Madison G. Associations between motor timing, music practice, and intelligence studied in a large sample of twins. Ann N Y Acad Sci. 2015;1337:125–9.

Botvin GJ. Substance abuse prevention: theory, practice, and effectiveness. Crime Justice. 1990;13:461–519.

Fredricks JA, Simpkins S, Eccles JS. Family socialization, gender, and participation in sports and instrumental music. In: Developmental pathways through middle childhood. Mahwah, NJ: Psychology Press; 2006. p. 53–74.

Maes HH, Neale MC, Kirkpatrick RM, Kendler KS. Using multimodal inference/model averaging to model causes of covariation between variables in twins. Behav Genet. 2020. https://doi.org/10.1007/s10519-020-10026-8 .

Sala G, Gobet F. When the music’s over. Does music skill transfer to children’s and young ado escents’ cognitive and academic skills? A meta-analysis. Educ Res Rev. 2017;20:55–67.

Salimpoor VN, Zald DH, Zatorre RJ, Dagher A, McIntosh AR. Predictions and the brain: how musical sounds become rewarding. Trends Cogn Sci. 2015;19:86–91.

Loui P, Patterson S, Sachs ME, Leung Y, Zeng T, Przysinda E. White matter correlates of musical anhedonia: Implications for evolution of music. Front Psychol. 2017;8:1664.

Salimpoor VN, van den Bosch I, Kovacevic N, McIntosh AR, Dagher A, Zatorre RJ. Interactions between the nucleus accumbens and auditory cortices predict music reward value. Science. 2013;340:216–9.

Zatorre RJ, Salimpoor VN. From perception to pleasure: music and its neural substrates. Proc Natl Acad Sci. 2013;110:10430–7. Suppl 2

Alluri V, Brattico E, Toiviainen P, Burunat I, Bogert B, Numminen J, et al. Musical expertise modulates functional connectivity of limbic regions during continuous music listening. Psychomusicology. 2015;25:443–54.

Vanyukov MM, Tarter RE, Kirisci L, Kirillova GP, Maher BS, Clark DB. Liability to substance use disorders: 1. Common mechanisms and manifestations. Neurosci Biobehav Rev. 2003;27:507–515.

Volkow ND, Morales M. The brain on drugs: from reward to addiction. Cell. 2015;162:712–25.

Wise RA. Dopamine and reward: the anhedonia hypothesis 30 years on. Neurotox Res. 2008;14:169–83.

Koob GF, Volkow ND. Neurocircuitry of addiction. Neuropsychopharmacology. 2010;35:217–38.

Mukherjee S, Coque L, Cao JL, Kumar J, Chakravarty S, Asaithamby A, et al. Knockdown of Clock in the ventral tegmental area through RNA interference results in a mixed state of mania and depression-like behavior. Biol Psychiatry. 2010;68:503–11.

Kaufling J. Alterations and adaptation of ventral tegmental area dopaminergic neurons in animal models of depression. Cell Tissue Res. 2019;377:59–71.

Small KM, Nunes E, Hughley S, Addy NA. Ventral tegmental area muscarinic receptors modulate depression and anxiety-related behaviors in rats. Neurosci Lett. 2016;616:80–5.

Kokal I, Engel A, Kirschner S, Keysers C. Synchronized drumming enhances activity in the caudate and facilitates prosocial commitment-if the rhythm comes easily. PLos One. 2011;6:e27272.

Goldstein A. Thrills in response to music and other stimuli. Physiological Psychol. 1980;8:126–9.

Moore E, Schaefer RS, Bastin ME, Roberts N, Overy K. Can musical training influence brain connectivity? Evidence from diffusion tensor MRI. Brain Sci. 2014;4:405–27.

Imfeld A, Oechslin MS, Meyer M, Loenneker T, Jancke L. White matter plasticity in the corticospinal tract of musicians: a diffusion tensor imaging study. Neuroimage. 2009;46:600–7.

Han Y, Yang H, Lv YT, Zhu CZ, He Y, Tang HH, et al. Gray matter density and white matter integrity in pianists’ brain: a combined structural and diffusion tensor MRI study. Neurosci Lett. 2009;459:3–6.

Schmithorst VJ, Wilke M. Differences in white matter architecture between musicians and non-musicians: a diffusion tensor imaging study. Neurosci Lett. 2002;321:57–60.

Hudziak JJ, Albaugh MD, Ducharme S, Karama S, Spottswood M, Crehan E, et al. Cortical thickness maturation and duration of music training: Health-promoting activities shape brain development. J Am Acad Child Adolesc Psychiatry. 2014;53:1153–61.

Gingras B, Honing H, Peretz I, Trainor LJ, Fisher SE. Defining the biological bases of individual differences in musicality. Philos Trans R Soc B. 2015;370:20140092.

Heath AC, Kessler RC, Neale MC, Hewitt JK, Eaves LJ, Kendler KS. Testing hypotheses about direction of causation using cross-sectional family data. Behav Genet. 1993;23:29–50.

Coleman JRI, Peyrot WJ, Purves KL, Davis KAS, Rayner C, Choi SW, et al. Genome-wide gene-environment analyses of major depressive disorder and reported lifetime traumatic experiences in UK Biobank. Mol Psychiatry. 2020;25:247353–1446.

Domingue BW, Liu HX, Okbay A, Belsky DW. Genetic heterogeneity in depressive symptoms following the death of a spouse: polygenic score analysis of the US Health and Retirement Study. Am J Psychiatry. 2017;174:963–970.

Barcellos SH, Carvalho LS, Turley P. Education can reduce health differences related to genetic risk of obesity. Proc Natl Acad Sci USA. 2018;115:E9765–E9772.

Howard DM, Adams MJ, Shirali M, Clarke TK, Marioni RE, Davies G, et al. Genome-wide association study of depression phenotypes in UK Biobank identifies variants in excitatory synaptic pathways. Nat Commun. 2018;9:1470.

Okbay A, Baselmans BM, De Neve JE, Turley P, Nivard MG, Fontana MA, et al. Genetic variants associated with subjective well-being, depressive symptoms, and neuroticism identified through genome-wide analyses. Nat Genet. 2016;48:624–33.

Walters RK, Polimanti R, Johnson EC, McClintick JN, Adams MJ, Adkins AE, et al. Transancestral GWAS of alcohol dependence reveals common genetic underpinnings with psychiatric disorders. Nat Neurosci. 2018;21:1656–69.

Niarchou M, et al. Unraveling the genetic architecture of music rhythm. https://www.biorxiv.org/content/10.1101/836197v1 . 2019 ; 29:S62.

Purcell S. Variance components models for gene-environment interaction in twin analysis. Twin Res. 2002;5:554–71.

Barch DM, Albaugh MD, Avenevoli S, Chang L, Clark DB, Glantz MD, et al. Demographic, physical and mental health assessments in the adolescent brain and cognitive development study: rationale and description. Dev Cogn Neurosci. 2018;32:55–66.

Uban KA, Horton MK, Jacobus J, Heyser C, Thompson WK, Tapert SF, et al. Biospecimens and the ABCD study: rationale, methods of collection, measurement and early data. Dev Cogn Neurosci. 2018;32:97–106.

Mullensiefen D, Gingras B, Musil J, Stewart L. The musicality of non-musicians: an index for assessing musical sophistication in the general population. PLos One. 2014;9:e89642.

Blood AJ, Zatorre RJ. Intensely pleasurable responses to music correlate with activity in brain regions implicated in reward and emotion. Proc Natl Acad Sci. 2001;98:11818–23.

Jenkins LM, Skerrett KA, DelDonno SR, Patrón VG, Meyers KK, Peltier S, et al. Individuals with more severe depression fail to sustain nucleus accumbens activity to preferred music over time. Psychiatry Res Neuroimaging. 2018;275:21–7.

Belfi AM, Loui P. Musical anhedonia and rewards of music listening: current advances and a proposed model. Ann N Y Acad Sci. 2020;1464:99–114.

Insel T, et al. Research domain criteria (RDoC): toward a new classification framework for research on mental disorders. Am Psychiatric Assoc. 2010;167:748–51.

Loo SK, McGough JJ, McCracken JT, Smalley SL. Parsing heterogeneity in attention-deficit hyperactivity disorder using EEG-based subgroups. J Child Psychol Psychiatry. 2018;59:223–31.

Zotev V, Mayeli A, Misaki M, Bodurka J. Emotion self-regulation training in major depressive disorder using simultaneous real-time fMRI and EEG neurofeedback. Neuroimage Clin. 2020;27:102331.

Yang AC, Jann K, Michel CM, Wang DJJ. Editorial: advances in multi-scale analysis of brain complexity. Front Neurosci. 2020;14:337.

Lin C, Lee SH, Huang CM, Chen GY, Ho PS, Liu HL, et al. Increased brain entropy of resting-state fMRI mediates the relationship between depression severity and mental health-related quality of life in late-life depressed elderly. J Affect Disord. 2019;250:270–7.

Bhattacharya J, Lee EJ. Modulation of EEG theta band signal complexity by music therapy. Int J Bifurc Chaos. 2016;26:1650001.

Carpentier SM, McCulloch AR, Brown TM, Faber S, Ritter P, Wang Z, et al. Complexity matching: brain signals mirror environment information patterns during music listening and reward. J Cogn Neurosci. 2020;32:734–45.

Etkin A. A reckoning and research agenda for neuroimaging in psychiatry. Am J Psychiatry. 2019;176:507–11.

Rissling AJ, Makeig S, Braff DL, Light GA. Neurophysiologic markers of abnormal brain activity in schizophrenia. Curr Psychiatry Rep. 2010;12:572–8.

Loo SK, Makeig S. Clinical utility of EEG in attention-deficit/hyperactivity disorder: a research update. Neurotherapeutics. 2012;9:569–87.

Boyle EA, Li YI, Pritchard JK. An expanded view of complex traits: from polygenic to omnigenic. Cell. 2017;169:1177–86.

Sivagnanam S, Yoshimoto K, Carnevale T, Nadeau D, Kandes M, Petersen T, et al. Neuroscience Gateway enabling large scale modeling and data processing in neuroscience research. In: Practice and Experience in Advanced Research Computing. Association for Computing Machinery, Portland, OR, USA ; 2020. p.510–513.

Download references

Acknowledgements

This work was supported by NIH grants DP2HD098859, R01AA028411, R61MH123029, R21DC016710, U01DA04112, and R03AG065643, National Endowment for the Arts (NEA) research lab grants 1863278-38 and 1855526-38, and National Science Foundation grant 1926794. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or National Endowment for the Arts. The authors would like to thank Navya Thakkar and Gabija Zilinskaite for their assistance.

Author information

Authors and affiliations.

Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA

• Daniel E. Gustavson

Vanderbilt Genetics Institute, Vanderbilt University Medical Center, Nashville, TN, USA

Daniel E. Gustavson, Reyna L. Gordon & Miriam D. Lense

Department of Otolaryngology – Head & Neck Surgery, Vanderbilt University Medical Center, Nashville, TN, USA

Peyton L. Coleman & Reyna L. Gordon

Swartz Center for Computational Neuroscience, Institute for Neural Computation, University of California, San Diego, La Jolla, CA, USA

John R. Iversen

Department of Human and Molecular Genetics, Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond, VA, USA

• Hermine H. Maes

Department of Psychiatry, Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond, VA, USA

Massey Cancer Center, Virginia Commonwealth University, Richmond, VA, USA

Vanderbilt Brain Institute, Vanderbilt University, Nashville, TN, USA

Reyna L. Gordon & Miriam D. Lense

The Curb Center, Vanderbilt University, Nashville, TN, USA

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Daniel E. Gustavson .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Gustavson, D.E., Coleman, P.L., Iversen, J.R. et al. Mental health and music engagement: review, framework, and guidelines for future studies. Transl Psychiatry 11 , 370 (2021). https://doi.org/10.1038/s41398-021-01483-8

Download citation

Received : 23 November 2020

Revised : 03 June 2021

Accepted : 10 June 2021

Published : 22 June 2021

DOI : https://doi.org/10.1038/s41398-021-01483-8

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

• Laura W. Wesseldijk
• Miriam A. Mosing

Translational Psychiatry (2023)

Effects of music-based interventions on cancer-related pain, fatigue, and distress: an overview of systematic reviews

• Ana Trigueros-Murillo
• Javier Martinez-Calderon
• Alberto Marcos Heredia-Rizo

Supportive Care in Cancer (2023)

A Systematic Review of Music-Based Interventions to Improve Treatment Engagement and Mental Health Outcomes for Adolescents and Young Adults

• Aaron H. Rodwin
• Rei Shimizu
• Michelle R. Munson

Child and Adolescent Social Work Journal (2023)

Heritability of Childhood Music Engagement and Associations with Language and Executive Function: Insights from the Adolescent Brain Cognitive Development (ABCD) Study

• Srishti Nayak

Behavior Genetics (2023)

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

We use cookies to give you the best experience possible. By continuing we’ll assume you’re on board with our cookie policy

• A Research Guide
• Research Paper Topics

120 Music Research Paper Topics

How to choose a topic for music research paper:.

Music Theory Research Paper Topics:

• The influence of harmonic progression on emotional response in music
• Analyzing the use of chromaticism in the compositions of Johann Sebastian Bach
• The role of rhythm and meter in creating musical tension and release
• Examining the development of tonality in Western classical music
• Exploring the impact of cultural and historical context on musical form and structure
• Investigating the use of polyphony in Renaissance choral music
• Analyzing the compositional techniques of minimalist music
• The relationship between melody and harmony in popular music
• Examining the influence of jazz improvisation on contemporary music
• The role of counterpoint in the compositions of Ludwig van Beethoven
• Investigating the use of microtonality in experimental music
• Analyzing the impact of technology on music composition and production
• The influence of musical modes on the development of different musical genres
• Exploring the use of musical symbolism in film scoring
• Investigating the role of music theory in the analysis and interpretation of non-Western music

Music Industry Research Paper Topics:

• The impact of streaming services on music consumption patterns
• The role of social media in promoting and marketing music
• The effects of piracy on the music industry
• The influence of technology on music production and distribution
• The relationship between music and mental health
• The evolution of music genres and their impact on the industry
• The economics of live music events and festivals
• The role of record labels in shaping the music industry
• The impact of globalization on the music industry
• The representation and portrayal of gender in the music industry
• The effects of music streaming platforms on artist revenue
• The role of music education in fostering talent and creativity
• The influence of music videos on audience perception and engagement
• The impact of music streaming on physical album sales
• The role of music in advertising and brand marketing

Music Therapy Research Paper Topics:

• The effectiveness of music therapy in reducing anxiety in cancer patients
• The impact of music therapy on improving cognitive function in individuals with Alzheimer’s disease
• Exploring the use of music therapy in managing chronic pain
• The role of music therapy in promoting emotional well-being in children with autism spectrum disorder
• Music therapy as a complementary treatment for depression: A systematic review
• The effects of music therapy on stress reduction in pregnant women
• Examining the benefits of music therapy in improving communication skills in individuals with developmental disabilities
• The use of music therapy in enhancing motor skills rehabilitation after stroke
• Music therapy interventions for improving sleep quality in patients with insomnia
• Exploring the impact of music therapy on reducing symptoms of post-traumatic stress disorder (PTSD)
• The role of music therapy in improving social interaction and engagement in individuals with schizophrenia
• Music therapy as a non-pharmacological intervention for managing symptoms of dementia
• The effects of music therapy on pain perception and opioid use in hospitalized patients
• Exploring the use of music therapy in promoting relaxation and reducing anxiety during surgical procedures
• The impact of music therapy on improving quality of life in individuals with Parkinson’s disease

Music Psychology Research Paper Topics:

• The effects of music on mood and emotions
• The role of music in enhancing cognitive abilities
• The impact of music therapy on mental health disorders
• The relationship between music and memory recall
• The influence of music on stress reduction and relaxation
• The psychological effects of different genres of music
• The role of music in promoting social bonding and cohesion
• The effects of music on creativity and problem-solving abilities
• The psychological benefits of playing a musical instrument
• The impact of music on motivation and productivity
• The psychological effects of music on physical exercise performance
• The role of music in enhancing learning and academic performance
• The influence of music on sleep quality and patterns
• The psychological effects of music on individuals with autism spectrum disorder
• The relationship between music and personality traits

Music Education Research Paper Topics:

• The impact of music education on cognitive development in children
• The effectiveness of incorporating technology in music education
• The role of music education in promoting social and emotional development
• The benefits of music education for students with special needs
• The influence of music education on academic achievement
• The importance of music education in fostering creativity and innovation
• The relationship between music education and language development
• The impact of music education on self-esteem and self-confidence
• The role of music education in promoting cultural diversity and inclusivity
• The effects of music education on students’ overall well-being and mental health
• The significance of music education in developing critical thinking skills
• The role of music education in enhancing students’ teamwork and collaboration abilities
• The impact of music education on students’ motivation and engagement in school
• The effectiveness of different teaching methods in music education
• The relationship between music education and career opportunities in the music industry

Music History Research Paper Topics:

• The influence of African music on the development of jazz in the United States
• The role of women composers in classical music during the 18th century
• The impact of the Beatles on the evolution of popular music in the 1960s
• The cultural significance of hip-hop music in urban communities
• The development of opera in Italy during the Renaissance
• The influence of folk music on the protest movements of the 1960s
• The role of music in religious rituals and ceremonies throughout history
• The evolution of electronic music and its impact on contemporary music production
• The contribution of Latin American musicians to the development of salsa music
• The influence of classical music on film scores in the 20th century
• The role of music in the Civil Rights Movement in the United States
• The development of reggae music in Jamaica and its global impact
• The influence of Mozart’s compositions on the classical music era
• The role of music in the French Revolution and its impact on society
• The evolution of punk rock music and its influence on alternative music genres

Music Sociology Research Paper Topics:

• The impact of music streaming platforms on the music industry
• The role of music in shaping cultural identity
• Gender representation in popular music: A sociological analysis
• The influence of social media on music consumption patterns
• Music festivals as spaces for social interaction and community building
• The relationship between music and political activism
• The effects of globalization on local music scenes
• The role of music in constructing and challenging social norms
• The impact of technology on music production and distribution
• Music and social movements: A comparative study
• The role of music in promoting social change and social justice
• The influence of socioeconomic factors on music taste and preferences
• The role of music in constructing and reinforcing gender stereotypes
• The impact of music education on social and cognitive development
• The relationship between music and mental health: A sociological perspective

Classical Music Research Paper Topics:

• The influence of Ludwig van Beethoven on the development of classical music
• The role of women composers in classical music history
• The impact of Johann Sebastian Bach’s compositions on future generations
• The evolution of opera in the classical period
• The significance of Mozart’s symphonies in the classical era
• The influence of nationalism on classical music during the Romantic period
• The portrayal of emotions in classical music compositions
• The use of musical forms and structures in the works of Franz Joseph Haydn
• The impact of the Industrial Revolution on the production and dissemination of classical music
• The relationship between classical music and dance in the Baroque era
• The role of patronage in the development of classical music
• The influence of folk music on classical composers
• The representation of nature in classical music compositions
• The impact of technological advancements on classical music performance and recording
• The exploration of polyphony in the works of Johann Sebastian Bach

• Writing a Research Paper
• Research Paper Title
• Research Paper Sources
• Research Paper Problem Statement
• Research Paper Thesis Statement
• Hypothesis for a Research Paper
• Research Question
• Research Paper Outline
• Research Paper Summary
• Research Paper Prospectus
• Research Paper Proposal
• Research Paper Format
• Research Paper Styles
• AMA Style Research Paper
• MLA Style Research Paper
• Chicago Style Research Paper
• APA Style Research Paper
• Research Paper Structure
• Research Paper Cover Page
• Research Paper Abstract
• Research Paper Introduction
• Research Paper Body Paragraph
• Research Paper Literature Review
• Research Paper Background
• Research Paper Methods Section
• Research Paper Results Section
• Research Paper Discussion Section
• Research Paper Conclusion
• Research Paper Appendix
• Research Paper Bibliography
• APA Reference Page
• Annotated Bibliography
• Bibliography vs Works Cited vs References Page
• Research Paper Types
• What is Qualitative Research

Receive paper in 3 Hours!

• Choose the number of pages.
• Select your deadline.
• Complete your order.

Number of Pages

550 words (double spaced)

Deadline: 10 days left

By clicking "Log In", you agree to our terms of service and privacy policy . We'll occasionally send you account related and promo emails.

Sign Up for your FREE account

Behind The Scenes of Video and Audio Research at the Library for the Performing Arts

One of a number of reel-to-reel tape players at the Library for the Performing Arts.

The New York Public Library for the Performing Arts is home to hundreds of thousands of archival audio and video recordings of theater, film, dance, and music. While many of these recordings have been digitized, many more have not, even though we're constantly working to digitize more material. However, the Library makes the non-digitized material that's on physical media accessible to anyone who visits us: to explore these archives, patrons with a Special Collections account can access video and audio in digital terminals on our third floor. This research floor is what the Library’s Assistant Manager of Research Access Services, Benjamin Moreno, calls “a hidden gem in New York City.”

But, because this material isn’t digitized, the Library plays the physical media for any patron that requests to watch or listen to it. How does it work? In this video and interview, staff explain how the process works from start to finish.

Patrons can come to the third floor and put in a request with a librarian for a recording that they want to listen to or watch. Once they put in a request—say, for a cassette tape, or a DVD—they are assigned a specific computer. Library staff then locate the media, and bring it to our playback room where we have all kinds of audio and video electronics. The media is played from there and piped up to the patron sitting on the third floor.

“People always think there are little green elves working in the basement,” laughed Moreno, “but there aren't!” When the staff receive a request, they make their way to the stacks. They find the physical recording, find the right equipment to play it, and then connect that to the computer. If patrons come in as a group and want to watch or listen to the same recording, staff can connect the same physical equipment downstairs to multiple computers on the third floor. 

If a patron is watching a DVD, they might be able to fully control play, pause, rewind, or fast-forward functions from the terminal, but if they are listening to a record on vinyl, they wouldn’t have that same control. Instead, they can use the chat function on these computers to make requests. 

Angela Rapp uses the computer system that connects patrons to the physical media they want to listen to or watch.

The chat system, Moreno explained, can be used for many different purposes. Sometimes, patrons use it to make requests about starting the audio or film, playing the next disk, or pausing. Other times, they simply want to talk to the Library staff.

"Sometimes people are just very overjoyed to get what they came here for, and they’re very amazed about how the system works," he said. They’ll ask questions and make conversation. The staff even has a wall of printouts of funny chats patrons or staff have sent through the years!

This media playback system is a special and important part of the Library, allowing patrons to access incredible archival footage and sound that doesn't exist anywhere else. It remains a space for nostalgia, historical preservation, and learning, allowing patrons to dive into the rich and wonderful archives. 

Moreno hopes that more people will visit the Library for the Performing Arts in the coming years, not just dedicated researchers, but also young people and students. 

"You might get blown away," Moreno said. "You can definitely come here and leave learning something special from you've heard or watched."

To get started, head to the Library for the Performing Arts' research page .

Help | Advanced Search

Computer Science > Artificial Intelligence

Title: vision language models are blind.

Abstract: Large language models with vision capabilities (VLMs), e.g., GPT-4o and Gemini 1.5 Pro are powering countless image-text applications and scoring high on many vision-understanding benchmarks. We propose BlindTest, a suite of 7 visual tasks absurdly easy to humans such as identifying (a) whether two circles overlap; (b) whether two lines intersect; (c) which letter is being circled in a word; and (d) counting the number of circles in a Olympic-like logo. Surprisingly, four state-of-the-art VLMs are, on average, only 56.20% accurate on our benchmark, with \newsonnet being the best (73.77% accuracy). On BlindTest, VLMs struggle with tasks that requires precise spatial information and counting (from 0 to 10), sometimes providing an impression of a person with myopia seeing fine details as blurry and making educated guesses. Code is available at: this https URL

Submission history

Access paper:.

• HTML (experimental)
• Other Formats

References & Citations

• Google Scholar
• Semantic Scholar

BibTeX formatted citation

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

• Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .

Advertisement

Supported by

What We Know About the Global Microsoft Outage

Airlines to banks to retailers were affected in many countries. Businesses are struggling to recover.

• Share full article

By Eshe Nelson and Danielle Kaye

Eshe Nelson reported from London and Danielle Kaye from New York.

Across the world, critical businesses and services including airlines, hospitals, train networks and TV stations, were disrupted on Friday by a global tech outage affecting Microsoft users.

In many countries, flights were grounded, workers could not get access to their systems and, in some cases, customers could not make card payments in stores. While some of the problems were resolved within hours, many businesses, websites and airlines continued to struggle to recover.

What happened?

A series of outages rippled across the globe as information displays, login systems and broadcasting networks went dark.

The problem affecting the majority of services was caused by a flawed update by CrowdStrike , an American cybersecurity firm, whose systems are intended to protect users from hackers. Microsoft said on Friday that it was aware of an issue affecting machines running “CrowdStrike Falcon.”

But Microsoft had also said there was an earlier outage affecting U.S. users of Azure, its cloud service system. Some users may have been affected by both. Even as CrowdStrike sent out a fix, some systems were still affected by midday in the United States as businesses needed to make manual updates to their systems to resolve the issue.

George Kurtz, the president and chief executive of CrowdStrike, said on Friday morning that it could take some time for some systems to recover.

How a Software Update Crashed Computers Around the World

Here’s a visual explanation for how a faulty software update crippled machines.

How the airline cancellations rippled around the world (and across time zones)

Share of canceled flights at 25 airports on Friday

50% of flights

Ai r po r t

Bengalu r u K empeg o wda

Dhaka Shahjalal

Minneapolis-Saint P aul

Stuttga r t

Melbou r ne

Be r lin B r anden b urg

London City

Amsterdam Schiphol

Chicago O'Hare

Raleigh−Durham

B r adl e y

Cha r lotte

Reagan National

Philadelphia

1:20 a.m. ET

We are having trouble retrieving the article content.

Please enable JavaScript in your browser settings.

Thank you for your patience while we verify access. If you are in Reader mode please exit and  log into  your Times account, or  subscribe  for all of The Times.

Thank you for your patience while we verify access.

Already a subscriber?  Log in .

Want all of The Times?  Subscribe .

• Expert Systems
• Biosignal Processing
• Biological Science

Affective content analysis of music video clips

• November 2011
• Conference: Proceedings of the 1st international ACM workshop on Music information retrieval with user-centered and multimodal strategies, Scottsdale, AZ, USA, November 28 - December 01, 2011
• This person is not on ResearchGate, or hasn't claimed this research yet.

• École Polytechnique Fédérale de Lausanne

Abstract and Figures

Discover the world's research

• 25+ million members
• 160+ million publication pages
• 2.3+ billion citations
• PATTERN RECOGN LETT

• Hyung-Jeong Yang
• Guee-Sang Lee

• Xihao Zhang
• Rushuang Zhou
• Zhiguo Zhang
• Ho-Hsiang Wu

• Juan Pablo Bello

• Zhenzhong Chen

• NEUROCOMPUTING
• Dong-Ki Jeong
• Hyoung-Gook Kim
• Jin Young Kim

• Sander Koelstra

• Ioannis Patras

• Paul. Ekman
• Proceedings of SPIE
• Rainer Lienhart
• Image Process
• Wael Abd-Almageed
• Simon Moncrieff

• Recruit researchers
• Join for free
• Login Email Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google Welcome back! Please log in. Email · Hint Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google No account? Sign up

Nursing aide turned sniper: Thomas Crooks' mysterious plot to kill Trump

BUTLER, Pa. – Donald Trump and would-be assassin Thomas Crooks started on their violent collision course long before the former president's political rally ended in gunshots and death.

Crooks, 20, was a one-time registered Republican, a nursing home worker with no criminal record, shy in school, and living in a decent middle-class neighborhood in suburban Pennsylvania with his parents. Trump, 78, was eyeing Crooks' state as a key battleground – but not in the way that anyone envisioned on Saturday.

Riding high on polls showing that he's got a strong chance of toppling President Joe Biden, the former president had been campaigning for reelection in swing states, and Pennsylvania is a key prize. Trump won the state in 2016 but lost it four years later.

And on July 3, Trump's campaign announced he would hold a rally at the Butler Farm Show grounds, about 30 miles north of Pittsburgh.

"Pennsylvania has been ravaged by monumental surges in violent crime as a direct result of Biden’s and Democrats’ pro-criminal policies," Trump's campaign said in announcing the event, noting that when he's elected, he'll "re-establish law and order in Pennsylvania!"

The Saturday attack on Trump turned the heated rhetoric of the 2024 presidential campaign freshly violent. Authorities said bullets fired from Crooks' AR-15 style rifle about 150 yards away grazed Trump's ear, killed a rally attendee as he dove to protect his family, and critically wounded two others. Secret Service agents killed Crooks moments later.

Attack planned well in advance

Investigators are still seeking Crooks' motive – despite his Republican leanings, he had donated recently to a progressive voter-turnout campaign in 2021 – but indicated he'd planned the attack well in advance.

The shooting marks the first assassination attempt against a former or current U.S. president since President Ronald Reagan was injured in a March 1981 shooting at a Washington, D.C., hotel. 

There are many questions about why Crooks turned into a would-be presidential assassin, firing indiscriminately into hordes of political supporters.

FBI special agent Kevin Rojek said on a call with media that law enforcement located "a suspicious device" when they searched Crooks' vehicle and that it's being analyzed at the FBI crime lab.

"As far as the actions of the shooter immediately prior to the event and any interaction that he may have had with law enforcement, we're still trying to flesh out those details now," Rojek said.

None of Crooks' shocked neighbors or high school classmates described him as violent or that he in any way signaled he was intent on harming Trump. Sunday morning, reporters and curious locals swarmed the leafy streets of the home where Crooks lived with his parents in Bethel Park, about 50 miles from the shooting scene.

Those who knew him described a quiet young man who often walked to work at a nearby nursing home. One classmate said he was bullied and often ate alone in high school.

Sunday morning, neighbor Cathy Caplan, 45, extended her morning walk about a quarter mile to glimpse what was happening outside Crooks’ home.“It came on the morning news and I was like ‘I know that street,’” said Caplan, who works for the local school district. "It feels like something out of a movie.”

Dietary aide turned deadly killer

Authorities say they are examining Crooks' phone, social media and online activity for motivation. They said he carried no identification and his body had to be identified via DNA and biometric confirmation.

Although no possible motive has yet been released, Crooks nevertheless embodies the achingly familiar profile of an American mass shooter: a young white man, isolated from peers and armed with a high-powered rifle. His attack was one of at least 59 shootings in the United States on Saturday, according to the Gun Violence Archive.

According to records and online posts of the ceremony, Crooks graduated from Bethel Park High School, about 42 miles from Butler County, on June 3, 2022. That same day, Trump met briefly with investigators at his Mar-a-Lago club in Florida as they examined whether he improperly took classified documents with him when he left the White House.

A classmate remembered Crooks as a frequent target of bullies. Kids picked on him for wearing camouflage to class and his quiet demeanor, Jason Kohler, 21, said. Crooks usually ate lunch alone, Kohler said.

Crooks worked as a dietary aide at the Bethel Park Skilled Nursing and Rehabilitation, less than a mile from his home. In a statement provided to USA TODAY on Sunday, Marcie Grimm, the facility's administrator, said she was "shocked and saddened to learn of his involvement."

Neighbor Dean Sierka, 52, has known Crooks and his parents for years. The families live a few doors apart on a winding suburban street, and Sierka’s daughter, who attended elementary, middle and high school with Crooks, remembers him as quiet and shy. Sierka said they saw Crooks at least once a week, often when he was walking to the nursing home from his parents' three-bedroom brick house.

"You wouldn’t have expected this," Sierka said. "The parents and the family are all really nice people."

"It's crazy," he added.

Secret Service role: Did they do enough?

Founded in 1865, the Secret Service is supposed to stop this kind of attack, and dozens of agents were present Saturday. As the former president and presumptive Republican presidential nominee, Trump's public appearances are managed by the Secret Service, which works with local law enforcement to develop security plans and crowd-management protocols.

In the days before the event, the agency's experts would have scouted the location, identified security vulnerabilities, and designed a perimeter to keep Trump and rally attendees safe. Congress and the Secret Service are now investigating how Crooks was able to get so close to the former president, and several witnesses reported seeing him in the area with the gun before Trump took the stage.

As the event doors opened at 1 p.m., the temperature was already pushing close to 90, and ticketed attendees oozed through metal detectors run by members of the Secret Service's uniformed division. Similar to airport security screenings, rallygoers emptied their pockets to prove they weren't carrying guns or other weapons.

Media reports indicate the Secret Service had in place, as usual, a counter-sniper team scanning the surrounding area for threats.

In an exclusive interview, former Secret Service Director Julia Pierson told USA TODAY that maintaining such a sniper security perimeter is part of the agency's responsibility for safeguarding protectees like Trump from harm. She said agents typically consider 1,000 yards to be the minimum safe distance for sniper attacks.

The Secret Service has confirmed that it is investigating how Crooks got so close to Trump, who took the stage shortly after 6 p.m. Officials say Crooks' rifle was legally obtained but have not yet released specifics.

Outside the venue at that time, Greg Smith says he tried desperately to get the attention of police. He told the BBC that he and his friends saw a man crawling along a roof overlooking the rally. Other witnesses said they also saw a man atop the American Glass Research building outside the official event security perimeter, well within the range of a 5.56 rifle bullet.

"We noticed the guy bear-crawling up the roof of the building beside us, 50 feet away from us," Smith told the BBC. "He had a rifle, we could clearly see him with a rifle."

Smith told the BBC that the Secret Service eventually saw him and his friends pointing at the man on the roof.

"I'm thinking to myself, why is Trump still speaking, why have they not pulled him off the stage?" Smith said. "Next thing you know, five shots rang out."

From his nearby deck, Trump supporter Pat English watched as the former president took the stage to Lee Greenwood's "God Bless the U.S.A.," and attendees raised their cell phones to record.

English had taken his grandson to see the rally earlier but left when it got too hot. From his deck, they listened as Trump began speaking at 6:05 p.m., backed by a crown of red-hatted MAGA supporters waving "fire Joe Biden" signs.

And then gunfire began.

Boom, boom, boom

"I heard a 'boom, boom, boom' and then screams,” English said Sunday. "I could see people running and the police run in."

Trump was saying the word "happened" as the first pop rang out. He reached up to grab his ear as two more shots echoed, and the crowd behind him – and Trump himself – ducked. Plainclothes Secret Service agents piled atop the president as a fusillade of shots rang out, apparently the Secret Service killing Crooks.

The crowd screamed, and the venue's sound system picked up the agents atop Trump planning to move the former president to safety. One yelled, "shooter's down. Let's move, let's move."

The agents then helped Trump back to his feet as they shielded him on all sides.

The sound system then picked up Trump's voice: "Wait, wait," he said, before turning to the audience and triumphantly raising his fist to yell "fight, fight" as the crowd cheered, blood streaming down his face.

By 6:14 p.m. Trump's motorcade was racing from the scene, and in a later statement, Trump's campaign said he was checked out at a local medical facility.

"I was shot with a bullet that pierced the upper part of my right ear," Trump said in a statement. "I knew immediately that something was wrong in that I heard a whizzing sound, shots, and immediately felt the bullet ripping through the skin. Much bleeding took place, so I realized then what was happening."

Firefighter 'hero' gunned down

Outside of the Butler Township Administration Office Sunday afternoon, Pennsylvania Gov. Josh Shapiro identified the rally attendee killed by Crooks as Corey Comperatore, a firefighter, father of two and longtime Trump supporter.

“Corey died a hero,” Shapiro said. “Corey dove on his family to protect them last night at this rally. Corey was the very best of us. May his memory be a blessing.”

Two other Pennsylvanians are still undergoing treatment for their injuries, Shapiro said.

Pennsylvania State Police identified two wounded attendees David Dutch, 57, of New Kensington, and James Copenhaver, 74, of Moon Township. Both are hospitalized and listed in stable condition. Shapiro said he spoke with the family of one victim and received a message from the other.

Biden spoke briefly with Trump on Saturday night, and the president condemned the assassination attempt as “sick.” He said there’s no place for political violence in the U.S. and called on Americans to unite together to condemn it.

But earlier in the week, Biden told campaign donors in a private phone call it was time to stop talking about his own disastrous presidential debate performance and start targeting Trump instead.

"I have one job and that's to beat Donald Trump," Biden said. "We're done talking about the (June 27) debate. It's time to put Trump in the bullseye."

Republicans across the country have used similar language to attack their opponents over the years, and political scientists say violent rhetoric used worldwide almost invariably leads to physical violence.

On Sunday, someone parked a truck-mounted electronic billboard at the gates to the Butler Farm Show grounds reading "Democrats attempted assassination," along with a picture of Trump clutching an American flag, his face overlaid with a bullseye crosshairs.

Authorities say they have not yet determined a motive for Crooks' attack. But in a statement, Trump declared the shooting an act of evil and thanked God for preventing the unthinkable.

"We will fear not, but instead remain resilient in our faith and defiant in the face of wickedness," Trump said.

And he said he'd be back on the campaign trail for the Republican National Convention in Milwaukee, which starts Monday.

"Based on yesterday’s terrible events, I was going to delay my trip to Wisconsin, and the Republican National Convention, by two days," Trump said on his Truth Social account Sunday, "but have just decided that I cannot allow a 'shooter,' or potential assassin, to force change to scheduling, or anything else."

Contributing: David Jackson, Aysha Bagchi, Christopher Cann, Bryce Buyakie, Emily Le Coz, Josh Meyer, USA TODAY Network

How the assassination attempt unfolded : Graphics, maps, audio analysis show what happened