Music making and wellbeing: A systematic review of methods, musical activities, and outcomes

Abstract

Research on music making and wellbeing reports positive effects across psychological, physiological, and social dimensions. A closer look, nevertheless, reveals that most of these studies solely focus on group singing and define wellbeing in various and occasionally limited ways. This systematic review aimed at identifying the different research designs of quantitative studies in the field, structuring the musical activities, and summarizing the wellbeing outcomes. A comprehensive literature search in five databases resulted in a total of 50 papers that were included in the current review. Results showed that the majority of studies used quasi‐experimental designs with active or passive control conditions. Group and individual singing were most frequently studied. As for psychological wellbeing, studies with positive or no evidence are nearly balanced, with the exception of choral studies in younger and middle‐aged adults that predominantly provided positive evidence. Physiological wellbeing was measured mostly in singing activities with mixed results. Social wellbeing was primarily reported for group singing activities, with predominantly positive evidence. Taken together, this review underlines the need for including a wider range of music making activities in research on wellbeing, a careful selection of the respective research designs, and a balanced use of wellbeing measures.

INTRODUCTION

Positive effects of music on individual's wellbeing and stress reduction have for long attracted the interest of researchers across various contexts such as music therapy, daily music listening, and music education (MacDonald, 2013). A major focus has been on the relationship between active music making and wellbeing, particularly at the amateur level, and there are numerous research studies reporting benefits of this relationship (Clift, Hancox, et al., 2010; Hallam & Creech, 2016; Johnson et al., 2017; Kreutz et al., 2004; Linnemann et al., 2017; Matlschweiger & Sattmann, 2016; Pentikäinen et al., 2023; Schladt et al., 2017; Williamson & Bonshor, 2019). Whereas the majority of studies focused on only one form of music making, namely, singing rather than instrument playing, the variety of research methods employed is vast, and, correspondingly, definitions and measures of wellbeing are manifold. This systematic review aims at providing an overview of the current state of research on music making and wellbeing by examining the different music making activities, the research designs used, and the different wellbeing outcomes in a non‐clinical and non‐therapeutic context with healthy adults.

The concept of wellbeing

The term “wellbeing” has been defined in various ways, and there exists a plethora of different approaches and models (e.g., Diener et al., 2017; Lawes & Eid, 2025). This systematic review proposes a multidimensional view of wellbeing with three main components: psychological, social, and physiological. These dimensions can be compared to the dimensions of the “biopsychosocial” model of health (Engel, 1977, p. 132) and are also expressed in the World Health Organization's definition of health as “[…] a state of complete physical, mental, and social well‐being and not merely the absence of disease and infirmity” (World Health Organization, 1948, p. 1). Thus, this systematic review follows a holistic approach by integrating various components of wellbeing.

In this context, psychological wellbeing refers to two often mentioned concepts of wellbeing: hedonic and eudaimonic wellbeing (Ryff et al., 2021). Hedonic wellbeing focuses on the pursuit of pleasure and happiness, whereas eudaimonic wellbeing refers to the fulfillment of human potential and personal growth (Ryan & Deci, 2001; Ryff et al., 2021). The idea of hedonic wellbeing is reflected in the concept of subjective wellbeing, which is based on three core components: positive affect, negative affect, and life satisfaction (Diener et al., 1999; Ryff et al., 2021). The idea of eudaimonic wellbeing is, among others, integrated in the PERMA model by Seligman (2011), one of the founders of positive psychology. The model combines five elements, of which one belongs to the hedonic perspective: Positive emotions, the hedonic element, refer to feelings of contentment and happiness (Goodman et al., 2018). Engagement can be compared to a state of flow that occurs when someone is completely absorbed in a task or experiences states of intense concentration and focus (Butler & Kern, 2016; Goodman et al., 2018). Relationships include the quality and quantity of social connections and refer to feelings of being loved and valued by a caring and supportive social network (Ascenso et al., 2018; Goodman et al., 2018). Meaning can be defined by a sense of direction and purpose, resulting in a feeling that one's life is worthwhile (Goodman et al., 2018). Finally, accomplishment refers to a sense of achievement, success, and mastery in life (Goodman et al., 2018). Taken together, psychological wellbeing represents a rather internal and personal conceptualization of wellbeing (Elliott et al., 2022; Keyes, 2002).

Social wellbeing, in comparison, refers to an external and social conceptualization of wellbeing, with emphasis on the social environment (Elliott et al., 2022; Keyes, 2002). Although the PERMA model incorporates relationships and thus underlines the importance of social interactions for wellbeing, this systematic review treats social wellbeing as a separate dimension of wellbeing. In this context, social wellbeing refers to the interaction and communication with the social environment, the quality of social relationships, and thus the capability to engage and flourish in social contexts (e.g., Elliott et al., 2022; Shapiro & Keyes, 2008). As music making is often realized in social groups, it seemed convincing to focus on the social dimension separately.

Finally, physiological wellbeing refers to bodily processes. In contrast to the two aforementioned dimensions of wellbeing that mostly rely on subjective measurements, physiological wellbeing includes objective measurements such as hormones (e.g., cortisol, oxytocin), enzymes (e.g., alpha‐amylase), or peripheral measures. For example, the stress markers cortisol and alpha‐amylase can be associated with the stress response (Engert et al., 2011; Takai et al., 2004; van Stegeren et al., 2008); hence, reduced levels are related to wellbeing. Furthermore, oxytocin is often related to wellbeing through its effects on calm, social interactions, trust, and reduced fear (IsHak et al., 2011). As research has provided evidence for a relation between bodily functions and subjective wellbeing (e.g., Howell et al., 2007; Kushlev et al., 2020), it is necessary to integrate this dimension into the proposed concept of wellbeing.

Taken together, this systematic review considers wellbeing to be a multidimensional construct (e.g., Dodge et al., 2012; Lawes & Eid, 2025) that takes into account internal and external as well as subjective and objective states of wellbeing. All three dimensions of wellbeing are referred to in current research on music making and wellbeing (e.g., Clift et al., 2008; Gick, 2011; Shoda et al., 2025; Williamson & Bonshor, 2019).

Research on music making and wellbeing

A recent random‐effects meta‐analysis (Cooper, 2025) investigated the mean effects of music participation on social and emotional measures of wellbeing across the lifespan. The author included 25 studies with a between‐subject design and seven potential moderators that might influence the effects of music participation such as age, active/passive control group, intervention duration, and musical activity. Analysis showed small to medium effects (g ¹  = .23) of music participation on wellbeing and no statistical influence of the investigated moderators. In other words, music participation increased social and emotional wellbeing compared to active or passive control groups independent of the moderator variables. While this meta‐analysis provides an overview of the effects of music making on emotional and social wellbeing, physiological effects were not considered. The current review aims at both differentiating between three wellbeing dimensions and providing an overview of different methods in quantitative research on music making and wellbeing, without being restricted to only one research design (i.e., between‐subject). This permits the inclusion of a broader range of relevant studies, reflecting the diversity of research designs within the field and thus providing a more comprehensive picture of the current state of research. Moreover, it is explicitly intended to differentiate between the specific musical activities in order to identify potential research gaps.

Singing and wellbeing

Singing has frequently been investigated in music making (e.g., Bullack et al., 2018; Good & Russo, 2022; Kreutz et al., 2004; Maury & Rickard, 2018). Krause et al. (2018) reviewed literature on music participation and wellbeing to integrate perceived benefits into a measurement scale. Half of the reviewed studies addressed singing, a situation that has not changed in recent years.

Beck et al. (2000) were among the first to measure the effects of professional singing on the physiological markers cortisol and secretory immunoglobulin A (S‐IgA) as indicators of immune function. S‐IgA increased during rehearsals and a concert condition, whereas cortisol only decreased during rehearsals. Following up on this study, Kreutz et al. (2004) investigated the effects of amateur choir singing on S‐IgA, cortisol, and emotional state compared to a listening condition. Singing, in comparison to listening, increased positive mood and S‐IgA and decreased negative mood. No changes in cortisol were found. Another study by Good and Russo (2022) used a within design to investigate the effects of group singing compared to individual singing on self‐reported mood, cortisol, and oxytocin. Cortisol decreased in both singing conditions, whereas increases in mood and oxytocin could only be found in the group singing condition, which the authors attributed to social factors.

Further studies reported positive effects of singing using nonphysiological methods. Herschel et al. (2022) compared an online choir condition with an online mindfulness program in relation to wellbeing and other cognitive variables. Wellbeing increased in both groups, whereas trait anxiety decreased. The decrease in state anxiety was especially strong in the choir condition. In a similar way, Maury and Rickard (2018) examined a singing group, an exercise group, and a discussion group in terms of mood, energy levels, and group cohesion before and after a regular session. All groups showed an increase in the dependent variables while there were no between‐group differences.

Several studies, on the other hand, resulted in less clear effects of singing on wellbeing. Davidson et al. (2014), for example, evaluated the benefits of a singing program for older adults on health and wellbeing. Analyses revealed no significant changes between pre‐ and post‐scores in social and psychological wellbeing dimensions. In a questionnaire study, Johnson et al. (2017) compared the wellbeing of older amateur choir singers with the wellbeing of a matched control group. Choristers scored significantly higher on physical but not psychological wellbeing compared to the control group.

Instrument playing and wellbeing

Research on musical activities other than singing (i.e., instrument playing) is scarce. A qualitative study by Williamson and Bonshor (2019) investigated different dimensions of wellbeing in brass bands. Results showed areas of overlap between reported wellbeing in brass band players and choristers, especially in physical, psychological, and social wellbeing, mainly due to enhanced respiratory function, body posture, and regular social interactions common to both musical activities. In contrast, Matlschweiger and Sattmann (2016) found that brass band playing differed from choral singing, theater playing, and listening to music in a concert. All activities showed positive influences on wellbeing in a pre–post rehearsal comparison, except for brass banding. In a similar vein, Toyoshima et al. (2011) investigated the stress‐reducing effects of 30‐min piano playing, clay molding, and calligraphy. Anxiety scores and cortisol decreased in all three conditions, with a significantly higher decrease of the latter in the piano condition. Thus, there is some evidence for a relation between instrument playing and wellbeing, yet results are less conclusive, and more research is needed.

Instrument playing seems to play a more important role in research on cognitive functioning and cognitive health. For instance, learning to play the piano resulted in, among others, improved attention, executive functions, verbal fluency, and processing speed in older adults (Bugos & Kochar, 2017; Seinfeld et al., 2013). Further positive effects have been found for memory capacity, especially the recall of verbal and visual information in musically trained individuals (Jakobson et al., 2008; Kilgour et al., 2000).

Another area of study concentrates on music making and “ill‐being” with a focus on, for example, Music Performance Anxiety (e.g., Kenny et al., 2014; Spahn et al., 2021) and physical pain resulting from music making (e.g., Rousseau et al., 2021; Silva et al., 2015). This research focuses on the improvement of misery (Seligman, 2010). In contrast, since the postulation of positive psychology (Seligman, 2010), researchers and practitioners increasingly focus on the improvement of wellbeing or “what made life worth living” (Seligman, 2010, p. 232), an approach that is also followed in the present systematic review.

The current systematic review

Previous research on active music making and wellbeing includes different musical activities, but mostly investigated singing. Several research designs were used that differ in their experimental approaches and the composition of the control conditions. In terms of concepts and definitions, wellbeing has been operationalized and measured across psychological, physiological, and social dimensions, yet these dimensions are typically investigated separately and findings partially contradict each other. To the best of our knowledge, there has been no systematic review that considers these three dimensions in one paper.

The purpose of this study was to provide an overview of the current state of quantitative empirical research on the relationship between active music making and different dimensions of wellbeing. A particular focus is on the research methods used, the musical activities investigated, and the evidence in the results. There are three main research questions:

1. Which quantitative research designs have been used for investigating the relationship between music making and wellbeing?

2. Which types of musical activities were included in the studies?

3. What evidence of music making on different dimensions of wellbeing has been reported?

METHODS

This Systematic Review was conducted according to the Preferred Reporting Items for Systematic reviews and Meta‐Analyses (PRISMA) statement 2020 including the PRISMA 2020 Checklist (Page et al., 2021).

Eligibility criteria

Several inclusion criteria were defined. A given paper of interest needed to report some type of active music making (i.e., singing or playing an instrument) of healthy adult participants. The reported outcomes should include at least one physiological and/or psychological measure of wellbeing and/or stress/anxiety. In order to assess the benefits of music making, it was necessary that the experimental design included a control group or was based on a pre–post measure. As one focus of this systematic review is on the variety of research designs, it was decided to focus on quantitative studies only and to exclude qualitative studies. Only peer‐reviewed papers were included.

The exclusion criteria comprised music therapy (especially for severe illnesses such as Parkinson's disease, dementia, or cancer), Music Performance Anxiety, dyadic relationships (i.e., mother–infant), qualitative research, and reviews. Since there were only very few studies of children and adolescents, which occasionally did not meet the abovementioned design selection criteria as revealed by the subsequent text screening, we decided to exclude these studies.

Information sources and search strategy

A comprehensive literature search was carried out in September 2024 using the following databases: Web of Science, PsycINFO, PubMed, RILM, and SAGE Journals. For the database PsycINFO, the two filters peer‐reviewed and humans were applied. The filter peer‐reviewed was applied to the database RILM. In addition, a hand search was conducted to retrieve further suitable articles. This additional hand search lasted until February 2025. As we aimed to identify papers that investigated the benefits of active music making on different forms of wellbeing and stress reduction, we used the following search term: (“making music” OR “playing music” OR singing OR “music* engagement” OR “music* activity” OR “music* performance”) AND (stress OR wellbeing OR “well‐being” OR cortisol OR amylase OR “heart rate”) NOT (therap* OR MPA OR “Music Performance Anxiety”). This search term was applied to the papers' title, abstract, and, if available, keywords.

Selection of papers

After the identification of duplicates (i.e., identical records identified in different databases), the remaining studies were imported into Rayyan, a web application for the initial screening of abstracts and titles (Ouzzani et al., 2016). No automation processes of the application were used, and all papers were screened by the authors. Abstracts and titles were prescreened in two stages by the first author (C.S.) and classified into the three categories Included, Maybe, and Excluded. While the first stage predominantly served to exclude papers with no relation to music or the outcomes of interest, the second stage focused on the exclusion criteria in more detail. The full‐text screening (stage three), containing the remaining less than 100 Included and Maybe papers, was performed independently by the first and second authors (C.S. and C.W.). Emerging discrepancies were resolved by discussion between the two authors. If papers were not available, authors were contacted and asked for a copy. In case of no reply, papers were excluded from the systematic review. Figure 1 presents the complete identification and screening process.

FIGURE 1.

PRISMA flow diagram of the search, screening, and inclusion process. Note: Some papers comprised more than one study.

Variables and data

After the full‐text screening, data of interest were extracted from the studies. In order to compare the selected studies according to our research questions, we categorized the data as follows: (1) study, (2) objectives, (3) participants, (4) musical activity, (5) conditions, (6) intervention and research design, (7) wellbeing measures, and (8) outcome (see Table S1). The different experimental designs are summarized in Table 1.

TABLE 1.

Summary of research designs and quality appraisal of the studies. Numbers in square brackets are used for the identification of studies in this review.

Research designs and authors	N	Conditions	MMAT score
Experimental design
1. Short term
1.1 Randomized controlled trial (RCT)
[1] Bullack et al. (2018, Exp. 1 + 2)	54, 49	Choral singing vs. non‐singing (listening)	3
[2] Sanal and Gorsev (2014)	70	Choral singing vs. unstructured free time	2
[3] Unwin et al. (2002)	81	Amateur group singing vs. listening to singing	3
1.2 Quasi‐experimental within design
[4] Beck et al. (2000)	41	Choir rehearsal vs. performance (professionals)	4
[5] Boyd et al. (2020)	59	Group singing vs. listening	4
[6] Fancourt et al. (2015)	15	Choir rehearsal vs. performance (professionals)	5
[7] Good and Russo (2022)	9	Group singing vs. solo singing	3
[8] Keeler et al. (2015)	4	Standard performance vs. improvised performance	3
[9] Kreutz et al. (2004)	31	Choral singing vs. listening	4
[10] Kreutz (2014)	21	Singing vs. chatting	4
[11] Schladt et al. (2017)	38	Choral singing vs. solo singing	5
[12] Stone et al. (2018)	9	Singing vs. dancing vs. reading vs. cycling	4
1.3 Quasi‐experimental mixed design
[13] Bowling et al. (2022)	71	Singing together vs. singing alone vs. speaking together vs. speaking alone	5
[14] Grape et al. (2002)	16	Amateur singers vs. professional singers, singing lesson	3
[15] Matlschweiger and Sattmann (2016)	183	Choir vs. theater vs. brass band vs. concert‐goers	5
[16a] Maury & Rickard (2018, Phase 1)	79	Choir vs. exercise vs. discussion	4
[17] Shoda et al. (2025)	36	Choir vs. go playing	4
[18] Toyoshima et al. (2011)	57	Piano playing vs. clay molding vs. calligraphy vs. silence	4
[19] Valentine and Evans (2001)	33	Choral singing vs. solo singing vs. swimming	4
2. Long‐term
2.1 RCT and crossover
[20] Coulton et al. (2015)	258	Group singing older adults vs. active WLG	4
[21] Galinha et al. (2021)	149	Group singing older adults vs. active WLG	5
[22] Galinha et al. (2022)	149	Group singing older adults vs. active WLG	5
[23] Herschel et al. (2022)	61	Online choir training vs. online mindfulness training (COVID)	4
[24] Ploukou and Panagopoulou (2018)	48	Percussion music playing vs. no intervention	2
[25] Schäfer (2023)	74	Online group singing vs. online discussion vs. WLG (COVID)	2
2.2 Quasi‐experimental mixed design
[26] Creech et al. (2013)	500	Music groups vs. other kinds of non‐music activities	3
[27] Hallam and Creech (2016)	500	Music groups vs. other kinds of non‐music activities	3
[28] Hallam et al. (2014)	500	Music groups vs. other kinds of non‐music activities	3
[29] Maury and Rickard (2022)	65	Choir vs. exercise	4
[30] Pearce et al. (2015)	135	Singing vs. creative craft vs. creative writing	5
[31] Pearce et al. (2016)	135	Singing vs. creative craft vs. creative writing	5
[32] Pentikäinen et al. (2023)	169	Choral singers vs. non‐singers	5
[33] Perkins & Williamon (2014, Study 1)	98	Instrumental lessons vs. non‐music learning lessons	5
[34] Seinfeld et al. (2013)	41	Piano learning group vs. other leisure activities	4
2.3 Quasi‐experimental one‐group pre–post
[35] Corvo et al. (2020)	41	Group singing older adults	5
[36] Davidson et al. (2014)	29	Group singing older adults	4
[37] Linnemann et al. (2017)	44	Choral singing (practice days + performances)	5
[38] Moss and O'Donoghue (2020)	54	Health service workplace choirs	4
[39] Taets et al. (2021)	25	Choral singing	4
[40] Viola et al. (2024)	23	Choral singing	4
[41a] Zhang (2022, Phase 2)	30	Group vocal training for non‐singers	4
Non‐experimental design
1. Cross‐sectional
[42] Bailey (2023)	543	Music ensemble participants vs. non‐music ensemble participants	5
[43] Caetano et al. (2019)	73	Choral singers vs. matched general population	4
[44] Johnson et al. (2017)	416	Older choir singers vs. matched general population	5
[45] Lamont and Ranaweera (2020)	957	Amateur musicians vs. amateur knitters	5
[46] Lonsdale and Day (2021)	194	Choral singing vs. other music and sport activities	5
[47a] Maury et al. (2022, Study 1)	190	Choir vs. exercise vs. other sedentary activities	5
[48] Pentikäinen et al. (2021)	162	Low activity choir singers vs. high activity choir singers vs. control group	5
[49] Pérez‐Aldeguer and Leganés (2014)	496	Choristers vs. non‐choristers	5
[50] Stewart and Lonsdale (2016)	375	Choral singers vs. solo singers vs. team sport players	5
[41b] Zhang (2022, Phase 1)	180	Singers vs. non‐singers	4
2. Other (ESM, observational)
[16b] Maury & Rickard (2018, Phase 2)	53	Choir vs. exercise	5
[47b] Maury et al. (2022, Study 2)	59	Mood and activities of a mixed group of participants	4

Note: N = total number of participants. Studies 21–22, 26–28, and 30–31 each used the same data set.

Abbreviations: ESM, Experience Sampling Method; MMAT, Mixed Methods Appraisal Tool; WLG, waiting‐list group.

Risk of bias assessment

In order to assess the quality of the included studies, we used the Mixed Methods Appraisal Tool (MMAT), version 2018 (Hong et al., 2018). The MMAT has been designed to evaluate the quality of empirical studies with different study designs. It contains five categories: (1) qualitative studies, (2) quantitative randomized controlled trials, (3) quantitative non‐randomized studies, (4) quantitative descriptive studies, and (5) mixed methods studies. Categories two to four were relevant for this systematic review. All categories are evaluated using five design specific questions with the following response options: Yes, No, and Can't tell (Hong et al., 2018). Thus, each study receives a score from 1 to 5, with 5 being the highest rating possible indicating low risk of bias and 1 being the lowest rating possible indicating high risk of bias. The MMAT was used as it facilitates the quality assessment of various study designs and enables comparisons between different designs while being both efficient and reliable (Pace et al., 2012).

RESULTS

A total of 2851 records were identified in the literature search process. After removing 1190 duplicates, the titles and abstracts of the remaining 1661 papers were screened, which yielded in the exclusion of another 1594 papers. Main reasons for exclusion were a missing relation to music and/or wellbeing, research topics primarily related to physical problems in music making, music listening, music interventions with people suffering from severe illness, and articles with a historical focus. Thus, 67 articles remained in the data set for full‐text screening, and the exclusion reasons at this stage are presented in Figure 1. In this context, no adequate design and no adequate outcome refer to studies that, for example, mixed artistic activities (e.g., singing and art exhibitions visits) or focused only on health effects (e.g., perceived health) without reflecting on participants' wellbeing. As a result, a total of 50 papers with 54 studies were included in the systematic review. The number of participants investigated in the studies ranged from a minimum of four to a maximum of 957. Studies were published between 2000 and 2024.

Regarding the quality appraisal of the included studies, 21 studies reached an assessment of 5/5, 21 studies scored 4/5, eight studies scored 3/5, and three studies received 2/5. Table 1 summarizes the overall score for each study. The detailed risk of bias assessment with the specific quality appraisal questions can be found in Table S2. Limitations of the studies are addressed in  Discussion section.

Research designs

Table 1 provides an overview of the different methodological approaches applied in the studies. A rough distinction can be made between experimental and non‐experimental designs. The experimental designs were further categorized into short‐term and long‐term (longitudinal) studies. In the current paper, interventions with one session for each condition were defined as short term. Long‐term studies had a minimum duration of 4 weeks. Quasi‐experimental studies (e.g., convenience sampling) constituted a subcategory of experimental studies and did therefore not represent a separate category.

Studies with short‐term designs comprised three randomized controlled trials (RCTs; studies 1–3) that all compared a singing condition to non‐music making control conditions (i.e., listening to music, unstructured free time). A follow‐up measure after 1week was only implemented by Unwin et al. (2002). The vast majority of studies included in this review employed a quasi‐experimental design meaning that allocation to the experimental or control group depended on preexisting group participation, such as music, theater, or sport. Nine quasi‐experimental studies employed a within‐participant design (studies 4–12) in which the study sample participated in all conditions. Here, only two studies compared singing with other active non‐music conditions (Kreutz, 2014; Stone et al., 2018). Two studies further compared choral singing with a listening condition (Boyd et al., 2020; Kreutz et al., 2004), whereas the majority of studies compared different music making settings (i.e., rehearsal vs. performance, group singing vs. solo singing, standard performance vs. improvisation). Five out of seven studies employed a mixed design (i.e., group as a between factor and time as a within factor; studies 13–19) and compared preexisting music groups with other preexisting non‐music control groups (e.g., swimming, clay molding, go playing). Grape et al. (2002) focused on two types of musicians and compared amateur singers with professional singers during a singing lesson, whereas Bowling et al. (2022) compared group and solo singing to group and solo speaking.

The six experimental long‐term RCTs (studies 20–25) included two studies with a 6‐month follow‐up (Coulton et al., 2015; Galinha et al., 2022). Galinha et al. (2021) implemented a crossover design with an intervention group and a waiting‐list group that switched roles after a 2‐month washout period. Two studies compared a randomized online singing condition with an online active control group (Herschel et al., 2022; Schäfer, 2023). Only one study compared a music group to a no‐intervention control group (Ploukou & Panagopoulou, 2018). Nine quasi‐experimental mixed design studies (studies 26–34) used intervention periods from 10 weeks up to 2 years. The music groups were all compared to other non‐music groups. Seven long‐term one‐group pre–post design studies (studies 35–41a) investigated different choir groups over periods of up to 3 months. Only the study by Corvo et al. (2020) implemented a 3‐month follow‐up measurement.

The non‐experimental designs (studies 41b–50) predominantly included cross‐sectional studies that investigated musicians and non‐musicians at one specific point in time. Six out of 10 studies compared musicians to non‐musicians (Caetano et al., 2019; Johnson et al., 2017; Pentikäinen et al., 2021; Pérez‐Aldeguer & Leganés, 2014; Zhang, 2022, Phase 1). The remaining four studies compared amateur musicians to participants in other physical and more sedentary activities such as exercise or knitting. An Experience Sampling Method was used by Maury et al. (2022, Study 2; study 47b) to collect data about daily activities and daily mood of participants engaged in musical and non‐musical activities. Another study by Maury and Rickard (2018, Phase 2; study 16b) used an observational methodology for the two groups of interest (choir and exercise).

Taken together, most research was conducted using a quasi‐experimental between‐ or within‐participant design that compared musical activities to other active or, to a lesser extent, passive control conditions.

Musical activities

An overview of the different musical activities investigated in the studies in relation to wellbeing is presented in Figure 2. The tree diagram illustrates the higher‐level musical activities (color coding) that branch out into more specific musical activities. Different circle sizes indicate the number of studies per category, with bigger circles representing more studies.

FIGURE 2.

Tree diagram of musical activities. Numbers in brackets indicate the respective studies listed in Table 1. Note: Studies 21–22, 26–28, and 30–31 each used the same data set.

The most prominent musical activity investigated is singing. A total of 47 singing activities were identified in the respective studies, compared to six instrument playing activities and six conditions that included diverse musical activities. It should be mentioned that three of the instrument playing activities were part of a variety of music conditions that also included singing and were reported separately in the tree diagram (Lonsdale & Day, 2021; Matlschweiger & Sattmann, 2016). The singing activities were further subdivided into group singing (40 conditions) and individual singing, including solo singing (five conditions) and singing lessons (two conditions). With the exception of one study (Grape et al., 2002), individual singing was never investigated in isolation but was always compared with group singing. The third level of the tree diagram shows the specific group singing activities. Choral singing was the topic of 24 studies, of which four studies included professionals and music students (Beck et al., 2000; Fancourt et al., 2015; Keeler et al., 2015; Sanal & Gorsev, 2014). A relatively large sample of 14 studies focused on choral singing with older people (mostly aged 60+), including specific singing interventions for the elderly (i.e., Sing4Health, Silver Song Club). Two studies investigated online singing groups during the COVID‐19 pandemic (Herschel et al., 2022; Schäfer, 2023).

Instrument playing has been subdivided into two different branches. The ensemble playing category includes four studies: A percussion workshop with oncology nurses (Ploukou & Panagopoulou, 2018), piano training with older adults (Seinfeld et al., 2013), a brass band rehearsal (Matlschweiger & Sattmann, 2016), and people “playing an instrument as part of a band/orchestra” (Lonsdale & Day, 2021, p. 1183) without any further specification. Two studies fell into the individual playing category: Toyoshima et al. (2011) studied amateur piano players, and Lonsdale and Day (2021) also included solo instrumentalists.

The third higher‐level musical activity is a combination of singing and instrument playing activities. The six studies mixed diverse musical activities (e.g., choir, orchestra, guitar, recorder, ukulele) and reported the final benefits of music making without distinguishing between the different musical activities. As well as for choral singing, the diverse musical activities also include a section focusing on older people (Creech et al., 2013; Hallam et al., 2014; Hallam & Creech, 2016; Perkins & Williamon, 2014). A younger sample was targeted in the studies by Bailey (2023) and Lamont and Ranaweera (2020).

Overall, these findings illustrate that singing is by far the most frequently investigated musical activity and that a high number of studies focused on music making in older people.

Wellbeing outcomes

The following section is divided into three parts. First, the assessment of the different wellbeing dimensions is outlined. Second, the three evidence categories are introduced and explained. Finally, the reported outcomes for the different musical activities on different dimensions of wellbeing are summarized according to the evidence categories.

Overview of methods

Tables 2 and 3 report the number of studies yielding different evidence on psychological, physiological, and social wellbeing. The specific studies included in each evidence category can be found in Table S3 and Table S4. Results are separated for experimental (quasi‐experimental and RCT) and non‐experimental studies. Psychological wellbeing was mainly measured using questionnaires on positive and negative affect, mood, quality of life, satisfaction with life, stress, anxiety, depression, and mental health (e.g., Positive and Negative Affect Schedule; Watson et al., 1988; Perceived Stress Scale; Cohen et al., 1983; State–Trait Anxiety Inventory; Spielberger et al., 1983; Satisfaction with Life Scale; Diener et al., 1985). Physiological wellbeing was predominantly assessed with measures of hormones (e.g., cortisol, oxytocin), proteins (e.g., immunoglobulin A), enzymes (e.g., alpha‐amylase), and other peripheral measures (e.g., heart rate, blood pressure). Studies with an emphasis on social wellbeing investigated the benefits of music making by focusing on social bonding, social connectedness, social cohesion, and loneliness. Commonly used scales to measure social wellbeing included the Inclusion of Other in Self Scale (IOS; Aron et al., 1992), the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet et al., 1988), and the UCLA Loneliness Scale (Hughes et al., 2004).

TABLE 2.

Total number of experimental studies (n = 42) yielding positive (+), mixed (+/o), or no (o) evidence for wellbeing measures. Read the table as follows: The first number refers to the total of experimental studies per category; the second number refers to the subset of RCT studies. For example, 21/3 indicates a total of 21 studies that include three RCT studies.

Musical activity	Psychological wellbeing			Physiological wellbeing			Social wellbeing
	+	+/o	o	+	+/o	o	+	+/o	o
Choral (n = 21/3)	16/3	2/0	2/0	5/0	5/0	3/2	3/1	−	2/0
Choral older adults (n = 10/3)	3/1	3/1	2/0	1/0	2/1	−	2/1	1/1	3/0
Choral online (n = 2/2)	2/2	−	−	−	−	−	1/1	−	−
Individual singing (n = 5/0)	3/0	1/0	1/0	2/0	2/0	1/0	−	−	−
Ensemble playing (n = 3/1)	2/1	−	1/0	−	−	−	−	−	1/0
Individual playing (n = 1/0)	1/0	−	−	1/0	−	−	−	−	−
Diverse activities (n = 0)	−	−	−	−	−	−	−	−	−
Diverse activities older adults (n = 4/0)	1/0	3/0	−	−	−	−	−	−	−

Note: Single studies can appear several times if various wellbeing measures and different musical activities had been investigated. Some studies used the same data set (see Table S3).

TABLE 3.

Total number of non‐experimental studies (n = 12) yielding positive (+), mixed (+/o), or no (o) evidence for wellbeing measures.

Musical activity	Psychological wellbeing			Physiological wellbeing			Social wellbeing
	+	+/o	o	+	+/o	o	+	+/o	o
Choral (n = 4)	1	1	2	−	−	−	−	−	−
Choral older adults (n = 6)	2	3	2	−	−	−	1	2	−
Choral online (n = 0)	−	−	−	−	−	−	−	−	−
Individual singing (n = 2)	−	−	2	−	−	−	−	−	−
Ensemble playing (n = 1)	−	−	1	−	−	−	−	−	−
Individual playing (n = 1)	−	−	1	−	−	−	−	−	−
Diverse activities (n = 2)	−	1	1	−	−	−	−	−	1
Diverse activities older adults (n = 0)	−	−	−	−	−	−	−	−	−

Note: Single studies can appear several times if various wellbeing measures and different musical activities had been investigated. Some studies used the same data set (see Table S4).

Types of evidence

These three outcomes were categorized in positive evidence (+), mixed evidence (+/o), and no evidence (o), based on the statistical significance of the quantitative results. Studies were assigned to the category positive evidence if participants in the music making group showed a significant improvement in the majority of the respective wellbeing variables between pre‐ and post‐measurement or if they scored higher in wellbeing than a non‐musical control group (e.g., cross‐sectional studies that compare musicians with non‐musicians). The mixed evidence category comprised all studies that showed mixed results on the respective wellbeing measures (i.e., positive and negative or positive and no effects). Studies in which music making did not affect wellbeing measures were labeled no evidence. As many studies of psychological musical wellbeing frequently measured changes in positive and negative affect (e.g., Bullack et al., 2018; Kreutz et al., 2004; Sanal & Gorsev, 2014), benefits regarding these two variables served as a decisive criterion for classification. With regard to physiological wellbeing measures, changes in cortisol (a frequently used marker for stress) were used as a key criterion for classification. No specific classification criterion was defined for social wellbeing measures.

Psychological wellbeing, as the most frequently assessed wellbeing measure, was investigated across all musical activities using experimental and non‐experimental designs. The measurement of physiological wellbeing was, with the exception of one study (Toyoshima et al., 2011), limited to experimental studies on singing. Social wellbeing was predominantly investigated with choirs using both experimental and non‐experimental designs.

Evidence in singing activities

Studies of choral singing investigated wellbeing outcomes in all three dimensions. While non‐experimental studies showed a relatively even distribution across positive, mixed, and no evidence categories, experimental studies, and in particular RCTs investigating psychological and social wellbeing, more often reported positive or partly positive evidence. Bowling et al. (2022), for example, compared four experimental conditions (i.e., singing together, speaking together, singing alone, speaking alone) and found that singing together resulted in the greatest benefits for positive affect and social connectedness compared with the other three conditions. Moreover, a significant main effect of time indicated that cortisol was significantly lower in the post‐measurement compared to the pre‐measurement across all conditions. Pearce et al. (2015) compared a singing group with non‐singing groups (i.e., crafts and creative writing) over 7 months and found overall increases in positive affect with a significantly greater increase for the singing condition. Closeness increased similarly in both groups; however, the singing group showed a greater increase after the first time point, indicating that singing might facilitate social bonding at the beginning. They called this phenomenon the “ice‐breaker effect” (Pearce et al., 2015, p. 1). Grape et al. (2002) reported mixed evidence in their comparison of professional and amateur singers before and after an individual singing lesson. They concluded that amateur singers experienced increases in wellbeing and enjoyment after singing lessons, whereas professional singers experienced more arousal.

As physiological wellbeing was nearly solely investigated in experimental studies on singing with a roughly equal distribution across all three evidence categories, these results will be outlined in more detail. Investigating different physiological markers, Bullack et al. (2018), for example, did not detect any differences in either alpha‐amylase or cortisol after a choir rehearsal. Sanal and Gorsev (2014) similarly found no changes in alpha‐amylase after a period of singing. Kreutz et al. (2004), in contrast, reported that choral singing leads to increases in salivary immunoglobulin A but did not find any differences in cortisol. Immunoglobulin A also increased in a study by Beck et al. (2000) who compared professional singers after a rehearsal and a concert condition. The measurement of oxytocin resulted in rather mixed results, ranging from increases to decreases, which might depend on specific study characteristics (e.g., group or individual singing, newly formed or preexisting group; Bowling et al., 2022; Kreutz, 2014; Schladt et al., 2017). Physical markers most frequently included heart rate and blood pressure (e.g., Valentine & Evans, 2001; Viola et al., 2024). Positive effects were observed for heart rate, whereas blood pressure showed no effects.

Evidence in instrument and diverse musical activities

Instrument playing activities and diverse musical activities focused on the measurement of psychological wellbeing. Positive evidence was found for percussion groups and group piano training (Ploukou & Panagopoulou, 2018; Seinfeld et al., 2013), individual piano playing (Toyoshima et al., 2011), and groups of diverse music learners (Perkins & Williamon, 2014). No evidence was found for orchestra/band musicians, solo musicians (Lonsdale & Day, 2021), a group of diverse amateur musicians (Lamont & Ranaweera, 2020), and brass bands (Matlschweiger & Sattmann, 2016), which the authors of the latter study attributed to dissatisfaction with the rehearsal session (Matlschweiger & Sattmann, 2016). Hence, both ensemble and individual playing conditions resulted in a diverse range of evidence. The only study including a physiological measure of wellbeing was conducted by Toyoshima et al. (2011), who compared piano playing to other creative arts. They found decreased levels of cortisol at post‐measurements, with a steeper decrease in the music than the control conditions. Two studies examined the benefits of instrumental and diverse forms of music making on social wellbeing (Lamont & Ranaweera, 2020; Seinfeld et al., 2013) but did not find positive evidence, irrespective of the experimental design used.

In summary, positive, mixed, and no evidence appear in all musical activities. RCTs resulted in positive and mixed evidence particularly for psychological and social wellbeing. Studies of choral singing investigated wellbeing outcomes in all three dimensions, with a considerably high number of positive evidence on psychological wellbeing for younger and middle‐aged adults.

DISCUSSION

Wellbeing has been defined as a multidimensional construct that includes psychological, physiological, and social dimensions. This systematic review aimed at providing an overview of the current research on music making and wellbeing, with a particular focus on the various research designs used, the different musical activities included, and the reported outcomes regarding the three dimensions of wellbeing. A systematic literature search resulted in 50 papers that were included in the current review. The majority of examined papers used a quasi‐experimental design with different control conditions. Group and individual singing were the most often studied musical activities. Regarding the outcomes in the psychological, physiological, and social dimensions, a broad range from no evidence to mixed and positive evidence was reported, with choral studies reporting the highest amount of at least partially positive evidence on all forms of wellbeing.

Research designs

To the best of our knowledge, this is the first systematic review focusing in detail on the research designs of quantitative studies on music making and wellbeing. Research designs were also mentioned in other thematically related reviews (e.g., Campbell et al., 2022; Clift, Nicol, et al., 2010; Sheppard & Broughton, 2020; Zhang et al., 2025); however, in these cases, the classification was kept rather broad (e.g., classification in observational and experimental studies, Campbell et al., 2022). The current systematic review, apart from a more detailed overview of the various designs, also considered the experimental and control conditions. According to the MMAT (Hong et al., 2018), the majority of the studies received high scores (79% with at least four out of five points), indicating low risk of bias across the several study designs.

Most of the included studies used quasi‐experimental designs with different control conditions or a one‐group pre–post measure. Thus, assignment to the experimental and the control condition was not randomized but depended on group membership (e.g., music group, sports group, theater group). Only nine studies used a RCT design. This underrepresentation of RCT designs is not surprising given the difficulty of randomization in an artistic and leisure activities context, where, depending on the musical activity in question, a certain level of expertise is required. Randomization is possible if no musical expertise is required, for example, in studies that investigate the learning of musical instruments, or musical activities that do not require any prior knowledge such as elementary singing or drumming (e.g., Herschel et al., 2022; Ploukou & Panagopoulou, 2018; Schäfer, 2023). Additionally, randomization within a music group might carry a certain bias, as frustration can arise when musicians are assigned to a non‐music control group that includes unstructured free time or passive listening to the rehearsed pieces (e.g., Bullack et al., 2018; Sanal & Gorsev, 2014). A recommendation by Dingle et al. (2019) about best practice research on group singing and wellbeing suggests that “randomization is preferred but where not feasible, researchers should include an appropriate control or comparison sample in their design” (p. 7). Indeed, most of the studies included in this systematic review compared musical activities with other non‐musical activities, such as theater playing, clay molding, discussion groups, or sport groups (e.g., Matlschweiger & Sattmann, 2016; Maury & Rickard, 2018; Toyoshima et al., 2011).

Regarding the intervention duration, studies were categorized into short‐term and long‐term (longitudinal) studies. These studies all used at least two measurement points (pre–post), with only six studies implementing a follow‐up measure from a week up to 2 years (Corvo et al., 2020; Coulton et al., 2015; Galinha et al., 2021; Galinha et al., 2022; Pentikäinen et al., 2023; Unwin et al., 2002). As two measurement points within a short duration of time do not permit conclusions about long‐term effects, follow‐up measures should be taken into account (cf. Dingle et al., 2019).

Musical activities

The musical activity most frequently studied in the 50 papers reviewed here was singing, and in particular, choral singing. This finding is in line with previous work, highlighting the strong research focus on choirs and singing as a musical activity (e.g., Krause et al., 2018; Williamson & Bonshor, 2019). There is also a strong focus on research dedicated to music making (mostly singing) in older adults, which emphasizes the importance of research on interventions helping to promote wellbeing in later life. This is especially important given the ongoing aging of highly developed societies (e.g., Seinfeld et al., 2013) and the emphasis on successful or active aging (e.g., Coulton et al., 2015; Davidson et al., 2014).

Outcomes on psychological, physiological, and social wellbeing

The quantitative outcomes summarized in this systematic review measured wellbeing across psychological, physiological, and social dimensions. Psychological wellbeing is the most often reported outcome across all musical activities. Outcome variables frequently addressed positive and negative affect, quality of life, and satisfaction with life, as well as reductions in anxiety, stress, and depression, employing a number of standardized measures and inventories. The evidence range is quite broad, with studies reporting outcomes from positive to mixed or no evidence of beneficial effects. The summarized evidence on psychological and social wellbeing is in line with the meta‐analysis by Cooper (2025), which reports small to medium‐sized effects of music participation on emotional and social measures of wellbeing. In the current review, choral studies most often reported positive or mixed evidence, whereas the distribution between positive and no evidence studies is nearly balanced for the remaining musical activities. In the case of no evidence studies (e.g., Davidson et al., 2014; Lamont & Ranaweera, 2020; Lonsdale & Day, 2021; Moss & O'Donoghue, 2020), possible explanations included sample characteristics, sample size, or the duration of the intervention. This demonstrates the complexity and significance of study organization as well as the importance of ongoing research on this topic.

It is important to consider that the number of variables taken into account differ considerably across studies. For example, Kreutz et al. (2004) focused on the emotional state by only using the PANAS, whereas Matlschweiger and Sattmann (2016) measured positive and negative affect, stress, and anxiety. Even if certain variables were prioritized, the higher the number of included variables, the higher the probability that studies were categorized into the mixed evidence category. Moreover, one important limitation to the studies is that a considerable number did not provide a theoretical framework or a working definition of the constructs of interest (e.g., wellbeing, quality of life), yet rather referred to previous work on music making and wellbeing. Defining the construct of interest is a crucial prerequisite for the selection of the dependent variables and strengthens clarity as well as comparability between studies (cf. Dingle et al., 2019). A second challenge refers to the limited focus on psychological wellbeing, neglecting physiological and social dimensions. Only two studies considered all three dimensions (i.e., Bowling et al., 2022; Bullack et al., 2018). Thus, most studies in the field of music making and wellbeing do not provide a holistic perspective on wellbeing but focus primarily on internal and personal processes. Although this dimension is undoubtedly important to consider, it should be complemented by additional relevant dimensions.

Physiological wellbeing was most commonly investigated in singing contexts. The objective measurements ranged from hormones, proteins, and enzymes to physical markers and thus cover a broad range of physiological markers. Results were mixed, with some markers showing increases (i.e., immunoglobulin A; e.g., Kreutz et al., 2004), some markers showing no changes (i.e., alpha amylase; e.g., Bullack et al., 2018), and other markers showing heterogeneous results (i.e., cortisol and oxytocin; e.g., Bowling et al., 2022; Kreutz, 2014; Schladt et al., 2017; Toyoshima et al., 2011). Objective measurements of physiological markers are important as they provide additional information about the effects of music making on the body and, consequently, on wellbeing, independent from subjective self‐reports. Potential inconsistencies in results could be related to lag times of biomarkers (Dingle et al., 2019), and collaborations between endocrinologists, immunologists, or biological scientists are recommended.

Social wellbeing was measured almost exclusively in choral studies. Results provided predominantly positive evidence, indicating that singing improved social variables such as cohesion and connectedness (e.g., Bowling et al., 2022; Maury & Rickard, 2018). As social connectedness and social relations in general contribute considerably to wellbeing (i.e., PERMA model; Seligman, 2011), they warrant greater investigation in other group music contexts such as orchestras, brass bands, or percussion groups, which are also shaped by social structures.

Limitations and future directions

First, we only focused on quantitative research. Although this decision was made with regard to our research questions and to permit a certain level of comparability, qualitative findings may offer insights into relevant underlying self‐concepts of music making and wellbeing. Second, due to the diversity of studies, research designs, and methods, we refrained from conducting a meta‐analysis and instead chose to provide an overview of a larger number of studies in comparison with Cooper (2025). Third, only research conducted with healthy adults was included. There is a large body of research on music making and wellbeing in the context of music therapy (e.g., Köhler et al., 2020) and with individuals suffering from severe illness such as cancer (e.g., Fancourt et al., 2016) that would require a different review.

Based on the results of this systematic review, there are different recommendations for future studies of music making and wellbeing. First, future research would benefit from a clear conceptualization of wellbeing along with corresponding measures. This would contribute to a better general understanding of wellbeing and a higher comparability between the studies. Second, wellbeing is often conceptualized as psychological wellbeing, focusing on internal and personal emotional states (e.g., Elliott et al., 2022; Keyes, 2002). Although this is one important dimension, bodily processes as well as external and social aspects should also be considered, thereby contributing to a more holistic perspective on wellbeing. Third, quasi‐experimental designs are an adequate standard when conducting research on the benefits of music making on wellbeing. RCTs should be used whenever possible, if potential biases (e.g., frustration of the non‐music control group) can be controlled or the musical activity does not require any prior training. Longitudinal studies should include follow‐up assessments to enhance the interpretation of the findings and to investigate long‐lasting effects of music making. Fourth, further research on musical activities beyond singing is needed. Amateur and professional musicians engage in a variety of activities, including brass bands or symphony orchestras. Since the benefits of choir singing partly result from physiological characteristics of singing (e.g., posture, respiration; Clift & Hancox, 2001), it is important to focus on other music making activities that go beyond these dimensions.

CONCLUSION

While music is often perceived to be intrinsically rewarding, for many individuals, positive effects on wellbeing are a major motivation for engaging in musical activities. Wellbeing has been investigated in psychological, physiological, and social dimensions. As shown in this review, the majority of studies focus on singing, leading to a need for research on instrument playing as well. Quasi‐experiments are the most common research design. Further studies could follow this line with a careful selection of the respective control conditions, including comparable activities with no relation to music. Although there is some positive evidence for the effects of music making on wellbeing, the considerable number of studies with mixed or no evidence calls for further research, since it cannot be ruled out that null‐results studies have not been published. Moreover, research clearly focuses on psychological wellbeing and should also refer to other forms of wellbeing to contribute to a more comprehensive picture of the relation between music making and wellbeing.

Taken together, this review provides a snapshot of the current state of growing research in this domain. If future research endeavors to systematically address the blind spots regarding musical activities, the critical points raised in terms of methods, and most importantly, to acknowledge the multifaceted nature of wellbeing, then a number of evidence‐based suggestions can be formulated for clinical and community contexts. For instance, a combination of social and physiological effects, though widely neglected in most research, may benefit vulnerable individuals in particular who may not engage in other activities. In this regard, care should be taken to not convey an elitist view of music making and to embrace a welcoming and low‐threshold approach.

ETHICS STATEMENT

Not applicable.

CONFLICT OF INTEREST STATEMENT

The authors declare that there is no conflict of interest.

Supporting information

Table S1.Overview of included studies.

Table S2.Risk of bias assessment Mixed Methods Appraisal Tool, version 2018 (Hong et al., 2018).

Table S3. Experimental studies (n = 42) yielding positive (+), mixed (+/o), or no (o) evidence for wellbeing measures. Numbers in bold indicate RCT studies.

Table S4. Non‐experimental studies (n = 12) yielding positive (+), mixed (+/o), or no (o) evidence for wellbeing measures.

Steinmacher, C. , & Wöllner, C. (2026). Music making and wellbeing: A systematic review of methods, musical activities, and outcomes. Applied Psychology: Health and Well‐Being, 18(4), e70191. 10.1111/aphw.70191

DATA AVAILABILITY STATEMENT

Data for this study are available in the Supporting Information and upon request.