Students’ Perceptions of Circle Singing as an Arts-Based Practice for Well-Being, Belonging and Learning: A Mixed-Methods Study in Higher Education

1. Introduction

The World Health Organization conceptualizes health as encompassing physical, mental and social well-being, extending beyond the absence of illness. This multidimensional understanding is reflected in the biopsychosocial approach underpinning the International Classification of Functioning, Disability and Health (ICF), which considers functioning in relation to health conditions and contextual factors (WHO, 2001-2018). It is also consistent with a holistic conception of health (Rizzo, 2016a) and an understanding of well-being as a balanced way of life sustained through relationships with others (Berzano, 2010).

Participation in the arts may support health and well-being by engaging aesthetic, imaginative, sensory, emotional and cognitive processes. Depending on the activity, it may also involve social interaction, physical movement and engagement with health-related issues and settings (Fancourt & Finn, 2019).

Arts-based practices span a continuum, from art therapy delivered by qualified therapists within a therapeutic relationship, to arts-in-health programmes, to participatory arts activities embedded in educational and community settings (Fancourt & Finn, 2019; Rizzo, 2016b). The present study sits at the non-clinical end of this continuum: Circle Singing was delivered within university courses by facilitators other than the course instructor, with educational rather than therapeutic aims, and without diagnostic or clinical outcomes. It is nonetheless relevant to the promotion of well-being for two reasons. First, belonging and academic anxiety are important aspects of students’ experiences in higher education (Ruan & Zhu, 2026; Idoiaga-Mondragon et al., 2025), while the vocal demands of teaching make voice management and vocal fatigue relevant considerations in teacher education (Roy et al., 2004). Second, collective vocal improvisation engages mechanisms proposed in the arts-and-health literature: breath-related bodily regulation, emotional activation, shared attention and interpersonal synchrony, and opportunities for social sharing and inclusion (Tarr et al., 2014; Vickhoff et al., 2013).

The relationship between music, health and medicine has long been investigated (MacDonald, 2012), with music therapy representing a specialized form of therapeutic practice (Alvin, 1975). In their systematic review of the role of music in promoting health and wellbeing, Viola et al. (2023) examine potential benefits across cognitive, physical, affective and social dimensions. The wider literature on musical training has also explored its relationship with speech processing and phonological skills (Overy, 2003; Patel, 2011), experience-dependent neural plasticity (Kraus et al., 2014; Thaut & Hodges, 2019) and auditory–motor connectivity (Palomar-García et al., 2017). These lines of enquiry address the cognitive and neural dimensions of musical engagement, complementing research concerned more directly with experiences of well-being, including emotional well-being in instrumental participation (Coffman & Adamek, 1999; Clift & Hancox, 2001).

Within this broader field, singing has attracted scholarly attention for its relationship with health (Hunter, 1999), with its integration of bodily, cognitive and expressive processes providing a basis for understanding its potential benefits (Welch, 2005). Research on group singing has examined relationships between musical structure, breathing and heart-rate synchronization (Vickhoff et al., 2013). Other studies have investigated perceived physical, emotional, social and spiritual benefits of choral singing (Clift & Hancox, 2010; Clift et al., 2010), alongside the potential of collective singing to support social relationships (Latimer, 2008; Silber, 2005).

The present study focuses on Circle Singing, a collective vocal improvisation practice designed by Bobby McFerrin, in which participants develop and combine vocal patterns through guided interaction. Collective improvisation can provide opportunities to develop musical communication skills (Monson, 1996) and practice coexistence through the negotiation inherent in shared musical performance (Seeger, 1977). It requires participants to understand and respond to their fellow performers’ actions through continuous adjustment and reciprocal exchange, coordinating individual contributions within a shared present (Schütz, 1964). Within this body of research, further empirical investigation is needed to clarify how specific singing practices may support well-being in different contexts. Although studies have explored singing in relation to social bonding and well-being, creative approaches such as vocal improvisation have received less attention within research on singing-based interventions (Cartier et al., 2026). Further investigation is therefore warranted into how participants experience the emotional and relational dimensions of collective vocal improvisation within higher education.

This article presents findings from a broader research project on Circle Singing conducted at Roma Tre University. The project involved 188 students enrolled in Educational Sciences and DAMS (Disciplines of the Arts, Music and Performing Arts) programmes, comprising an experimental group (n = 104) and a control group (n = 84). Previous analyses identified improvements in music perception skills and English pronunciation accuracy following the intervention (Pantano & Rizzo, 2026). Extending this research, the present investigation focuses on the experimental group’s perceptions of the intervention’s social, emotional and educational significance, with particular attention to peer relationships, anxiety associated with assessment and vocal performance, and confidence in speaking and singing in English. The study draws on a questionnaire administered before and after the intervention and a supplementary post-intervention questionnaire, both including Likert-type items, alongside three focus groups conducted in January 2026. The qualitative analysis addresses three domains: emotional well-being and engagement; social and relational experiences; and learning and professional development. Together, these data offer insight into how participants experienced Circle Singing and understood its relevance to future educational practice.

This study therefore addressed three research questions: (RQ1) What changes did students report in their comfort speaking and singing in English in front of others and in their vocal well-being, and how did they perceive the workshop’s contribution to their sense of group belonging? (RQ2) How did students describe the emotional and relational processes that characterized their experience of Circle Singing? (RQ3) How did they relate these experiences to their perceived readiness to use Circle Singing in their future educational practice?

2. Materials and Methods

2.1. Study Design and Sample

The broader research project adopted a mixed-methods design incorporating a non-randomized, quasi-experimental intervention. Data were collected at Roma Tre University between October 2025 and January 2026. The quantitative component followed a pre-test–post-test design involving an experimental group (EG) and a control group (CG), complemented by post-intervention questionnaires and focus groups with EG participants. Students were recruited voluntarily from pre-existing university courses and were not randomly assigned to groups. For the outcomes analyzed in this article, the design was a single-group pre–post design, complemented by an exit questionnaire and focus groups. The control group participated only in the pre- and post-intervention assessments of music perception and English pronunciation, reported separately, and did not complete the questionnaires examined here or participate in the focus groups.

The broader project comprised 188 students: 104 in the EG and 84 in the CG. The EG included 96 women and 8 men, with a mean age of 23.2 years (range: 19–61). Participants were enrolled in Primary Teacher Education (n = 71), Pedagogical Sciences (n = 15), and DAMS—Disciplines of the Arts, Music and Performing Arts (n = 18). The CG included 80 women and 4 men, with a mean age of 19.5 years (range: 18–26). Most were enrolled in the EduNido programme (n = 83), while one participant attended Primary Teacher Education. The present study focuses exclusively on the EG, examining participants’ perceptions of the intervention’s emotional, relational and educational significance through questionnaire responses and three focus groups. All participants provided informed consent before taking part in the research, including consent to the audio recording and automated transcription of the focus groups.

2.2. Intervention and Research Phases

The research comprised design, baseline assessment, intervention, post-intervention assessment, qualitative follow-up and data analysis. During the design phase, musical materials, English-language excerpts and assessment procedures were selected and adapted to the objectives of the broader project. Baseline data were collected in October 2025 through questionnaires and digital assessments of music perception skills and English pronunciation. The Circle Singing laboratory comprised 15 hours of activities delivered between October and December 2025. Each sessions lasted approximately one hour and followed a recurring sequence of vocal preparation, pattern learning, collective performance, closure and reflection.

The sessions were led by a facilitator with formal training in music education who was not responsible for teaching or assessing the courses. The course professor attended every session and sang alongside the students but did not lead the activities; she was also responsible for conducting the final course examination. Participants stood in a circle and were organized into subgroups according to compatible vocal ranges, both during vocal preparation and Circle Singing. Vocal preparation included breathing exercises, lip trills, vocal slides and short melodic exercises. Following the facilitator’s instructions, participants reproduced short rhythmic and melodic patterns using vowels, non-lexical syllables, scale-degree numbers or words. Major and minor pentatonic material was used to introduce patterns suitable to nonprofessional singers.

Collective performance involved a maximum of 20 participants performing at a time. An initial vocal pattern incorporating selected English lyrics was taught to the whole group and subsequently assigned to a subgroup for continuous repetition. Additional patterns were introduced and allocated to other subgroups, gradually producing a texture of simultaneous vocal parts constituting the circle song. Participants maintained their own patterns while listening to the surrounding ensemble. The number and complexity of the parts were adjusted to the group’s developing competence.

A shared gestural vocabulary enabled the facilitator to regulate entries, pauses, repetition, tempo, dynamics and closure. Activities including call-and-response exchanges, alternating subgroups, interlocking patterns, and guided variation were organized so as to combine repetition and variation, while accommodating different levels of vocal and musical experience. Each performance ended with a clearly signaled closure, followed by a short reflective phase focusing on the music and linguistic materials performed. Participants were invited to identify, reproduce or, where appropriate, transcribe aspects of the material, including melodic patterns, metre, rhythm and form.

Post-Intervention assessment took place in December 2025 and included repeated questionnaires, a supplementary final questionnaire, and the music-perception and pronunciation assessments administered at baseline. Three focus groups followed in January 2026 to explore participants’ experiences of the intervention.

2.3. Data Collection, Instruments and Procedures

The broader research project investigated musical and linguistic outcomes alongside the perceived social, learning, physical and emotional effects of the intervention on participants. Music perception and English pronunciation were assessed before and after the intervention using Mini-PROMS (Law & Zentner, 2012) and SpeechAce (Nguyen et al., 2025). Mini-PROMS assessed melody, tuning, accent and tempo perception, while SpeechAce provided English pronunciation scores based on participants’ spoken recordings across four intermediate lessons provided by the digital tool. The present contribution focuses on participants’ perceptions of the intervention’s social, physical, emotional and educational impact, while analyses of musical and linguistic performance form the subject of separate manuscripts (Pantano & Rizzo, 2026).

Two questionnaires, both developed and conducted in Italian, collected self-assessments and evaluations of the Circle Singing experience. The first was administered before and after the intervention, enabling comparisons of repeated items; the second was administered only after the intervention and included items that could be meaningfully answered only once participants had completed the experience. Questionnaire responses were linked across PRE and POST using a self-generated code based on participants’ initials, date of birth and sex. The code was used solely for matching responses. The course professor did not have access to individual responses, and participants were informed of this arrangement.

The quantitative component was complemented by three focus group sessions, each lasting approximately 50 minutes, conducted remotely and audio-recorded. Participants were recruited voluntarily from the experimental group and allocated to three focus groups. The total sample consisted of 12 participants: four from DAMS and 8 from Educational Sciences. The first focus group comprised 2 DAMS students and 3 Educational Sciences students; the second comprised one DAMS student and 3 Educational Sciences students and third comprised one DAMS student and 2 Educational Sciences students. The discussions followed a moderator-led format in which topics were introduced to stimulate participants’ reflections and exchanges (Acocella, 2005). The moderator encouraged contributions from all participants, ensuring that everyone had an opportunity to discuss each topic. The focus group material was subsequently subjected to thematic analysis involving coding and the development and interpretation of themes and main analytical domains.

2.4. Quantitative Data Analysis

The first questionnaire comprised 28 items covering participants’ backgrounds, self-perceived abilities, confidence and educational beliefs. Seven items addressed perceived English pronunciation in speaking and singing, voice management, breath control, and the perception of melody, rhythm and tuning. Six collected information on current and recent English-language study, professional status as a musician or singer, and formal professional musical training. A further six items explored comfort when speaking or singing in English before a group, perceived self-efficacy to design and lead school singing workshops, and comfort using the spoken and singing voice in teaching. Seven items addressed beliefs about the contribution of music to pronunciation and of group singing to engagement and peer-bonding, evaluations of previous English-language and pronunciation learning experiences, and the perceived need for practical English-language workshops in schools. The remaining two items addressed vocal fatigue associated with teaching activities and concerns about managing the voice in the classroom.

The post-intervention questionnaire comprised 16 items. Two addressed perceived improvements in group belonging and reductions in anxiety when pronouncing English, while two concerned perceived improvements in the use of the singing and speaking voice. Five explored linguistic benefits, including spelling-to-sound relationships, discrimination of minimal pairs, speech sound production, retention of pronunciation learnt during the activities, and vocabulary development. One examined whether feedback from the digital assessments had increased awareness of personal strengths and areas for improvement. Three addressed overall perceptions of musical and linguistic development and knowledge and use of the vocal apparatus. The remaining three concerned the intention to use Circle Singing in future educational practice and its perceived suitability for teaching music and English in primary and secondary schools. Items concerning professional status as a musician or singer, or attendance at private or public English language courses, used a dichotomous response format. Items relevant to this study were assigned codes to facilitate analysis and are presented in Table 1 alongside their meanings and thematic areas, as specified in the original codebook accompanying the dataset. Response formats varied according to item content: items concerning professional status as a musician or singer, or attendance at private or public English-language courses, used a dichotomous (yes/no) format, while the remaining items used Likert-type scales^[1].

Questionnaire items were analyzed according to their level of measurement rather than their numerical coding, with response categories assigned ascending numerical codes (1–5) to facilitate statistical processing. This coding establishes a rank order among categories but does not imply that the psychological distance between adjacent categories is equal (Jamieson, 2004). Consistent with established measurement guidance, single ordinal items were therefore not treated as continuous, interval-level variables. Because the present questionnaire items were not designed and validated as indicators of a single latent construct, no composite scores were computed, and each item was analyzed individually at the ordinal level. Descriptive statistics for ordinal items are reported as absolute and relative frequencies (n, %) together with median and interquartile range (Md [Q1–Q3]), rather than means and standard deviations. For pre–post comparisons of paired ordinal items, the Wilcoxon signed-rank test was used, as it evaluates the rank order of within-subject differences without requiring the assumption of a normally distributed continuous variable. Effect size was estimated using the matched-pairs rank-biserial correlation, calculated from the sum of positive and negative ranks yielded by the Wilcoxon test output, as r_rb = (W⁺ − W⁻)/(W⁺ + W⁻), where W⁺ is the sum of positive ranks and W⁻ is the sum of negative ranks; a positive value indicates a prevalence of change toward higher POST scores. “Don’t know” responses were treated as missing answers and reported separately in the frequency distributions rather than being incorporated into the ordinal scale. Tests were two-sided, with α = .05. The Holm correction was applied to the group of 15 pre- to post-intervention comparisons of agreement-scale items; the seven self-assessed competence items were analyzed as a separate set. Of the 15 items, 11 showed statistically significant changes, while 4 did not. The four non-significant items addressed participants’ evaluation of their past English learning experiences, both at school and in private academy contexts^[2]. Given their limited relevance to the scope of this article, they are not discussed further in the text and not reported in Table 2.

For interpretive purposes, and to support integration with the qualitative findings from the focus groups, selected items were grouped into thematic areas defined a priori on conceptual grounds. Each item within a thematic area is therefore reported and interpreted individually. All analyses were conducted using IBM, SPSS vers. 30.0.0.0

2.5. Thematic Analysis of Focus Groups

Three focus group sessions were conducted in January 2026, following the Circle Singing intervention. The focus groups were moderated by the facilitator who had led the Circle Singing sessions; the course professor was not present at any of them, and all participants had already taken the course examination. Participants were informed that their contributions would not be shared with the professor. The sessions explored participants’ experiences, thoughts and feelings through discussions guided by topics introduced by the moderator. The focus groups served as a complementary means of inquiry, exploring in greater depth dimensions also addressed by the quantitative component (Bertrand et al., 1992, p. 199). Adopting the role of an “interventionist moderator” (Macnaghten & Myers, 2004, p. 71), a structured discussion inquired three analytical domains (Table 3), balancing the need to limit digressions with opportunities for participants to elaborate on their experiences and perspectives, facilitating moving on to the next topic once the preceding one had been sufficiently explored. Group composition combined elements of homogeneity and heterogeneity (Acocella, 2005): participants shared the experience of the same intervention, while each group included students from both DAMS and Educational Sciences programmes. This arrangement brought together students from different educational backgrounds, some of whom already knew one another while others did not, with the aim of balancing familiarity and openness to different perspectives and encouraging the sharing of personal experiences and opinions (Krueger, 1994).

The first analytical domain concerned physical and emotional experiences of the laboratory, including awareness of vocal technique, breathing and bodily involvement, as well as emotional engagement and activation during Circle Singing. Participants were also invited to reflect on how these experiences related to their motivation and course attendance. The second area addressed social and relational experiences, including peer bonding and relationships with the Circle Singing facilitator and the professor who participated in the laboratory. The third concerned learning and professional development: participants compared Circle Singing with previous experiences of learning English, particularly pronunciation, and reflected on their perceived capabilities as future educators teaching music and English. The analytical process was informed by literature concerning focus group and thematic analysis (Bertrand et al., 1992; Braun & Clarke, 2006). MAXQDA supported transcription and the organization of the material. The workflow comprised: transcription and checking; familiarization and annotation; coding; theme development and interpretation; and written reporting. Audio recordings were imported into MAXQDA, where automated transcripts were generated. The resulting texts were subsequently checked manually against the recordings to correct transcription errors, with particular attention to words and expressions in languages other than Italian. Repeated words were omitted, and inaudible passages were explicitly marked. Conventional punctuation was used to support readability, without additional notation of suprasegmental features. During familiarization, the recordings were listened to alongside repeated reading of the transcripts. Notes documented how the moderator introduced topics, how participants responded and when the discussion moved between topic areas. In each session, approximately 38 of the 50 minutes were devoted to well-being, emotional engagement and relationships with peers and teaching staff. The remaining approximately 12 minutes addressed perceived English-language learning outcomes, comparisons with previous educational experiences and perceived self-efficacy in relation to future teaching. Different colors were used to highlight passages relevant to general analytical domains, organizing the material and supporting familiarization through an initial inventory of discussion points. Red identified passages concerning participants’ relationships with the professor, including perceived changes in anxiety associated with the professor’s role as an examiner. Blue marked accounts of physical well-being related to voice use, while green highlighted emotional engagement and arousal during Circle Singing. Yellow identified passages concerning anxiety associated with English pronunciation and participants’ perceived self-efficacy as future teachers. Magenta marked practical comments, suggestions about the intervention and other observations not covered by these areas. This preliminary organization of the material preceded more detailed coding and the subsequent development and interpretation of themes. Following familiarization, deductive codes derived from the research questions were developed and applied manually to relevant passages across the three transcripts. The analysis followed a theoretically driven approach, focusing on aspects of participants’ experiences relevant to the study’s analytical interests (Braun & Clarke, 2006, p. 84). Coded extracts were collated and compared across the dataset to identify patterns of meaning and develop themes addressing the research questions. This process was complemented by an examination of recurring words and their meanings in context across the three transcripts. A detailed feedback has been provided to participants through a written report conveyed by email to present and share the results of the thematic analysis.

3. Results

As previously noted, detailed analyses of music perception and English pronunciation outcomes are the subject of separate manuscripts arising from the broader research project, reported in Pantano and Rizzo (2026). These outcomes involve a comparison group and differ from those examined here; they are not used to interpret the present findings.

The following subsections present the questionnaire results and focus group findings, focusing on participants’ perceptions and experiences of the intervention, with particular attention to well-being, peer relationships, perceived self-efficacy, engagement in higher education and the overall Circle Singing experience.

3.1. Questionnaire Results

Individual questionnaire items were analyzed separately, and the findings were organized into eight thematic areas reflecting the topics explored in the study (Table 2). Pre–post analyses included 15 items asking participants to indicate their level of agreement with specific statements and seven items assessing their perceived skills. Of the 15 agreement items, 11 are reported in Table 2 because of their relevance to the focus of the present article. Ten items administered only at the end of the intervention were analyzed descriptively. The thematic areas provided a framework for presenting and interpreting the findings and were not treated as composite measures.

For clarity, the narrative presentation brings these areas together under three broader headings: (1) self-perceived pronunciation, vocal and musical perception skills; (2) perceived teaching readiness, educational beliefs, overall learning and development, and anticipated educational use of Circle Singing; and (3) vocal well-being and emotional and relational experiences.

These broader headings guide the presentation of the findings and do not represent composite scales or empirically validated dimensions. The first analytical area concerns students’ self-perceived competence in English pronunciation, vocal management, and music perception. Among the 104 participants in the experimental group, Wilcoxon signed-rank tests indicated statistically significant PRE–POST changes across all seven items (all p < .001). For each item, more participants reported higher scores at POST than lower scores. Rank-biserial correlations ranged from rᵣᵦ = .568 for voice management (VML) to rᵣᵦ = .762 for English pronunciation when speaking (EPLS). Median ratings for EPLS and rhythm perception (RPL) increased from 3 (Intermediate) at PRE to 4 (Good) at POST. For the remaining items, medians remained at 3, despite statistically significant shifts towards higher response categories.

The second analytical block concerns participants’ sense of adequacy in using their voice — in speech, singing, and breath control — within specific teaching contexts. All eight items in this analytical block showed statistically significant pre- to post-intervention changes following Holm correction (all p_Holm < .05). Participants reported greater comfort speaking and singing in English in front of a group (CSEG and CCEG), with median ratings increasing from 2 to 3 for both items (r_rb = .86 and .80, respectively). Perceived ability to design and lead school music and singing workshops also increased markedly (DMSL and DRTL), with medians rising from 2 to 3 (both Holm-adjusted p < .001) and correspondingly large effect sizes (r_rb = .93 and .95, based on 91 and 92 valid pairs, respectively). At POST, 42 participants (40.4% of the full experimental group) selected one of the two highest agreement categories for DMSL, compared with 5 (4.8%) at PRE; for DRTL, the corresponding figures increased from 5 participants (4.8%) to 43 (41.3%). These descriptive percentages use all 104 participants as the denominator, including those who selected ‘I don’t know. Comfort using the singing voice in teaching (CUVS) increased from a median of 2 to 4, and comfort using the speaking voice in teaching (CUVP) increased from 3 to 4, both with large effects (r_rb = .89 and .66, respectively). Reported vocal fatigue (VFAS) and concern about managing the voice in the classroom (WVMC) decreased significantly (p_Holm = .015 and .008, respectively). For VFAS, the median remained at 2 despite a shift toward lower ratings, while for WVMC the median decreased from 3 to 2. The corresponding rank-biserial correlations were −.51 and −.46, with negative values reflecting favorable change on these negatively worded items. Overall, participants’ accounts of increased perceived adequacy in situated, active voice use converged consistently across all eight indicators considered.

The last analytical area examines participants’ self-reported outcomes addressing physical/phonatory well-being and emotional/relational well-being. For physical/phonatory well-being, perceived improvement in the use of the singing and speaking voice (SVI_FINAL and SPVI_FINAL, respectively) was considered alongside pre- to post-intervention changes in vocal fatigue (VFAS) and concern about voice management (WVMC). For emotional/relational well-being, perceived improvement in group belonging (GB) and reduction in performance-related anxiety about English pronunciation (PAR) were considered alongside pre- to post-intervention changes in comfort speaking and singing English in front of a group (CSEG and CCEG, respectively). WVMC was also considered within this second area as a cross-cutting indicator, given its relevance to concerns about voice management in a relational, classroom setting.

A consistent pattern of convergence emerged across indicators. From pre- to post-intervention, comfort speaking and singing English in front of a group increased significantly (CSEG and CCEG, both p_Holm < .001), while concern about managing the voice decreased significantly (WVMC, p_Holm = .008). At post-intervention, the majority of participants also reported a strong sense of improvement in group belonging (82.7% agreement) and a reduction in performance-related anxiety (61.6% agreement). For physical/phonatory well-being, vocal fatigue decreased significantly (VFAS, p_Holm = .015), and a majority of participants reported improved use of the singing voice (66,4%) and the speaking voice (56.7%). These indicators are treated as belonging to a priori conceptual thematic areas rather than as evidence for a validated “emotional/relational well-being” or “physical well-being” scale.

3.2. Thematic Findings from the Focus Groups

The thematic analysis generated twenty-four codes, using descriptive labels formulated by the researcher, and organized into six themes reflecting recurring patterns of meaning (Braun & Clarke, 2006). Participants’ accounts described positive experiences across the three analytical domains, alongside emotional challenges and ambivalent responses captured within the six themes. Table 3 presents these themes and codes alongside 24 illustrative quotations from participants. The following subsections describe and interpret each theme.

3.2.1. Finding Renewed Energy and Motivation Within the University

The first theme concerned participants’ accounts of well-being during the intervention. In responding to questions about their emotional experiences, participants spontaneously connected the enjoyment and energy associated with Circle Singing to their motivation to attend classes. One participant described how anticipating the shared singing activity had changed their feelings about attending university:

“Monday became my favorite day of the week. I looked forward to Monday. I was really happy to come because I knew we would have that moment of singing together.” (FG A1, segment 18)

This account connects the emotional value of shared singing with a positive anticipation of university attendance. The experience was also valued as an opportunity for practical participation within an academic setting. As another excerpt illustrates:

“I am always happy to see workshops at university. Real workshops, I mean: the opportunity to actually try out a method that is innovative—not new, perhaps, but innovative in the university context.” (FG C2, segment 26)

The emphasis on “real workshops” suggests that the opportunity to put a method into practice was a meaningful feature of the experience. Participation was therefore associated both with the enjoyment of singing and with the value attributed to learning through practical activity. Attendance also featured in the account of a participant whose work commitments usually limited their presence at university:

“First of all, I want to say that I do not normally attend classes because my work gives me very little opportunity to be there. In fact, I hardly attend any courses.” (FG B2, segment 6)

Reflecting on attending university almost every Monday, the participant emphasized the relational dimension of the course, suggesting a connection between this theme and “Strengthening Peer Relationships through Shared Singing”:

“Having ended up at university practically every Monday, and having seen not only the subject matter but also how this activity was presented, it was clear that relationships were at the centre. […] In my view, that is the most important aspect of any course of study, or indeed any educational experience.” (FG B3, segment 50)

Together, these accounts illustrate how emotional engagement, opportunities for practical learning and the relational experience of the course were connected to participants’ reported motivation and attendance.

3.2.2. Negotiating the Emotional Demands of Participation

This theme captured the emotional demands of participating in Circle Singing, particularly the initial embarrassment of singing in front of others and uncertainty about one’s musical background. Positive experiences sometimes involved working through these concerns. One participant remarked, “There is no time to feel embarrassed or to let embarrassment hold you back” (FG A4, segment 41), while another described singing before others as “a challenge that I personally feel I have overcome” (FG C1, segment 16). One participant’s account illustrated this process particularly clearly. Initially, comparing their musical background with that of other students raised doubts about their ability to participate. Reflecting afterwards, however, the participant recognized the emotional benefits of leading a Circle Song and improvising vocally. Eye contact during shared singing and the active participation of teaching staff were described as helping to connect initially unfamiliar people and reduce perceived interpersonal distance (FG B3, segments 26–30). Alongside these challenges, collective singing offered opportunities for relaxation and a sense of community. One participant associated well-being with the shared experience of singing and the opportunity to interact beyond the usual routine of sitting at a desk and taking notes (FG C3, segment 54). These accounts suggest that negotiating embarrassment was intertwined with the relational experience of participation, connecting this theme with those concerning peer relationships and closeness to teaching staff.

3.2.3. Enhancing Relationships with Peers Through Shared Singing—Closer Relationships with Teaching Staff

Stronger social connections featured prominently in participants’ accounts, alongside emotional well-being and motivation to attend the course. These experiences concerned relationships with both peers and teaching staff. The Circle Singing sessions were led by a facilitator who also served as the professor’s assistant, while the professor participated in every session, singing alongside the students. Regarding peer relationships, participants valued the opportunity to share emotions and interact through practical musical activity. One participant associated collective singing with relaxation and a sense of community, contrasting this experience with the usual routine of sitting at a desk and taking notes (FG B4, segment 73). Another described how singing encouraged interaction through eye contact and smiles:

“We had to interact in some way, even without speaking. […] We looked at one another and smiled because we recognized that we were making an effort to sustain the notes you asked us to sing, while also enjoying ourselves.” (FG C2, segment 24)

This account highlights how shared musical effort became an occasion for mutual recognition and enjoyment. The connection between attendance and relationship building was also explicit: “It was the only course I attended in person from beginning to end, and the only course where I formed good relationships” (FG B1, segment 14). Participants’ descriptions also suggested that belonging to the group could coexist with individual expression. One student stated, “I was able to express my individuality while remaining in harmony with the others” (FG C1, segment 8). This formulation echoes the organization of Circle Singing, in which individual vocal contributions form part of a collective performance. Shared participation also shaped perceptions of the facilitator and professor. Reflecting on the final performance, one participant explained:

“Performing together in the final performance also helped a great deal, because it was as though I had seen you both in a different context.” (FG C1, segment 6)

Another described feeling closer to the professor and developing a relationship with her (FG C1, segment 12). The professor’s willingness to participate was particularly significant in an account linking her involvement to trust:

“She joined in herself, and that set an example. It made me think, ‘OK, I can trust her.’” (FG B2, segment 20)

These accounts suggest that making music together provided opportunities to experience teaching staff beyond their usual instructional roles, contributing to a sense of familiarity and trust.

3.2.4. Physical Well-Being Through Teacher Training—Taking Learning into Educational Practice

This final subsection examines participants’ accounts of physical well-being alongside their reflections on learning and professional development. Participants experienced Circle Singing both as learners and as prospective practitioners considering its educational applications. Before the collective singing activities, preparatory exercises provided instruction in vocal technique, breathing, English pronunciation and musical skills. The analysis therefore considers two interconnected aspects: perceived physical benefits associated with using the voice, breath and body, and the relevance of this learning to participants’ current abilities and future teaching practice.

Participants reported noticing differences in their voice and singing, suggesting that the learning extended beyond the laboratory into everyday professional practice. One participant described learning “how to prepare the voice for lessons and how to help it recover afterwards”, alongside musical skills they had not previously possessed (FG C2, segment 22). Indeed, for a participant already working in education, these exercises had become part of their daily routine:

“At six in the morning, in the car, I do the exercises so that I feel much better when I get to work in the classroom.” (FG A1, segment 20)

Other accounts focused on the possibility of leading musical activities in future educational settings. One participant reflected on their previous uncertainty:

“Without this workshop, if I had been asked to teach music during a period of supply teaching, I really would not have known where to start.” (FG C3, segment 74)

These reflections connected the experience of participating in Circle Singing with a greater sense of readiness to use it in teaching. As one participant expressed it:

“Now that we have finished the course, I think we all have more confidence and experience to repeat the activity and let children experience it too.” (FG B4, segment 98)

The accounts thus linked perceived vocal well-being with the acquisition of practical resources and confidence in their educational application.

4. Discussion

Taken together, the questionnaire findings and focus group accounts suggest that participants experienced Circle Singing as an opportunity for learning, social connection and greater comfort in using their voices within higher education. These findings can be situated within the broader literature on the contribution of the arts, and of music specifically, to psychological well-being through positive emotion, engagement, relationships, meaning, and accomplishment (Adiasto et al., 2022; Fuentes-Sánchez et al., 2025); to support participation and inclusion (Booth & Ainscow, 2014; Rizzo, 2014, 2016b; Sawyer, 2012; WHO, 2018); to act as a socio-relational enhancer (Tarr, Launay, & Dunbar, 2014); and to foster cognitive and linguistic development (Cooper, 2020; Habib et al., 2016; Kraus et al., 2014; Patel, 2011).

The items assessing English pronunciation, voice management, breath control and musical perception all showed significant PRE–POST increases among the 104 participants in the experimental group (all p < .001; rank-biserial correlations ranging from .568 to .762). These changes indicate more favorable assessments of participants’ own abilities following the workshop. They resonate with the focus group domain “Learning and professional development”, in which participants valued practical engagement with musical activities as a complement to theoretical instruction. The exit questionnaire provided a consistent descriptive picture: 80.8% selected one of the two highest response categories for perceived musical growth, 70.2% for linguistic growth and 77.9% for increased knowledge and use of the vocal apparatus. The findings also suggest that participants felt better prepared to apply their learning in teaching contexts. Perceived ability to design and lead school music and singing workshops increased substantially, with median ratings rising from 2 to 3 for both DMSL and DRTL and large rank-biserial effect sizes (.928 and .948, respectively; both Holm-adjusted p < .001).

Participants also reported greater comfort using their singing and speaking voices in teaching, with particularly pronounced change for the singing voice. The exit questionnaire further indicated favorable views of the methodology’s educational applicability. The two highest response categories were selected by 86.5% for intended future use of Circle Singing, 91.3% for its perceived validity in primary education and 92.3% for its perceived validity in secondary education. These evaluations suggest that participants regarded the experience as relevant beyond the immediate university setting. Participants’ dual position — as learners experiencing the activities firsthand and as prospective teachers considering their educational use — appears to support this interpretation.

The emotional and relational findings help contextualize these perceptions of readiness, addressing also themes as anxiety and emotional engagement. Comfort speaking and singing English in front of a group increased significantly, with median ratings rising from 2 to 3 for both CSEG and CCEG (both Holm-adjusted p < .001). In the exit questionnaire, 61.5% selected the two highest response categories for the statement that the workshop had reduced their performance anxiety in English pronunciation. This retrospective evaluation complements the repeated assessments of comfort, although it does not constitute a PRE–POST measure of anxiety. Focus group accounts illuminate the experiences underlying these responses. Participants described initial embarrassment, uncertainty about their musical background and concerns about performing in front of others. Some subsequently associated greater ease with shared singing, mutual attention and the active participation of teaching staff.

The relational dimension was also prominent in the exit questionnaire, where 82.7% selected the two highest categories for improved group belonging. Accounts of exchanged glances, shared effort and enjoyment suggest that relationships developed through the activity itself.

The findings concerning vocal experience are particularly relevant to teacher education, linking professional practice and teacher training to physical well-being. Participants reported significantly less vocal fatigue after a school day (VFAS; Holm-adjusted p = .015) and less concern about managing their voice in class (WVMC; Holm-adjusted p = .008). These items capture related but distinct dimensions: VFAS addresses a perceived physical experience, whereas WVMC concerns apprehension about classroom voice use. The unchanged median for VFAS does not contradict its significant result, which reflects the distribution of paired responses rather than the median alone. In the exit questionnaire, 66.3% selected the two highest categories for improved use of the singing voice and 56.7% for improved use of the speaking voice. Focus group accounts connected these perceptions with practical knowledge about preparing the voice before teaching and supporting its recovery afterwards. Such experiences suggest that vocal preparation was meaningful both as part of the musical activity and as a resource for teaching. The relevance of this dimension is reinforced by Roy et al. (2004), who reported a higher lifetime prevalence of voice disorders among teachers than non-teachers (57.7% versus 28.8%).

Finally, the emotional engagement described in the focus groups was accompanied by favorable questionnaire responses concerning belonging and inclusion. Whereas GB captured participants’ perceived improvement in group belonging, GSAI, expressed the belief that group singing promotes engagement and active inclusion. Participants described enthusiasm for attending lessons, stronger peer relationships and a desire to attend the course consistently. Some also reported feeling “energized” despite the sessions taking place at the end of the day (FG B1, segment 10). These accounts suggest that shared singing was experienced as both energizing and socially meaningful, aligning with literature describing collective singing experience as producing euphoria or “singers’ high” (Jacob et al., 2009). Because Circle Singing was implemented as a non-clinical educational practice, the findings should be interpreted in terms of well-being promotion rather than therapeutic effectiveness. The pattern reported by participants—greater ease in front of others, closer relationships with peers and teaching staff, and less vocal fatigue—is consistent with mechanisms proposed for participatory arts, including embodied regulation through breath and voice, emotional activation, shared attention and opportunities for social sharing (Tarr et al., 2014; Vickhoff et al., 2013). These remain interpretive hypotheses: the single-group design and self-report data do not allow the proposed mechanisms to be tested. Future studies could combine validated measures of well-being with observational or physiological indicators of interpersonal synchrony. Taken together, the questionnaire responses and focus group accounts suggest that participants perceived Circle Singing as fostering peer bonding and positive emotions through the shared experience of singing. Participants also reported reduced anxiety related to English pronunciation, greater comfort using their singing voice in teaching, less vocal fatigue and greater awareness of breathing and voice use. As these findings derive from self-reports in a single group, they describe perceived changes rather than demonstrated causal effects. Finally, participants regarded Circle Singing as a valuable pedagogical tool for use in future academic and professional contexts and as a means of supporting English language learning, while endorsing the view that group singing promotes engagement and inclusion.

5. Limitations

Several limitations should be considered when interpreting these findings. First, participation was voluntary, and participants were recruited from existing university courses at a single institution. Self-selection may therefore have favored students with a particular interest in musical activities or collaborative learning. The predominantly female sample and the specific disciplinary context further limit the generalizability of findings. The focus groups involved 12 volunteers, representing 11.5% of the experimental group: four from DAMS (33.3%) and eight from Educational Sciences (66.7%). Given the small sample and voluntary participation, their accounts cannot be assumed to represent the experiences of the experimental group as a whole. Second, the design for the questionnaire outcomes was a single-group pre–post study without a comparison group. Reported changes may therefore reflect maturation, repeated testing, regression to the mean, seasonal variation (particularly for vocal fatigue, measured between October and December), response shift in self-assessment, or non-specific effects of taking part in any active, group-based workshop. Third, the questionnaires assessed participants’ perceptions rather than objectively measured competence, teaching effectiveness, or clinical indicators of well-being. Responses may have been influenced by social desirability, expectations about the intervention, and the relationships developed with teaching staff. The course professor attended the sessions, singing alongside the students, and later conducted the course examination, while the focus groups were moderated by the facilitator who had led the sessions. Questionnaires were completed anonymously, and the professor did not have access to individual responses; however, a self-generated code was used to pair each participant’s pre- and post-intervention responses. Transcripts were generated automatically and reviewed by the research team, which also carried out the coding; this team included members involved in teaching or delivering the intervention. Fourth, the questionnaire items addressed distinct aspects of experience and were analyzed individually. Their thematic organization reflected conceptual proximity rather than empirically validated psychometric dimensions. No composite scores were calculated, and Cronbach’s α was not computed for these groupings, as they were not designed as scales measuring a single underlying construct. The findings should therefore be interpreted in relation to the specific content of each item, rather than as evidence of improvement on a general scale of well-being or competence. “I don’t know” responses were treated as missing in paired analyses, resulting in different numbers of valid pairs across items. If participants with complete responses differed systematically from those without them, the estimates may not fully represent the experimental group. Fifth, convergence between questionnaires and focus groups strengthens the contextual interpretation of the findings, but both sources remain based on participants’ reports and may share similar response biases. Finally, no follow-up assessment was available to establish whether the reported changes persisted or translated into subsequent teaching practice. Perceived readiness and intentions to use Circle Singing should therefore be distinguished from its actual implementation and effectiveness in educational settings. Future research could combine validated measures of specific well-being constructs with qualitative accounts, independent observations of teaching practice, and longer-term follow-up.

6. Conclusions

This article has presented findings from a mixed-methods study conducted between October 2025 and January 2026 at Roma Tre University, examining Circle Singing as an educational approach combining collective vocal improvisation, disciplinary learning and social interaction. Within a broader research project investigating musical perception and English pronunciation, this contribution focused on participants’ perceptions of well-being, peer-bonding, competence and readiness for teaching. Questionnaire responses were integrated with a thematic analysis of three focus groups to explore both reported changes and the meanings participants attributed to their experiences.

The findings suggest that participants valued Circle Singing as an opportunity to develop confidence in their abilities, establish stronger peer relationships and engage with teaching staff through shared musical activity. PRE–POST comparisons indicated more favorable self-assessments of musical, linguistic and vocal skills, greater comfort using the voice in social and teaching contexts and lower self-reported vocal fatigue and concern about classroom voice management. Focus group accounts contextualized these changes, highlighting enjoyment, renewed motivation to attend classes, the negotiation of initial embarrassment and the value of practical experience for future teaching.

The study’s contribution lies in showing how participants experienced learning, emotional engagement and relationships as interconnected aspects of collective singing. Although the design does not establish causal effects or demonstrate lasting improvements in well-being or teaching competence, the convergence of questionnaire findings and qualitative accounts supports further investigation of Circle Singing in higher education and teacher preparation. Future research should examine the persistence of these reported changes and their transfer into educational practice.