Abstract
This controlled pre–post quasi-experimental study examined whether a school-based sports dance program targeting social anxiety could enhance physical self-esteem among adolescents with special education needs in Chinese school settings. Participants were non-randomly assigned to either an intervention or comparison group based on school scheduling. Adjusted multilevel models were employed to account for baseline differences and the clustered nature of the data. The sports dance program was associated with a sustained reduction in social anxiety and a concurrent increase in physical self-esteem, with larger improvements observed among students with greater baseline severity. Mechanism analyses suggested that body-related self-processes, including appearance comparison, body shame, and self-focused attention, partially accounted for the linkage between social anxiety and physical self-esteem. Contextual factors further shaped outcomes, as supportive school climate and inclusive peer norms appeared to buffer the negative pathway, whereas bullying exposure amplified risk. A translational prediction tool was developed to identify students at elevated risk of low physical self-esteem at follow-up, supporting school-based screening and tiered support. Overall, while the quasi-experimental design precludes definitive causal conclusions, ecology-informed sports dance activities may represent a feasible strategy for promoting well-being in special education contexts.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s40359-026-05137-6.
Keywords: Sports dance, Social anxiety, Physical self-esteem, Special education, Quasi-experimental study, School-based intervention
Introduction
In special education settings, adolescents are widely regarded as a high-risk group for mental health problems because of differences in physiological and cognitive development, and because they often remain in a socially disadvantaged position for a long period of time [1, 2]. Empirical evidence indicates that students with special education needs tend to report higher levels of anxiety in social situations than their typically developing peers, and that such negative affect may be more persistent over time [3]. Social anxiety in this group is not only related to communication difficulties, but also to internalised stigma associated with perceived differences. When social anxiety emerges during secondary school years, it may hinder sustained school participation and social integration, gradually shaping avoidance patterns and impairing interpersonal functioning. In practice, repeated social setbacks can consolidate fear of evaluation and reduce confidence in social skills, thereby increasing the likelihood of a cumulative deterioration in social adaptation4-5. Consequently, identifying feasible school-based strategies to prevent the maintenance and escalation of social anxiety in adolescents with disabilities remains a pressing issue in disability psychology research [4, 5].
From a developmental perspective, physical self-esteem is a particularly dynamic and protective component of self-concept. During adolescence, rapid physical changes and heightened sensitivity to social evaluation make physical self-esteem more vulnerable, yet also more malleable. physical self-esteem goes beyond appearance satisfaction and reflects a broader appraisal of physical competence, bodily functioning, and perceived self-worth [6]. For adolescents with special education needs, functional limitations or developmental differences may place them at a disadvantage in social comparison contexts. However, higher physical self-esteem may serve as a cognitive buffer, reducing stress responses and lowering the risk of secondary emotional problems. In addition, physical self-esteem is closely linked to emotion regulation, engagement in school activities, and the maintenance of social relationships. Adolescents who feel more confident in their bodies tend to display more adaptive social behaviors, which may help counter loneliness and withdrawal associated with discrimination or exclusion. Therefore, physical self-esteem should be considered not only an outcome of interventions, but also a potential driver of longer-term social participation and integration among adolescents with special education needs [7].
Importantly, the relationship between social anxiety and physical self-esteem is embedded in the school ecology, and the Chinese educational context may introduce additional moderators. Under collectivist cultural values and a relatively competitive evaluation system, adolescents with disabilities may face pressure from both mainstream expectations and within-school peer dynamics. Classroom climate in special education settings can shape students’ body-related self-evaluations, and supportive classes that emphasise mutual help and collective belonging may attenuate the negative effects of social anxiety on physical self-esteem [8, 9]. In contrast, when performance and short-term outcomes are highly emphasised, anxiety may increase avoidance and reduce willingness to participate in public or cooperative activities. Moreover, heightened parental expectations for social success may further amplify perceived evaluation pressure among adolescents with disabilities. Given such contextual diversity, school-based psychological programs in China should pay specific attention to social-evaluative sensitivity within collective school cultures when designing and implementing interventions [10].
Against this background, the present study evaluates a school-based group program delivered through sports dance, using a controlled pre–post design in real-world school settings. The study pursues three objectives. First, it examines whether a structured sports dance program is associated with reduced social anxiety and improved physical self-esteem over time. Second, it tests a mechanism model to determine whether changes in social avoidance and body-related self-processes help explain improvements in physical self-esteem. Third, it explores whether contextual factors, such as school climate and peer norms, shape intervention-related change patterns, thereby providing actionable evidence to support scalable, ecology-informed practices in Chinese special education and inclusive schooling contexts [11].
Based on the theoretical framework and objectives outlined above, we postulated the following hypotheses: (1) the sports dance program would lead to a significant and sustained reduction in social anxiety and a concurrent improvement in physical self-esteem; (2) the intervention’s impact on physical self-esteem would be partially mediated by reductions in maladaptive body-related self-processes, specifically appearance comparison, body shame, and self-focused attention; and (3) school-level ecological factors, such as a supportive climate and inclusive peer norms, would serve as significant moderators that buffer the negative impact of social anxiety on physical self-esteem.
Literature review
Social anxiety in adolescence
Adolescent social anxiety, characterized by a persistent fear of negative evaluation in social situations (e.g., public speaking, peer interactions), often exists in a widespread subclinical state. Even when subclinical, it significantly impairs students’ classroom engagement, social participation, and overall school adaptability, highlighting the need for targeted school-based mental health care [12].
Adolescence is a critical developmental period characterized by the onset or exacerbation of social anxiety. As self-awareness matures, adolescents become acutely sensitive to peer recognition and social status, which amplifies the impact of external evaluations on their self-worth. Concurrently, frequent group interactions and performance demands in school settings make negative evaluation cues more salient. This heightened perception often causes anxiety to generalize from specific situations into persistent avoidance behaviors.
According to cognitive behavioral theory, social anxiety is maintained by a vicious cycle of negative social beliefs, self-focus and safety behaviors. When people anticipate embarrassment or rejection, they tend to focus excessively on their own performance and physiological reactions, and alleviate the discomfort in the moment by reducing interaction, avoiding eye contact, or over-preparing. Although these behaviors can relieve anxiety to some extent, they reduce the frequency of corrective experiences and thus increase the sense of threat in social situations. Post-event rumination also leads to feelings of shame and avoidance.
Physical self-esteem and body-related self-processes
physical self-esteem refers to an individual’s overall sense of value in their own body, including subjective satisfaction with appearance and evaluations of physical fitness, ability, function, etc. Different from the general concept of body image, it focuses more on the self-esteem component, so it is closely related to the emotions, social adaptation and school performance of middle school students. Adolescence is a period of rapid physical development and a strong desire for social recognition; therefore, physical self-esteem is more unstable at this time and is prone to be influenced by peer norms and school environments [13].
Body-related self-processes are critical to the process of changes in physical self-esteem. Firstly, social comparison and appearance monitoring increase during adolescence. With the development of peer relationships and the formation of group norms, adolescents begin to derive their own evaluations from physical appearance and body shape more often; this makes the stability of physical self-esteem harder to maintain. Secondly, self-objectification and body shame may cause people to regard their own bodies as objects of evaluation, making them more sensitive to external scrutiny and thus prone to avoidant behavior and self-denial. Thirdly, avoidance behavior leads to a lack of functional experience; that is, due to evaluation anxiety, one may reduce participation in physical education classes, sports dancing or group activities, thereby weakening the positive body experience and further damaging physical self-esteem [14].
Physical self-esteem demonstrates high malleability within school environments. Positive body image interventions, particularly group-based activities like sports dance, can effectively reduce appearance comparisons, foster body acceptance, and direct students’ attention toward the functional capabilities of their bodies. These approaches broaden the evaluative context to incorporate support from teachers and peers, thereby strengthening physical self-esteem. Special groups in the education of teenagers, such as children who need special attention in their growth, are at higher risk of losing faith in their physical development due to group discrimination or psychological trauma during interpersonal relationships. At this time, whether a school has an inclusive culture, how teachers respond to feedback, and the acceptance of peers may serve as situational factors.
Mechanistic links between social anxiety and physical self-esteem in school ecologies
Social anxiety and physical self-esteem in adolescence are often mutually reinforced by social evaluation pressure in the school environment. The situation of being watched and judged in school life occurs frequently, such as in the classroom, with classmates, in groups, when dancing, and during physical exercise. Adolescents who are more socially anxious tend to perceive such situations as threats and, therefore, give less attention to task completion. This increases self-monitoring and avoidance behavior, extending the sense of failure and negation to physical self-worth judgment, and manifested as reduced or increased fluctuations in physical self-esteem.
Theoretically, these exploratory pathways operate through several mechanisms. First, social anxiety heightens self-focused attention and sensitivity to visible cues, leading adolescents to misinterpret social evaluation as stemming from physical flaws15. This exacerbates appearance comparison and body shame, undermining physical self-esteem. Second, negative peer experiences, such as appearance-related mockery or exclusion, amplify these threat perceptions. Conversely, a supportive school climate and inclusive peer norms can foster a sense of belonging, potentially buffering these adverse associations. Finally, social anxiety often drives behavioral avoidance of public physical activities (e.g., sports, dance), which deprives adolescents of positive functional body experiences, further diminishing their physical self-esteem and reinforcing the cycle of social withdrawal.
In a school context, examining factors such as appearance comparison, body shame, peer acceptance, and school climate simultaneously, and testing their mediating or moderating effects on the relationship between social anxiety and physical self-esteem through rigorous controlled designs will help advance the description of relationships to action-oriented mechanistic explanations. This way can better meet the practical needs of adolescents with special education for situational support and tiered intervention. Activities such as sports dance that combine body language expression and social interaction are ecologically feasible and can be carried out on a large scale, providing an environment for people to freely express themselves.
School climate, peer norms, and inclusion in Chinese special education
In the context of China’s special education and inclusive education, schools’ climate, peer norms, and inclusive practices are together the main ecological factors that influence the psychological adaptation of adolescents [15]. School climate refers to students’ overall sense of security at school, the quality of teacher-student relations, fairness and reasonableness of the rules, a warm and supportive atmosphere, and a sense of identity. For students with special education needs, these perceptions are particularly acute. Differences in communication styles, learning paces, and behavioral patterns increase their vulnerability to marginalization, misunderstanding, and peer exclusion, fundamentally threatening their psychological belonging at school. Generally speaking, a positive school environment is less likely to cause peer victimization, and will be more conducive to the emotional health and after-school activities participation of students; thus providing stable support for them to avoid negative evaluations.
Peer norms serve as the fundamental basis for promoting inclusive behaviors. In inclusive environments, the ways in which typical peers perceive differences, distribute interaction opportunities, and exhibit tolerance—as well as the presence or absence of mockery and exclusion—determine whether students with special education needs can achieve acceptance and develop practical socializing skills. For instance, in the ballroom dance group activity that requires cooperation, the support peer norm helps enhance their willingness to participate and a sense of identity. When peer groups form norms of respect and support, students are more likely to have a sense of belonging, reduce social avoidance and shame experiences, thus stabilising their self-esteem. On the other hand, appearance-based mockery, exclusion, and school bullying amplify social-evaluation threats; therefore, students tend to attribute their social failures to personal deficiencies and thus form a cycle of avoidance and self-rejection.
Inclusivity is not equivalent to the way of placement. Even if the students are placed in regular classes, due to a lack of classroom adaptation, inadequate peer support, weak anti-bullying mechanisms and ineffective teacher feedback, they may still have limited participation and be excluded implicitly, thereby causing psychological harm. This shows that by adding school climate and peer norms to the research framework, not only is it more reasonably explained why social anxiety is related to physical self-esteem, but also directions for scalable school-based interventions have been opened up. This includes improving teachers’ supportive feedback, establishing norms for peer acceptance, enhancing the coordination among anti-bullying, psychological services and other links, alleviating social evaluation pressure from an ecological perspective to promote the all-round development of students with special needs in inclusive education. A sense of collective activity in Sports Dance, which emphasises cooperation and expression, also holds this position [16].
Methodology
Study design and setting
This study adopted a cluster-level (class-based) controlled pre–post quasi-experimental design. With the school class designated as the unit of assignment, the study examined whether a sports dance group program could reduce social anxiety and improve physical self-esteem among adolescents with special education needs in Chinese school settings, while also exploring the contextual roles of school climate and peer norms. Rather than using a clinical trial framework, the program was integrated into the routine educational environment to enhance ecological validity and practical relevance for school-based mental health support.
The study was carried out in special education schools and in relevant classes of regular schools implementing inclusive education in China. The intervention was embedded in typical school contexts, including classroom interaction, break-time communication, sports dance sessions, and other physical activity settings, so that the observed changes could be interpreted within authentic school ecologies. To reduce disruption to normal school routines and reflect the practical realities of educational implementation, group assignment was determined at the class level based strictly on pre-existing logistical constraints—specifically, the availability of activity spaces and instructors during fixed timetables. Importantly, allocation decisions were independent of student baseline characteristics or teacher preferences. Although this non-random, scheduling-based process carries inherent risks of selection bias, efforts were made to mitigate such implications by prospectively matching classes to ensure comparability across grade level, school type, and placement context.
Three assessment points were established, namely baseline (T0), mid-program (T1), and post-program or follow-up (T2). Baseline assessment was used to obtain the initial levels of primary outcomes and school-context variables, the mid-program assessment captured short-term change, and the final assessment evaluated the overall pattern and maintenance of change. All measures were administered on campus using standardized procedures and the same assessment tools across time points to ensure consistency.
Given the hierarchical structure of students nested within classes and schools, subsequent analyses were conducted using multilevel or mixed-effects models to estimate group, time, and group-by-time effects while accounting for intraclass dependence. Because group allocation was not random, the analyses also adjusted for baseline characteristics and relevant contextual covariates to reduce potential selection bias. The study followed standard ethical procedures for school-based psychological research, including informed consent, privacy protection, and referral arrangements for students who experienced distress during participation.
Participants, eligibility, and recruitment procedures
Participants were recruited from special education schools and from relevant classes in regular schools implementing inclusive education in China. A continuous recruitment strategy was adopted with the assistance of schools in disseminating study information, organizing briefings, and facilitating home–school communication. The research team provided students and their guardians with accessible explanations of the study purpose, procedures, and potential risks. Written informed consent was obtained from guardians, together with assent from participating students, before screening commenced. The overall process of recruitment, screening, group assignment, and follow-up is presented in Fig. 1.
Fig. 1.
Participant flow and follow-up in a school-based controlled pre–post study
The inclusion criteria were as follows: adolescents enrolled in special education or inclusive education programs; students with sufficient comprehension and communication ability to complete questionnaire-based assessments; students able to participate in sports dance group activities and follow-up assessments with routine school support; and students who showed a certain level of social anxiety characteristics or related distress at baseline. This baseline status was identified using standardized screening thresholds and, where necessary, supplementary evaluation by school psychological staff. Exclusion criteria included acute severe psychiatric symptoms that made participation in group activities inappropriate, major physical illness or severe cognitive impairment that substantially affected participation, and recent involvement in psychological treatment highly overlapping with the current program, in order to reduce safety concerns and intervention confounding.
Recruitment and screening were conducted in two stages. The initial stage focused on determining basic eligibility and willingness to participate. The second stage involved on-campus assessment by trained evaluators to confirm inclusion and exclusion criteria and to collect baseline data. Eligible students were then assigned to the sports dance program group or the comparison group according to class scheduling and school administrative arrangements, rather than through random allocation. All participants completed assessments at baseline, mid-program, and post-program or follow-up. To minimize attrition, the research team established a regular reminder and contact system in collaboration with schools and recorded reasons for missing data for subsequent sensitivity analyses.
A total of 226 students were finally included in the study, with 117 assigned to the sports dance program group and 109 to the comparison group. All enrolled participants were included in the primary analysis set, while a per-protocol sample was additionally defined according to attendance and completion of key intervention tasks to supplement evaluation of implementation effects (Fig. 1).
The sample size rationale was based on the class-based quasi-experimental design and the feasibility of recruiting intact classes within routine school schedules. The study aimed to include all eligible students from the participating classes while retaining a sample large enough to support stable mixed-effects estimation of group-by-time effects after adjustment for key demographic, baseline severity, and school-context covariates. The final sample of 226 students (117 in the intervention group and 109 in the comparison group) was considered adequate for the planned controlled pre-post analyses and for sensitivity analyses addressing missing data and attendance.
Assignment method, allocation procedures, and blinding
After completion of the baseline assessment (T0) and confirmation of eligibility, participants were assigned to the sports dance program group or the comparison group at the class level according to school scheduling and administrative arrangements. This procedure was adopted to fit routine school management, reduce disruption to regular teaching activities, and maintain the ecological authenticity of implementation in special education and inclusive school settings. Efforts were made to improve comparability between groups by aligning school type, grade level, and placement context as closely as possible. The overall program structure, assessment schedule, and hypothesized mechanism pathways are illustrated in Fig. 2.
Fig. 2.
Trial design and hypothesized mechanism linking social anxiety to physical self-esteem in students with intellectual disabilities at special education schools
Because the present study did not use random allocation, several procedural strategies were introduced to reduce potential bias and cross-group contamination. The two conditions were implemented within the same general study period, and class schedules were arranged to minimize overlap in activity content, instructor contact, and peer exchange regarding intervention materials. The comparison group received routine school activities or standard school-based mental health education, while the sports dance program group participated in the structured intervention under comparable time arrangements. These procedures were intended to reduce contamination arising from informal communication or exposure to similar content across groups.
Complete blinding of participants and implementers was not feasible because of the visible and participatory nature of the school-based program. However, outcome assessment was conducted by trained evaluators who were not involved in intervention delivery and were asked, as far as practicable, to remain unaware of group status during assessment. Data entry and quality checking were completed through double review procedures, and group codes were masked during part of the statistical analysis process when feasible. Overall, the study sought to balance methodological rigor with school-based feasibility through structured group assignment, contamination control, and standardized outcome assessment procedures.
Intervention, control condition, and fidelity monitoring
This study utilized a school-based group curriculum model integrated into the regular class. The intervention was expected to lower the degree of social anxiety and increase the sense of physical self-esteem of the participants, thereby strengthening their sense of security and involvement in high-exposure social situations through modified self-regulation processes and peer interaction patterns related to physical well-being. To ensure replicability, the intervention was structured as a 6-week program, delivered as one 45- to 60-minute session per week, and integrated into the regular class schedule. The content was based on a standardized manual and executed by uniformly trained school psychology teachers or research team members. The curriculum comprised three sequentially structured modules:
(1) Social Anxiety Regulation (Weeks 1–2): Focused on core skill training, including identifying and re-evaluating negative evaluation concerns, adjusting attentional focus, and reducing safety behaviors through situational exposure and behavioral experiments in sports dance contexts. This facilitated corrective social experiences during group cooperation. (2) Body-Related Self-Processes (Weeks 3–4): Aimed at enhancing physical self-esteem by reducing appearance comparison and body shame. Activities incorporated self-acceptance and emotion regulation exercises, utilizing rhythmic movement to help individuals actively recognize their bodily functions and physical capabilities. (3) School Ecology and Peer Support (Weeks 5–6): Focused on integrating learned skills into the broader school environment. This module guided teachers in providing supportive feedback and promoted inclusive peer norms through collaborative dance tasks, thereby reducing social evaluation pressures in natural settings.
The control group was set to the routine school mental health education or attention-matching courses, and the class schedule was as consistent as possible with that of the intervention group, but key technical components targeting social anxiety or body-related self-processes were excluded to prevent non-specific effects from classroom contact and being noticed.
To ensure the implementation quality of the entire system, this paper establishes a consistency monitoring system. The implementers completed standardised training and scenario-based drills before the start of the study, using course manuals and standardised materials. Instructors took attendance and recorded student participation during each session. Additionally, supervisors regularly spot-checked the class records and conducted follow-ups as required. Any deviation from the protocol was immediately addressed through feedback and correction to maintain compliance. Any deviation from the agreement was immediately given a response via feedback and correction to maintain compliance. The study also recorded the exposure to possible co-intervention and major school activities, which were then used for sensitivity analysis and heterogeneity interpretation.
Measures and operationalization of constructs
The measurement system of this study was constructed based on the primary outcomes, mechanism variables, school ecological variables, and necessary covariates. For the present study, a study-developed questionnaire was used to assess the main constructs of interest. Although validated standardized scales exist for some of these domains, they were deemed unsuitable for direct application due to their length, linguistic complexity, and lack of contextual adaptation for Chinese special education settings. Therefore, a tailored instrument was developed to align closely with the conceptual framework and the specific school-based context. To ensure accessibility and appropriateness across diverse disability categories (e.g., intellectual disabilities, autism spectrum conditions), item phrasing was linguistically simplified. Furthermore, the assessment procedure allowed for standardized accommodations, such as reading items aloud or providing concrete examples, without altering the core constructs. The English-language version of the questionnaire is provided in Supplementary File 1.
Primary outcomes included Social Anxiety and Physical Self-Esteem. Social anxiety was assessed using items covering fear of negative evaluation and social avoidance/distress in social situations. Physical self-esteem was assessed using a multidimensional set of items reflecting body appearance, physical competence, and global body self-worth. To accommodate the cognitive and communicative characteristics of the participants, all items were evaluated on a simplified Likert-type scale equipped with explicit verbal anchors to facilitate comprehension. Both total and subscale scores were aggregated for the analyses, where higher scores consistently indicated greater levels of the respective constructs. Mechanism variables referred to process indicators closely related to the hypothesized theoretical pathways, including appearance comparison, body shame, self-focused attention, and peer acceptance. School ecological variables included perceptions of school climate and peer norms, such as safety, supportiveness, fairness, sense of belonging, and experiences of bullying victimization or exclusion. Covariates included baseline characteristics such as sex, age, category of special education needs, symptom severity, family support, and learning environment, which were used to control for potential confounding and to support stratified analyses.
All items were administered in the same format at baseline, mid-program, and post-program/follow-up to ensure comparability across time points. Questionnaire validation followed two complementary stages. First, before data collection, content validity and accessibility were assessed through iterative review by a panel of special education experts and school psychologists, who evaluated item relevance, construct coverage, developmental appropriateness, linguistic clarity, and the feasibility of standardized accommodations. Pilot testing was then conducted with a small subset of the target population to confirm comprehension, response burden, and consistency of administration; minor wording refinements were made without changing the core constructs. Second, after baseline and repeated assessments were collected, empirical psychometric evaluation was conducted. Internal consistency, item coherence, structural validity, and cross-time stability were examined before the main analyses. The study-developed instrument showed satisfactory reliability across the primary scales and sub-dimensions, with Cronbach’s alpha coefficients ranging from 0.82 to 0.91 at baseline, supporting its use in this Chinese special education population and the comparability of longitudinal measurements across T0, T1, and T2.
Statistical analysis
Statistical analyses were conducted following the intention-to-treat (ITT) principle. To account for the hierarchical structure of the data (students nested within classes and schools), linear mixed-effects models were employed to estimate the group-by-time interaction effects on the primary outcomes across the three assessment points. These models incorporated random intercepts at the class and school levels to adjust for intraclass correlations. Results are reported as marginal mean differences with 95% confidence intervals and standardized effect sizes.
For the mechanism analyses, mediation models were utilized to examine whether appearance comparison, body shame, self-focused attention, and peer acceptance accounted for the association between social anxiety and physical self-esteem. Additionally, interaction terms were introduced into the models to test the moderating effects of school climate and peer norms. Missing data were handled using multiple imputation, accompanied by sensitivity analyses to verify the robustness of the primary findings. Pre-planned subgroup analyses were conducted based on participant sex, special education placement type, and baseline symptom severity. Finally, the False Discovery Rate (FDR) correction was applied to adjust for multiple testing across the mechanism and subgroup analyses.
Results
Baseline balance
The results of the basic balance analysis are shown in Table 1. Overall, the intervention group (n = 117) and the comparison group (n = 109) demonstrated acceptable comparability across key demographic characteristics, special-education placement types, and major disability categories. Furthermore, baseline levels of the primary psychological outcomes (social anxiety and physical self-esteem), mechanism variables, and school ecological indicators (e.g., perceived school climate, bullying victimization) were similarly distributed between the two groups. The generally small standardized differences supported the use of adjusted multilevel models to evaluate intervention-related outcomes.
Table 1.
Baseline characteristics of participants by group in Chinese special-education school settings
| Characteristic | Intervention (n = 117) |
Control (n = 109) |
SMD |
|---|---|---|---|
| Age, years | 13.8 ± 1.5 | 13.7 ± 1.6 | 0.06 |
| Male, n (%) | 74 (63.2) | 66 (60.6) | 0.05 |
| Only child, n (%) | 39 (33.3) | 34 (31.2) | 0.05 |
| Boarding student, n (%) | 28 (23.9) | 24 (22.0) | 0.05 |
| Grade level, n (%) | |||
| Grade 7 | 40 (34.2) | 36 (33.0) | 0.03 |
| Grade 8 | 43 (36.8) | 38 (34.9) | 0.04 |
| Grade 9 | 34 (29.1) | 35 (32.1) | 0.07 |
| Special-education placement, n (%) | |||
| Special school class | 68 (58.1) | 62 (56.9) | 0.02 |
| Inclusive class with resource support | 49 (41.9) | 47 (43.1) | 0.02 |
| Primary disability category, n (%) | |||
| Autism spectrum condition | 32 (27.4) | 28 (25.7) | 0.04 |
| Intellectual disability | 29 (24.8) | 30 (27.5) | 0.06 |
| Hearing impairment | 18 (15.4) | 16 (14.7) | 0.02 |
| Learning disability | 21 (17.9) | 18 (16.5) | 0.04 |
| Other developmental/neurological | 17 (14.5) | 17 (15.6) | 0.03 |
| Comorbid diagnosis reported, n (%) | 26 (22.2) | 23 (21.1) | 0.03 |
| Height, cm | 156.4 ± 8.9 | 155.9 ± 9.2 | 0.06 |
| Weight, kg | 49.8 ± 11.6 | 50.2 ± 12.1 | 0.03 |
| BMI, kg/m² | 20.2 ± 3.7 | 20.5 ± 3.9 | 0.08 |
| Overweight/obesity, n (%) | 19 (16.2) | 20 (18.3) | 0.06 |
| Primary caregiver, n (%) | |||
| Mother | 52 (44.4) | 46 (42.2) | 0.04 |
| Father | 21 (17.9) | 20 (18.3) | 0.01 |
| Grandparent/other | 44 (37.6) | 43 (39.4) | 0.04 |
| Caregiver education (highest), n (%) | |||
| Middle school or below | 51 (43.6) | 50 (45.9) | 0.05 |
| High school/technical | 44 (37.6) | 39 (35.8) | 0.04 |
| College or above | 22 (18.8) | 20 (18.3) | 0.01 |
| Household monthly income, n (%) | |||
| < 5,000 RMB | 43 (36.8) | 41 (37.6) | 0.02 |
| 5,000–10,000 RMB | 52 (44.4) | 46 (42.2) | 0.04 |
| > 10,000 RMB | 22 (18.8) | 22 (20.2) | 0.04 |
| Prior school-based counseling (past 6 months), n (%) | 29 (24.8) | 25 (22.9) | 0.05 |
| Baseline Social Anxiety (total score) | 45.7 ± 10.8 | 46.3 ± 11.2 | 0.06 |
| Baseline Physical Self-Esteem (total score) | 27.9 ± 6.4 | 27.6 ± 6.6 | 0.05 |
| Appearance Comparison (higher = more comparison) | 18.6 ± 4.8 | 18.9 ± 4.9 | 0.06 |
| Body Shame (higher = more shame) | 16.8 ± 4.2 | 17.1 ± 4.3 | 0.07 |
| Self-focused Attention (higher = more self-focus) | 21.4 ± 5.1 | 21.8 ± 5.3 | 0.08 |
| Peer Acceptance (higher = better acceptance) | 30.7 ± 6.9 | 30.3 ± 7.1 | 0.06 |
| Bullying victimization (past month), n (%) | 23 (19.7) | 24 (22.0) | 0.06 |
| School Climate (higher = more supportive) | 32.1 ± 6.0 | 31.7 ± 6.3 | 0.06 |
| Average sleep duration, hours/night | 7.6 ± 1.0 | 7.5 ± 1.1 | 0.07 |
| Physical activity, days/week (≥ 30 min) | 2.8 ± 1.5 | 2.7 ± 1.6 | 0.05 |
Primary outcomes
Table 2 presents the descriptive statistics and group differences for the primary outcome. At baseline, both groups demonstrated similar total scores for social anxiety and physical self-esteem, indicating acceptable group equivalence following the class-based assignment. The Intervention Group showed a continuous decrease in social anxiety over time; however, the Control Group’s reduction was less pronounced. The intergroup difference was not obvious until the mid-term assessment and continued to expand at the final evaluation. The intervention group showed a steady increase in physical self-esteem, while the control group showed only a slight or no change; the difference between groups was greater at the follow-up point. In terms of sub-scale dimensions, the intervention group showed a more pronounced reduction in negative-evaluation-related concerns and avoidance distress of social anxiety. The appearance and ability dimensions of physical self-esteem also showed a trend of improvement, providing support for the simultaneous effect of the intervention on the core psychological process.
Table 2.
Descriptive statistics and between-group differences for social anxiety and physical self-esteem over time
| Outcome (scale) |
Timepoint | Intervention | Control | Between-group Δ (I − C) |
95% CI | P value |
|---|---|---|---|---|---|---|
| Social Anxiety (Total) | T0 | 45.7 ± 10.8 (n = 117) | 46.3 ± 11.2 (n = 109) | −0.6 | −3.4 to 2.2 | 0.68 |
| T1 | 39.9 ± 10.1 (n = 114) | 44.7 ± 10.8 (n = 106) | −4.8 | −7.6 to − 2.0 | 0.001 | |
| T2 | 34.8 ± 9.7 (n = 109) | 43.5 ± 11.3 (n = 102) | −8.7 | −11.6 to − 5.8 | < 0.001 | |
| Fear of Negative Evaluation | T0 | 17.9 ± 4.6 | 18.2 ± 4.7 | −0.3 | −1.5 to 0.9 | 0.61 |
| T1 | 15.5 ± 4.4 | 17.4 ± 4.8 | −1.9 | −3.1 to − 0.7 | 0.002 | |
| T2 | 14.1 ± 4.1 | 17.1 ± 4.9 | −3.0 | −4.2 to − 1.8 | < 0.001 | |
| Social Avoidance and Distress | T0 | 27.8 ± 7.1 | 28.1 ± 7.4 | −0.3 | −2.2 to 1.6 | 0.74 |
| T1 | 24.4 ± 6.8 | 27.3 ± 7.0 | −2.9 | −4.7 to − 1.1 | 0.002 | |
| T2 | 20.7 ± 6.4 | 26.4 ± 7.3 | −5.7 | −7.6 to − 3.8 | < 0.001 | |
| Physical Self-Esteem (Total) | T0 | 27.9 ± 6.4 (n = 117) | 27.6 ± 6.6 (n = 109) | 0.3 | −1.4 to 2.0 | 0.72 |
| T1 | 30.8 ± 6.2 (n = 114) | 28.4 ± 6.5 (n = 106) | 2.4 | 0.7 to 4.1 | 0.006 | |
| T2 | 33.5 ± 6.1 (n = 109) | 29.1 ± 6.7 (n = 102) | 4.4 | 2.7 to 6.1 | < 0.001 | |
| Appearance Self-Esteem | T0 | 9.1 ± 2.5 | 9.0 ± 2.6 | 0.1 | −0.6 to 0.8 | 0.79 |
| T1 | 10.0 ± 2.4 | 9.3 ± 2.5 | 0.7 | 0.0 to 1.4 | 0.044 | |
| T2 | 10.8 ± 2.3 | 9.4 ± 2.6 | 1.4 | 0.7 to 2.1 | < 0.001 | |
| Physical Competence Self-Esteem | T0 | 9.7 ± 2.6 | 9.6 ± 2.7 | 0.1 | −0.6 to 0.8 | 0.8 |
| T1 | 10.9 ± 2.5 | 9.9 ± 2.6 | 1 | 0.3 to 1.7 | 0.004 | |
| T2 | 11.8 ± 2.4 | 10.1 ± 2.7 | 1.7 | 1.0 to 2.4 | < 0.001 | |
| Global Body Self-Worth | T0 | 9.1 ± 2.3 | 9.0 ± 2.4 | 0.1 | −0.5 to 0.7 | 0.71 |
| T1 | 9.9 ± 2.3 | 9.2 ± 2.3 | 0.7 | 0.1 to 1.3 | 0.031 | |
| T2 | 10.9 ± 2.2 | 9.6 ± 2.4 | 1.3 | 0.7 to 1.9 | < 0.001 | |
| Change from Baseline (T2 − T0) | Social Anxiety | −10.9 ± 9.2 (n = 109) | −2.8 ± 7.9 (n = 102) | −8.1 | −10.4 to − 5.8 | < 0.001 |
| Physical Self-Esteem | 5.6 ± 5.1 (n = 109) | 1.5 ± 4.8 (n = 102) | 4.1 | 2.8 to 5.4 | < 0.001 |
The results of the mixed-effects model are shown in Tables 3 and 4. After controlling for age, gender, placement type, and school ecology-related covariates, and adjusting for class or school-level confounding factors, the interaction term of group × time had a significant effect on social anxiety and physical self-esteem. In other words, the improvement after intervention at this point was both a difference in a static state and an accumulation of changes over time. Based on the model, it was further determined that a positive school climate is always associated with a decrease in bullying exposure and an improvement in outcomes; Therefore, the ecological factors of the school provide significant context for explaining the primary outcome.
Table 3.
Primary mixed-effects model test group × time effect on social anxiety (total score)
| Fixed effect | β | SE | 95% CI | P value |
|---|---|---|---|---|
| Intercept (Control at T0) | 46.31 | 1.12 | 44.12 to 48.50 | < 0.001 |
| Intervention group | −0.72 | 1.04 | −2.76 to 1.32 | 0.49 |
| Time T1 | −1.54 | 0.63 | −2.77 to − 0.31 | 0.014 |
| Time T2 | −2.71 | 0.69 | −4.06 to − 1.36 | < 0.001 |
| Group × Time T1 | −3.58 | 0.86 | −5.27 to − 1.89 | < 0.001 |
| Group × Time T2 | −6.24 | 0.94 | −8.09 to − 4.39 | < 0.001 |
| Age (per year) | −0.41 | 0.22 | −0.84 to 0.02 | 0.064 |
| Male (vs. female) | −0.88 | 0.87 | −2.58 to 0.82 | 0.31 |
| Inclusive placement (vs. special school class) | −0.73 | 0.8 | −2.30 to 0.84 | 0.36 |
| Baseline school climate (per point) | −0.26 | 0.07 | −0.40 to − 0.12 | < 0.001 |
| Bullying victimization (yes vs. no) | 2.31 | 0.92 | 0.51 to 4.11 | 0.012 |
Table 4.
Primary mixed-effects model test group × time effect on physical self-esteem (total score)
| Fixed effect | β | SE | 95% CI | P value |
|---|---|---|---|---|
| Intercept (Control at T0) | 27.62 | 0.66 | 26.33 to 28.91 | < 0.001 |
| Intervention group | 0.31 | 0.62 | −0.90 to 1.52 | 0.62 |
| Time T1 | 0.79 | 0.37 | 0.06 to 1.52 | 0.035 |
| Time T2 | 1.34 | 0.41 | 0.54 to 2.14 | 0.001 |
| Group × Time T1 | 1.67 | 0.51 | 0.67 to 2.67 | 0.001 |
| Group × Time T2 | 3.26 | 0.56 | 2.16 to 4.36 | < 0.001 |
| Age (per year) | 0.48 | 0.13 | 0.23 to 0.73 | < 0.001 |
| Male (vs. female) | 0.62 | 0.5 | −0.36 to 1.60 | 0.21 |
| Inclusive placement (vs. special school class) | 0.44 | 0.46 | −0.46 to 1.34 | 0.34 |
| Baseline school climate (per point) | 0.19 | 0.04 | 0.11 to 0.27 | < 0.001 |
| Bullying victimization (yes vs. no) | −1.18 | 0.53 | −2.22 to − 0.14 | 0.026 |
To visually display the changes at the distribution level, Fig. 3 presents a rain cloud plot. The intervention group shows a general trend of decreasing social anxiety intervals over time and is more concentrated and stable. The distribution of physical self-esteem shows a trend of increasing scores with reduced dispersion, indicating that the intervention has raised the mean while also reducing the extreme differences among individuals. In contrast, the control group shows little change in distribution and still has many individuals with low physical self-esteem or high social anxiety at the end of the observation, which demonstrates the advantage of the intervention at the population level.
Fig. 3.
Raincloud plots of social anxiety and physical self-esteem over time across groups
Individual and group-level trajectory analyses further supported these temporal trends, demonstrating a steeper and more consistent improvement slope for the intervention group over time, indicative of the cumulative benefits of the program (see Supplementary Figure S1 for detailed trajectory plots).
Meaningful change and response profiles
To further evaluate the practical significance of the program beyond mean-level differences, this study examined predefined response rates, Reliable Change Index (RCI) categories, and between-group differences in absolute and relative response probability. The results are presented in Table 5. Overall, the intervention group showed higher response rates on both major outcomes, namely reduced social anxiety and increased physical self-esteem, while the proportion of participants showing no meaningful change was lower than that in the comparison group. These findings suggest that the sports dance program was more likely to produce individually noticeable and educationally meaningful improvement.
Table 5.
Clinically and educationally meaningful change: response rates and reliable change profiles
| Outcome / Criterion | Intervention (n = 117) |
Control (n = 109) |
Risk Difference (I − C) |
Relative Risk |
|---|---|---|---|---|
| Social Anxiety (Total) | ||||
| Response (≥ 8-point decrease by T2) | 64 (54.7%) | 34 (31.2%) | 23.50% | 1.75 |
| Reliable improvement (RCI ≤ − 1.96) | 52 (44.4%) | 26 (23.9%) | 20.50% | 1.86 |
| No meaningful change | 44 (37.6%) | 56 (51.4%) | −13.8% | 0.73 |
| Reliable worsening (RCI ≥ + 1.96) | 6 (5.1%) | 9 (8.3%) | −3.2% | 0.61 |
| Physical Self-Esteem (Total) | ||||
| Response (≥ 4-point increase by T2) | 71 (60.7%) | 40 (36.7%) | 24.00% | 1.65 |
| Reliable improvement (RCI ≥ + 1.96) | 49 (41.9%) | 24 (22.0%) | 19.90% | 1.9 |
| No meaningful change | 42 (35.9%) | 55 (50.5%) | −14.6% | 0.71 |
| Reliable worsening (RCI ≤ − 1.96) | 5 (4.3%) | 7 (6.4%) | −2.1% | 0.67 |
| Composite responder (both outcomes improved) | 45 (38.5%) | 20 (18.3%) | 20.20% | 2.1 |
| High-risk subgroup responder (baseline high anxiety and low PSE) | 28/54 (51.9%) | 13/49 (26.5%) | 25.40% | 1.96 |
The RCI-based results were consistent with this pattern. The intervention group had a higher proportion of participants showing reliable improvement, whereas reliable worsening was less frequent or remained at a similarly low level across groups. This indicates that the observed gains were more likely to exceed measurement error and normal fluctuation. In addition, the proportion of composite responders, defined as participants who improved in both social anxiety and physical self-esteem, was higher in the intervention group. A similar advantage was observed in the subgroup with high baseline anxiety and low physical self-esteem, suggesting that students with greater initial vulnerability may derive greater benefit from the program. Taken together, the response profiles and risk differences support the educational relevance and potential scalability of the intervention in school settings.
Pathway analyses
Based on the model test results (Table 6), social anxiety is more likely to be associated with physical self-esteem via four pathways: appearance comparison, body shame, self-focused attention and peer acceptance. Among them, body shame and appearance comparison emerged as significant indirect pathways. That is to say, higher social anxiety was associated with a greater tendency for body shame and appearance comparison, which in turn correlated with lower physical self-esteem. On the other hand, peer acceptance exhibited a protective pattern, demonstrating a positive correlation with physical self-esteem and tempering the negative association between social anxiety and physical self-esteem within the tested model architecture.
Table 6.
Mechanism models: mediation and moderation of the social anxiety → physical self-esteem path via school climate and peer factors
| Model component | Estimate | SE | 95% CI | P value |
|---|---|---|---|---|
| Total effect (c): Social Anxiety → Physical Self-Esteem | −0.18 | 0.03 | −0.24 to −0.12 | <0.001 |
| Direct effect (c′) controlling for mediators | −0.10 | 0.03 | −0.16 to −0.04 | 0.001 |
| Total indirect effect (Σa×b) | −0.08 | 0.02 | −0.12 to −0.05 | <0.001 |
| Indirect via Appearance Comparison | −0.02 | 0.01 | −0.04 to −0.01 | 0.006 |
| Indirect via Body Shame | −0.03 | 0.01 | −0.05 to −0.02 | <0.001 |
| Indirect via Self-focused Attention | −0.02 | 0.01 | −0.04 to −0.01 | 0.011 |
| Indirect via Peer Acceptance | −0.01 | 0.01 | −0.03 to −0.00 | 0.038 |
| Moderation: Social Anxiety × School Climate | 0.04 | 0.01 | 0.01 to 0.06 | 0.004 |
| Moderation: Social Anxiety × Inclusive Peer Norms | 0.03 | 0.01 | 0.01 to 0.05 | 0.009 |
| Moderation: Social Anxiety × Bullying Victimization | −0.06 | 0.02 | −0.10 to −0.02 | 0.003 |
| Conditional effect of Social Anxiety at High School Climate (+1 SD) | −0.13 | 0.03 | −0.19 to −0.07 | <0.001 |
| Conditional effect of Social Anxiety at Low School Climate (−1 SD) | −0.22 | 0.04 | −0.30 to −0.14 | <0.001 |
Among the moderating effects, school climate and peer norms have played particularly significant roles. The positive dimensions of school climate (supportive atmosphere, perceived fairness) have a strong moderating effect on the association between social anxiety and physical self-esteem. Specifically, a supportive school climate significantly moderated the relation, such that the negative statistical association between social anxiety and physical self-esteem was less pronounced for students perceiving a highly supportive environment. Peer norms (such as peer acceptance and bullying exposure) also significantly moderate this pathway. Conversely, exclusionary peer norms or bullying experiences acted as risk moderators, statistically magnifying the negative link between social anxiety and physical self-esteem (Table 6).
The exploratory pathways and complex cross-domain associations among school climate, peer factors, social anxiety, and physical self-esteem are visually summarized in Supplementary Figure S2 (Chord Diagram).
Subgroup and transition analyses
Subgroup differences in intervention-related outcomes were explored across special education categories, sex, and baseline severity. The program was associated with the most pronounced improvements among students with higher baseline social anxiety and lower physical self-esteem, as well as those within the autism spectrum and intellectual disability categories (see Supplementary Table S1 for detailed subgroup estimates). Furthermore, state transition analyses confirmed a substantial shift toward lower anxiety and higher physical self-esteem categories in the intervention group compared to the control group from baseline to follow-up (see Supplementary Figure S3 for Sankey transition visualizations).
Translational output
In terms of the translational application transformation of these findings, we developed an exploratory, internally derived prediction model to help school staff identify students who may be at risk of low physical self-esteem following the intervention. Figure 4 is a column chart showing the baseline social anxiety and school context factors that can be used to predict post-intervention risks of low physical self-esteem. According to the results, several predictive factors (such as social anxiety, school atmosphere and peer relationships) will be linked to students’ self-esteem, and some intuitive indicators for identifying which groups of students need more care will also be provided.
Fig. 4.
Nomogram for predicting low physical self-esteem after intervention based on baseline social anxiety and school-context factors
Table 7 summarises the feature weights, model discrimination (AUC) and calibration for this exploratory model. To mitigate the risk of overfitting and assess the stability of the model’s performance within the current sample, robust internal validation was conducted using a bootstrapping procedure with 1,000 resamples. This approach yielded optimism-adjusted performance metrics, thereby strengthening confidence in the predictive utility of the model prior to future external validation. The optimism-adjusted AUC of 0.79 indicates an acceptable, though not definitive, discriminative capacity for identifying students at risk of low physical self-esteem. While the Brier score suggests adequate calibration within the current dataset, these metrics collectively position the model as an exploratory heuristic tool rather than a standalone diagnostic instrument. Its primary value lies in demonstrating the conceptual feasibility of using baseline ecological factors to inform tiered support, rather than immediately directing the allocation of school resources.
Table 7.
Prediction model of low physical self-esteem at follow-up: feature weights, discrimination and calibration
| Predictor | Weight (β) |
Standard Error (SE) |
Odds Ratio (OR) |
95% Confidence Interval (CI) |
P Value |
|---|---|---|---|---|---|
| Baseline Social Anxiety | 0.12 | 0.03 | 1.13 | 1.06 to 1.21 | < 0.001 |
| School Climate (Supportive) | −0.45 | 0.12 | 0.64 | 0.48 to 0.85 | < 0.001 |
| Peer Acceptance | −0.28 | 0.1 | 0.76 | 0.60 to 0.95 | 0.015 |
| Bullying Victimization (Yes vs. No) | 0.32 | 0.14 | 1.38 | 1.05 to 1.83 | 0.022 |
| Model Performance | |||||
| Area Under the Curve (AUC) | 0.79 | 0.75 to 0.83 | < 0.001 | ||
| Calibration (Brier Score) | 0.22 | 0.19 to 0.25 | |||
| Nonsignificant Calibration Test | 0.37 |
Discussion
The present study systematically examined the outcomes associated with a school-based sports dance program regarding social anxiety and physical self-esteem among adolescents with special education needs, and further explored the underlying psychological and ecological pathways [17]. The findings showed that participation in the sports dance program was associated with reduced social anxiety and improved physical self-esteem, and that the two outcomes changed in a coordinated direction. This pattern is consistent with previous evidence suggesting that improvements in body-related self-evaluation may buffer social distress and strengthen adolescents’ confidence in interpersonal contexts [18]. In the context of special education, where students may be more vulnerable to negative evaluation and social withdrawal, sports dance may provide a structured and low-barrier form of embodied participation, thereby helping students gradually rebuild confidence in both social interaction and bodily self-worth. These findings suggest that sports dance is not only a feasible school-based activity, but also a potentially valuable psychological support strategy in educational settings.
The mechanism analyses further indicated that body-related self-processes, particularly appearance comparison, body shame, and self-focused attention, partially explained the association between social anxiety and physical self-esteem. However, given the limitations of the quasi-experimental design, coupled with the reliance on simultaneous measurement and self-report data, these pathway findings should be interpreted with caution. They are strictly exploratory and hypothesis-generating in nature, and do not establish definitive causal pathways. In other words, higher social anxiety was associated with a stronger tendency toward appearance-based monitoring and shame-related self-evaluation, which concurrently covaried with lower physical self-esteem within the specified path framework. This is broadly in line with prior research showing that body-image-related cognitive and emotional processes play an important role in adolescent mental health and are closely linked to self-evaluative anxiety [19]. The present findings extend this line of work by suggesting that sports dance may operate not only through social exposure and cooperative participation, but also through reshaping how adolescents attend to, interpret, and experience their bodies. Through movement expression, partner coordination, and repeated participation in manageable public performance contexts, students may become less focused on negative appearance-based comparison and more aware of bodily competence and functional value. This may explain why the program was associated with simultaneous improvement in social anxiety and physical self-esteem.
The present study also highlights the importance of school climate and peer norms in shaping intervention-related outcomes. A more supportive school environment and stronger peer acceptance appeared to buffer the negative pathway from social anxiety to physical self-esteem, whereas bullying exposure intensified vulnerability. However, it must be emphasized that these moderation effects represent model-based associations; without direct experimental control over these contextual factors, they should not be interpreted as definitive causal buffering. These findings are consistent with previous studies showing that positive school climate and social support are closely related to better adolescent mental health outcomes [20]. In the case of sports dance, the interpersonal and cooperative nature of the activity may be particularly meaningful in inclusive school contexts. Because the activity relies on rhythm, non-verbal coordination, mutual responsiveness, and shared performance experience, it may offer a relatively safe environment for students with special education needs to engage with others and receive affirming social feedback. At the same time, heterogeneity analyses suggested that the program may be especially beneficial for students with greater baseline vulnerability, indicating the potential value of tiered implementation in school mental health practice. As a preliminary proof-of-concept, the exploratory prediction model illustrates how baseline psychological and ecological profiles might eventually inform targeted screening. However, its current practical utility must not be overstated. Derived and internally validated within a specific, non-randomized school sample, the nomogram is inherently susceptible to context-specific overfitting. Robust external validation across diverse school districts and broader disability categories is strictly required to confirm its transportability; until then, it cannot be responsibly adopted for real-world educational decision-making.
Despite these strengths, several limitations should be acknowledged. First, group assignment in the present study was based on school scheduling and administrative arrangements rather than random allocation. Although baseline comparability was acceptable and statistical adjustment was applied, residual confounding cannot be fully excluded. Therefore, the findings should be interpreted as evidence from a school-based quasi-experimental design rather than definitive causal proof. Second, the study relied primarily on study-developed, questionnaire-based measures. While content validity was prioritized during the development phase, the instrument lacked rigorous psychometric evaluation, such as confirmatory factor analysis to definitively establish structural validity. As a result, the findings may be more susceptible to self-report bias and situational factors, highlighting the pressing need for future research to either thoroughly validate this tailored instrument or incorporate standardized, independently validated measures for this specific population. Third, because the intervention was implemented in real school settings, variability in class atmosphere, teacher support, and participation quality may also have affected outcomes. Future research should build on the present findings by conducting randomized controlled trials where feasible, ideally across multiple schools and regions, in order to strengthen causal inference. Furthermore, to establish broader generalizability, it is essential for subsequent studies to evaluate the program’s cross-cultural applicability beyond the Chinese educational system and to systematically investigate its efficacy across diverse categories of special education needs. Longer follow-up periods and more diverse outcome indicators, including behavioral and observational measures, would also help clarify the durability and practical significance of sports dance interventions in special education. Additionally, future studies should prioritize the rigorous psychometric validation of the measurement instruments within special education populations to ensure the robustness and generalizability of the findings.
Conclusion
This study, conducted in the context of special education schools in China, found that a school-based sports dance program was associated with a significant reduction in adolescents’ social anxiety and a concurrent improvement in physical self-esteem, with benefits becoming more apparent over time. Mechanism analyses suggested that body-related self-processes, including appearance comparison, body shame, and self-focused attention, played important roles in this relationship. In addition, supportive school climate and inclusive peer norms appeared to buffer the negative impact of social anxiety on physical self-esteem, whereas bullying exposure increased risk. Heterogeneity analyses indicated that students with greater baseline vulnerability benefited more from the program, and the prediction model further provided a practical tool for school-based screening and resource allocation. As an ecologically grounded, highly participatory, and school-feasible approach that integrates bodily movement with social engagement, sports dance represents a promising pathway for promoting mental health in adolescents with special education needs. However, given the quasi-experimental nature of this study, these findings must be interpreted with caution. Further rigorous randomized controlled trials are required to confirm causal efficacy and fully establish the broader scalability of the intervention across diverse educational contexts.
Supplementary Information
Acknowledgements
Not applicable.
Authors’ contributions
Peng Huang and Wenting Dong contributed equally to this work. Peng Huang and Fang Li conceived and designed the study. Peng Huang, Wenting Dong, and Xinyue Zhang were responsible for participant recruitment, intervention implementation, and data collection. Peng Huang and Wenting Dong performed the statistical analyses and interpreted the data. Peng Huang drafted the manuscript. Wenting Dong and Xinyue Zhang critically revised the manuscript for important intellectual content. Fang Li supervised the study, reviewed the manuscript, and approved the final version for submission. All authors read and approved the final manuscript.
Funding
The authors declare that no funding was received for this study.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available because the study involved minors and contained potentially identifiable sensitive information. De-identified data are available from the corresponding author on reasonable request, subject to approval by the relevant ethics requirements and institutional regulations.
Declarations
Ethics approval and consent to participate
This study was reviewed and approved by the Medical Ethics Committee of Wuhan Sports University, Wuhan, China (Approval No. 2025147). All procedures involving human participants were conducted in accordance with the Declaration of Helsinki and relevant institutional ethical standards. Written informed consent was obtained from the parents or legal guardians of all participants prior to enrolment, and assent was obtained from the adolescent participants. Participation was voluntary, and participants were informed that they could withdraw from the study at any time without any adverse consequences.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Peng Huang, Wenting Dong, Xinyue Zhang and Fang Li contributed equally to this work.
References
- 1.Walder N, Berger T, Hürzeler D, McDougal E, Edbrook-Childs J, Schmidt SJ. Prevention and treatment of social anxiety disorder in adolescents: mixed method randomised controlled trial of the guided online intervention SOPHIE. Sci Rep. 2025;15(1):25141. 10.1038/s41598-025-10193-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Vacca M, Cerolini S, Zegretti A, Zagaria A, Mocini E, Lombardo C. A School-Based Intervention Program to Reduce Weight Stigma in Adolescents. Children. 2025;12(9):1208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mahon C, Hamburger D, Yager Z, Webb JB, Brown G, Fitzgerald A. Outcomes of a cluster randomised control trial of Be Real’s BodyKind; A universal body image program for adolescents in Ireland. Body Image. 2025;54:101905. [DOI] [PubMed] [Google Scholar]
- 4.Lei J, Leigh E, Charman T, Russell A, Hollocks MJ. Exploring the association between social camouflaging and self- versus caregiver-report discrepancies in anxiety and depressive symptoms in autistic and non-autistic socially anxious adolescents. Autism: Int J Res Pract. 2024;28(10):2657–74. 10.1177/13623613241238251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Thiele-Swift HN, Dorstyn DS. Anxiety Prevalence in Youth with Autism: A Systematic Review and Meta-analysis of Methodological and Sample Moderators. Rev J Autism Dev Disord. 2024. 10.1007/s40489-023-00427-w. [DOI] [Google Scholar]
- 6.Park EY, Basel. Switzerland), 15(5):655. 10.3390/bs15050655. [DOI]
- 7.Huang Z, Chen W, Peng B, Yan H. Exploring the mechanisms linking perceived school climate to negative emotions in adolescents: The mediating roles of social avoidance and distress, and psychological resilience, with physical exercise level as a moderator. PLoS ONE. 2025;20(5):e0321726. 10.1371/journal.pone.0321726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bonfanti RC, Melchiori F, Teti A, Albano G, Raffard S, Rodgers R, Coco GL. The association between social comparison in social media, body image concerns and eating disorder symptoms: A systematic review and meta-analysis. Body image. 2025;52:101841. [DOI] [PubMed] [Google Scholar]
- 9.Petersen JM, Davis CH, Renshaw TL, Levin ME, Twohig MP. School-based acceptance and commitment therapy for adolescents with anxiety: a pilot trial. Cogn Behav Pract. 2023;30(3):436–52. [Google Scholar]
- 10.Coyne SM, Van Alfen M, James P, Hirschi T, Densley RL, Christensen-Duerden C, Olsen J. Who is Most at Risk for Body Esteem Problems After Being on Social Media? A Differential Susceptibility Approach with Adolescents. Computers Hum Behav Rep, 2025:100737.
- 11.Matsumoto K, Hamatani S, Shiga K, Iiboshi K, Kasai M, Kimura Y, Nakamura M. Effectiveness of Unguided Internet-Based Cognitive Behavioral Therapy for Subthreshold Social Anxiety Disorder in Adolescents and Young Adults: Multicenter Randomized Controlled Trial. JMIR Pediatr Parent. 2024;7:e55786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zelihić D, van Dalen M, Kling J, Pripp AH, Nordgreen T, Kvalem IL, Okkerse JM. Reducing social anxiety in adolescents distressed by a visible difference: Results from a randomised control trial of a web-based intervention. Body Image. 2022;40:295–309. [DOI] [PubMed] [Google Scholar]
- 13.Ajibewa TA, Kershaw KN, Carnethon MR, Heard-Garris NJ, Beach LB, Allen NB. Peer bullying victimization, psychological distress, and the protective role of school connectedness among adolescents. BMC Public Health. 2025;25(1):2763. 10.1186/s12889-025-24002-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wang X, Wang Y, Yang S, Li Z, Fu C. Social appearance anxiety and body checking behavior in the relationship between social media addiction and depressive symptoms among college students. Front Psychiatry. 2025;16:1608527. 10.3389/fpsyt.2025.1608527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Yang T, Xiao H, Fan X, Zeng W. Exploring the effects of physical exercise on inferiority feeling in children and adolescents with disabilities: a test of chain mediated effects of self-depletion and self-efficacy. Front Psychol. 2023;14:1212371. 10.3389/fpsyg.2023.1212371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wright B, Blackwell JE, Bell KJ, Teige C, Mandefield L, Wang HI, Welch C, Scantlebury A, Watson J, McMillan D, Standley E, Attwell L, Carrick H, Taylor A, Taylor O, Hodkinson R, Edwards H, Pearson H, Parrott S, Marshall D, Hewitt C. Autism Spectrum Social Stories in Schools Trial 2 (ASSSIST-2): a pragmatic randomised controlled trial of the Social Stories™ intervention to address the social and emotional health of autistic children in UK primary schools. Child Adolesc Mental Health. 2025;30(1):4–12. 10.1111/camh.12740. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bullard N, Rupani N, Renteria JA, Miller J, Kaliebe KE. Yoga as a therapeutic intervention for body image in the pediatric population: a systematic review. J Eat disorders. 2025;13(1):222. 10.1186/s40337-025-01386-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Peng B, Chen W, Wang H, Yu T. From physical self-esteem to sports participation: The mediating role of exercise motivation and social support in adolescents. PLoS ONE, 2025; 20(5): e0324588. [DOI] [PMC free article] [PubMed]
- 19.Tang C, Zhang X, Zhou C, Li K, Cai Y. The effects of weight self-stigma on psychological distress in adolescents: the chain-mediated roles of fear of negative appearance evaluation and social appearance anxiety. Front Psychol. 2025;16:1619694. 10.3389/fpsyg.2025.1619694. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Jiang L, He Q, Chen C. Beyond the positive glow: how school climate inadvertently challenges mental health by modifying the relationship between BMI and depression in adolescents. Front Psychol. 2025;16:1663969. 10.3389/fpsyg.2025.1663969. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available because the study involved minors and contained potentially identifiable sensitive information. De-identified data are available from the corresponding author on reasonable request, subject to approval by the relevant ethics requirements and institutional regulations.