Introduction

To realize the developmental objectives of the “Great Health” era, the Healthy China 2030 Planning Outline advocates for "the maintenance and protection of public health across all stages of life. "Under the “health priority” development paradigm, advancing health interventions becomes the most rational and cost-effective strategy to address the “window period” of adolescent sub-health. The physical fitness of adolescents stands as a cornerstone in building a strong sports nation and reflects the overall strength of the country, bearing significant implications for the nation’s future and the advancement of its ethnic groups. Consequently, it is a collective responsibility and an undeniable duty for society to ensure a robust and healthy reserve of young people’s physical fitness1. However, due to the absence, misalignment, and displacement of physical education within families, communities, and schools, children and adolescents have not developed a strong sports awareness or engaged in regular exercise2. This has led to insufficient physical activity levels and a continuous decline in physical fitness among Chinese children and adolescents, consequently creating a situation where the development of sports for children and adolescents remains lagging and there exists significant health inequality3.

In 2021, the State Council’s General Office emphasized the importance of integrating classroom instruction with extracurricular activities and engaging families, communities, and social organizations4. However, the lack of operational guidance in the document resulted in practical discrepancies. Research indicates that children and adolescents exhibit insufficient physical activity and sports awareness, with their physical fitness and health being less than optimal. This is primarily due to their limited independent exercise behaviors, which are influenced by both internal and external factors5. Given China’s current economic and social context, the primary factor influencing the level of physical activity is the national fitness environment6. The environment acts as a catalyst for action by reinforcing existing behavioral intentions. The immediate surroundings directly affect the exercise behavior of children and adolescents. Improvements in their environment lead to increased exercise behavior7. Consequently, the environment serves as a driving force for the development of youth physical fitness activities.

The theory of human adaptation to the environment posits that a match between humans and their surroundings enhances individual behavioral efficiency8. Focusing on the sports environment as a means to address the physical activity patterns of children and adolescents is a focal point in the realm of physical health, as well as an effective strategy to tackle health disparities among children and adolescents. Existing research has analyzed youth physical activity patterns not only through individual variables such as community environment, family capital, and school facilities, but also by exploring the sports environment from model construction, collaborative governance, and system innovation, offering theoretical and practical insights for further research. However, given that various sub-environments often exist separately yet coexist, each influencing children and adolescents based on their unique characteristics, a single-perspective approach can lead to isolated and incomplete outcomes. Moreover, discussions regarding the collaborative mechanisms among family, community, and school that influence youth physical activity are insufficient, necessitating a comprehensive examination of these factors. Drawing on data from the China Education Panel Study (CEPS), this paper employs the Hierarchical Linear Model (HLM) to analyze the influence of familial, communal, and academic factors on adolescent physical activity, and explores potential mechanisms at play. Compared with previous studies, innovations are mainly made in the following aspects: (1) Starting from the dual structure of the internal and external education system, this paper explores the influence of family, community, and school on young people’s physical exercise behavior, not only examining the individual impact of environmental factors but also testing the moderating effects of different levels through hierarchical analysis; (2) This study utilizes large-scale longitudinal tracking data and employs a nested model to explain synergistic effects. The data is highly representative and comprehensive, enabling a more accurate reflection of the mechanisms at play within schools, communities, and families. Consequently, this study yields more reliable conclusions. It not only enriches the research on the relationship between young people’s physical exercise behavior and the environment, but also explores the mechanisms influencing young people’s physical exercise behavior from the perspectives of family, community, and school, thereby providing theoretical support for addressing disparities in young people’s physical health.

Literature review and research hypotheses

Bronfenbrenner was the pioneer in proposing the ecological systems theory, which underscores that individual growth and development are inherently influenced by environmental factors9. He posits that the environment possesses a nested structure, indicating that an individual’s development is situated within various levels of environmental systems. The theory categorizes an individual’s environmental system into four distinct levels: microsystem, mesosystem, exosystem, and macrosystem (see Fig. 1). The microsystem pertains to the immediate context of individual activities and interactions, representing the closest system with which individuals engage during their developmental process. For adolescents, their families, communities, and schools constitute the microsystems integral to their development10. Bronfenbrenner emphasized that relationships within the system are bidirectional, indicating that while adults and the environment shape the responses of children and adolescents, the key physiological attributes, personality, and abilities of these individuals also influence adult behavior and the environment. Ecosystem theory posits that positive interactions between microsystems can foster optimal individual development. Consequently, it is crucial to encourage these interactions to yield a lasting and positive influence on adolescents. The introduction of microsystems like family, community, and school offers theoretical tools for analyzing the behavioral mechanisms of adolescent physical exercise. This article will examine studies on adolescent physical exercise behavior from these three perspectives.

Fig. 1
figure 1

Microsystems-environment.

Full size image

Family microsystem: family environment and adolescent physical exercise behavior

The family, as the primary agent for early socialization of adolescents, is the starting point for their physical activity behaviors, playing an irreplaceable role in guiding their health. Not only can it guide and support adolescents’ physical exercise behaviors from the levels of sports cognition and participation11, but more importantly, it exerts a lasting impact on their physical exercise behaviors through intergenerational intervention12. The family has been proven to be a positive predictor of physical activity in children and adolescents13, with parental encouragement, support, role models, and attitudes positively correlating with adolescents’ participation in sports activities. The family environment is the most influential microenvironment in the development of adolescents, alongside school, providing an environment that stimulates physical development, including being taught and learning basic skills14. Sports become an extension of family life and a part of family education. The family environment encompasses a variety of factors, including a family’s material conditions and social status, as well as the relationships, behavioral patterns, and emotional exchanges among family members15. In existing research, parents’ levels of education, the family’s economic status, and social class are often considered key indicators in measuring the family environment16. This study aims to explore the correlation between several key dimensions of the family environment, such as the harmony of family relationships, the educational levels of parents, and parent–child relationship, with adolescents’ participation in physical exercise. The theory of family social capital, as a vital tool for analyzing the family microsystem, posits parent–child relationships within the family as indicators for measuring family social capital, unveiling the essence of family social capital both conceptually and methodologically. It posits that the “strong connections” formed within parent–child relationships are the primary indicators for measuring family social capital17, whereas the integrity of the family structure merely represents a prerequisite for the construction of parent–child relationships. The research results show that the fundamental reason for the lack of family social capital is the neglect and structural absence of parents in the lives and education of their children. At the same time, Marjoribanks et al. defined the concept as the strength of the relationship between parents and children in the family based on this research, and used parental expectations and investment as measurement indicators15. Therefore, this study takes parent–child relationship as a factor to measure family social capital, and uses adolescents’ self-evaluation of parent–child relationship as a measurement method to detect its impact on adolescents’ physical fitness behavior. Some scholars have also synthesized previous research results and taken parental educational expectations as an indicator to measure family social capital18. This study believes that educational expectations, as the most important expectations of parents, have been embedded in family social capital and become an important measurement indicator. Therefore, parental educational expectations are taken as an important indicator to measure family environment in this study. Family social capital significantly influences adolescents’ engagement in physical activities by offering support, information, and opportunities. Research indicates that adolescents from families with larger social networks tend to exercise more frequently, and those with stronger parent–child relationships also show higher rates of physical activity19. The family sports atmosphere is particularly influential, with a high correlation to adolescents’ participation in sports, suggesting that multiple facets of family social capital positively impact adolescents’ involvement in sports activities20.

Meanwhile, the educational level of parents positively and significantly influences the participation of adolescents in sports, which is also an important reason for the stratification of physical exercise21. The physical exercise behavior of adolescents is usually closely related to their parents’ occupation and educational level, and has a positive and positive impact22. Parents with higher educational levels have a deeper understanding of the importance of physical fitness and the health of their children, and play an important role in shaping their children’s attitudes, cognition, and value systems. The higher the educational level of parents, the better the accessibility and affordability of sports facilities, mainly manifested in providing positive experiences and support, and cultivating children’s lifelong interest in sports participation. Because when parents realize the importance of athletic ability, they will consciously increase outdoor time to cultivate exercise habits23. Therefore, children whose parents have a high level of education are more likely to participate in sports activities in their leisure time. Moreover, well-educated parents are more likely to encourage their children to participate in organized sports activities, while parents with lower educational levels, in addition to affecting the family environment that adolescents are exposed to, are also constrained by unhealthy family beliefs24. The research results show that the educational level of parents affects the social and physical characteristics of the family environment, and the influence of the family environment on exercise behavior tendency is its key root cause, with a continuous impact effect. Therefore, this study selects the highest educational level of parents as an important indicator to measure the family environment. Based on the existing research results, this paper proposes the following hypotheses:

  • H1: Family environment has a significant positive impact on adolescent physical exercise behavior.

  • H1a: Parent–child relationship has a significant positive impact on adolescent physical exercise behavior.

  • H1b: Educational expectations have a significant positive impact on adolescent physical exercise behavior.

  • H1c: The highest educational level of parents has a significant positive impact on the physical exercise behavior of adolescents.

Community microsystem: community environment and adolescent physical exercise behavior

The community serves as an extension of the family environment. As the fundamental unit of the city, it is the environment that children and adolescents interact with the most intimately, and it is also the primary venue for daily activities25. The community offers diverse support and assistance in material and spiritual aspects, which, through its geographical, social mobilization, and resource integration strengths, positively influences the development of health behaviors among teenagers, thereby serving as an inexhaustible driving force for their physical fitness activities26. Community environment refers to the integrated manifestation of natural and social conditions within and surrounding a community. It encompasses various aspects, including the natural environment, physical environment, social environment, and cultural environment, as well as safety considerations. These factors collectively influence the quality of life for residents. A positive community environment is crucial for promoting resident health, enhancing community cohesion, and increasing resident satisfaction. The flexibility and convenience of the community offer a platform for communication and relaxation in fitness activities, enhancing the accessibility to and participation in sports fitness activities among teenagers, and compensating for the deficiencies in family physical education. Research has indicated that the community sports environment positively impacts the physical exercise behaviors of teenagers7. As the most frequently used daily activity venues, the size of community fitness facilities positively influences the guidance of physical fitness behaviors among children and adolescents27. Within the community spaces, interpersonal networks facilitate mutual support and services in terms of materials, resources, and psychology. Community activity venues reduce the distance between families and sports facilities, and the unstructured sports activities within the community enhance the participation rate of teenagers. Studies suggest that the community sports environment, including sports equipment and safety, enhances the value of teenagers’ sports activities and subtly influences their levels of activity during after-school and other non-school periods28. However, participation in non-organized sports activities within the community may be influenced by socioeconomic status. The likelihood of children and adolescents participating in non-organized sports activities is related to their surrounding environment, such as family, community, and level of safety. Communities with higher socioeconomic status typically provide more sports facilities, while low-income communities face limitations in the affordability and accessibility of sports facilities. Therefore, as a gathering place of culture, the potential cultural atmosphere, humanistic support, educational support, and other factors in the community play a guiding role in supporting the behavior of children and adolescents29. There is a complex dynamic relationship between the community environment and the physical exercise behavior of adolescents, and the construction of the humanistic environment in the community is particularly important for the development of healthy behaviors among children and adolescents. Firstly, communities with high levels of safety are positively correlated with spontaneous active play. The rules for active play in the community are more stringent, and a good community environment provides safety guarantees for active participation. Secondly, the availability of community recreational facilities is positively correlated with the participation of children and adolescents in sports, which has a positive impact on their participation in sports30. It can be seen that the community environment that affects the physical fitness behavior of adolescents should also consider the social status, public security, community development, and humanistic environment of community residents. Therefore, this paper proposes the following hypotheses:

  • H2: The community environment has a significant positive impact on the physical exercise behavior of adolescents.

School microsystem: school environment and adolescent physical exercise behavior

The school environment encompasses the overall context in which students learn and engage in activities within the campus. Broadly speaking, the school environment refers to all elements that influence student growth and development, which can be broadly categorized into two major types: physical and psychological environments. The physical environment primarily includes aspects such as campus planning, the allocation of classrooms and laboratories, and the provision of sports facilities. The psychological environment, on the other hand, encompasses non-material factors such as the political atmosphere, academic ambiance, and the ethos and spirit of the school. The school environment has a subtle yet profound impact on the mental and physical development of students. Reproduction theorists believe that schools are places that produce class inequality. Educational sociological research has found that schools have a significant impact on the production of unequal educational outcomes. As a crucial spatial ___domain within microsystems, schools have a strong penetrative influence on the formation of adolescent physical exercise behavior, with an irreplaceable authority. Differences in school atmosphere and conditions may exert an influence independent of the family environment31. The school environment will serve as an inexhaustible source of motivation and power for enhancing adolescent health. The school sports environment plays a crucial role in stimulating students motivation to exercise and significantly impacts their level of physical activity participation. Research from environmental psychology indicates that both natural and social environments profoundly influence individual behavior. Consequently, having well-equipped sports facilities, fostering a positive sports atmosphere, and offering a variety of sports activities can effectively spark students enthusiasm for physical activity. Studies have shown that the school sports environment can positively predict adolescents exercise motivation and their engagement in physical activities32. Beyond directly promoting the adoption of a healthy lifestyle, the school sports environment also indirectly influences the formation of healthy habits by enhancing motivation for exercise and encouraging regular physical activity routines33. The school sports atmosphere, which combines the spirit of sports with material resources, helps to develop students awareness and appreciation of physical activity.

The home–school relationship refers to the communication, exchange, and cooperation that occur during the educational process. A harmonious and healthy home–school relationship is an important part of the environment for teenagers’ growth, is an intrinsic requirement of education, and promotes the development of teenagers. Research shows that families who actively communicate with schools and teachers have a positive impact on students’ academic performance, while families who do not do so negatively affect their children’s academic performance34. Variability in home–school communication can hinder students’ performance at school. Studies have suggested that parental involvement in school governance can improve student outcomes35, with collaborative home–school interaction being conducive to developing adolescents’ sense of self-efficacy and mitigating the “disadvantages” associated with socioeconomic status36. Students with a close home–school relationship outperform those with a distant one37. Furthermore, a harmonious home–school relationship, serving as a stabilizer for students’ well-being, not only promotes academic achievement but also benefits students’ physical and mental health38, positively impacting their academic and emotional development39. Thus, the home–school relationship, as a micro representation of the school environment, plays a positive role in creating an educational environment, with a good home–school relationship being a safeguard for the development of healthy behaviors in teenagers.

Both Bourdieu’s exposition of the reproduction of the elite class and P. Willis’s discussion on the reproduction of the working class reveal the relationship between schools and educational outcomes and their impact on the shaping of individuals’ futures. However, due to geographical distribution differences and the proximity enrollment system, there is a disparity in the composition of school classes. Coupled with the persistent unequal distribution of educational resources between urban and rural areas, there are regional differences in the level of governance within schools35. Studies have shown that the ___location of schools has a predictive effect on students’ academic performance40. Therefore, the administrative level of the school may also be a factor influencing teenagers’ physical activity behavior.

School atmosphere is a critical component of the school environment, referring to the enduring and stable characteristics of the humanistic environment within a school that are collectively experienced or perceived by students and staff, exerting a profound influence on their psychological and behavioral patterns. As the central venue for adolescent social interaction, the behavioral patterns and psychological tendencies of peer groups in schools have a significant group contagion effect, markedly influencing adolescents’ consciousness and behavior. This effect spreads through interactions among peer groups, leading to the so-called "peer contagion." Psychological research indicates that undesirable behaviors can be transmitted through group effects within school networks, subsequently affecting adolescents’ psychological states and behavioral patterns41. Driven by peer effects, a negative school atmosphere may persist within group networks and subtly influence adolescents’ psychological emotions and behavioral changes under the influence of conformity and group identity. The quality of the school atmosphere is directly related to the healthy growth of adolescents. A positive school atmosphere can promote the comprehensive development of students and stimulate their enthusiasm; in contrast, a negative atmosphere may lead to problematic behaviors among students. The school atmosphere also has a potential impact on adolescents’ physical fitness behaviors. Studies show that the campus atmosphere is a key factor affecting the mental health development of middle school students and has significant predictive power for students’ school performance. Moreover, the positive influence of the school atmosphere on youth physical exercise has been confirmed through cross-lagged analysis, indicating that optimizing the school atmosphere plays an active role in promoting youth physical exercise42. Therefore, the construction and maintenance of a positive school atmosphere are essential for the mental and physical health and the promotion of physical fitness behaviors among adolescents. In this study, due to limitations in data acquisition, the school atmosphere is defined by the frequency of occurrences such as fighting and other uncivilized behaviors within the school.

The Outline of the "Healthy China 2030" Plan points out the need to "ensure the allocation of sports venues and equipment in schools." Social capital such as sports venues shows a direct correlation with sports behavior43. School sports venues and facilities are the material basis for students’ physical exercise behavior, the guarantee for the survival and development of school sports, and the key to achieving quality education. With the high attention of national policies, the venues and equipment of primary and secondary schools have been significantly improved. However, due to uneven resource distribution, there are still urban–rural disparities, regional development imbalances, and other differences that restrict the development of primary and secondary school sports. Existing research has shown that teacher resources, venues and equipment, and regional imbalanced development have become difficulties that restrict the governance of school sports in the new era17. In addition, the requirements for the breadth and depth of physical education under the “double reduction” policy have further highlighted this issue, exacerbating the unfairness of school physical education and the inequality of student health. In summary, school environmental factors affect the development of adolescent physical exercise behavior. This article proposes the following hypothesis:

  • H3: The school environment has a significant positive impact on adolescents’ physical exercise behavior.

  • H3a: The home–school relationship has a significant positive impact on adolescents’ physical exercise behavior.

  • H3b: The administrative level of the school has a significant positive impact on the physical exercise behavior of adolescents.

  • H3c: The school atmosphere has a significant positive impact on adolescent physical exercise behavior.

  • H3d: Sports facilities have a significant positive impact on adolescent physical exercise behavior.

Simultaneously, the influence of school environments on family environments is multidimensional and complex. By placing greater emphasis on physical activities and underscoring their pivotal role in students holistic development, schools effectively shape family attitudes and participation in sports. Research suggests that optimizing the school sports environment not only reinforces family support for physical activities but also enhances students core physical literacy44. Moreover, the establishment of a trilateral collaborative model that involves family–school–community strengthens the foundation of home-based sports, establishes the core of school sports, and extends and complements community sports, thereby facilitating resource sharing and exploiting complementary advantages45. This family–school cooperation model has a positive impact on adolescents physical fitness behaviors by guiding families in exercise and reinforcing supervision, evaluation, and incentive mechanisms for extracurricular activities. Therefore, the following hypothesis is proposed:

  • H4: The impact of family environment on adolescent physical exercise behavior is moderated by school environment.

School sports are a crucial means of promoting adolescent physical activity, and their focus on sports directly impacts community residents awareness and engagement in physical fitness. By offering a variety of sports courses and activities, schools not only spark students interest in sports and develop their athletic abilities but also serve as a conduit to bring sports culture into families and communities, thereby enhancing the overall sports environment within the community46. Moreover, the deep integration of schools with communities facilitates resource sharing; for example, opening sports facilities to the public during non-teaching hours enriches community sports offerings and improves accessibility. Schools can further ignite community residents enthusiasm for participation and reinforce the appeal of community sports by organizing events and activities47. Ultimately, the synergistic collaboration between schools, families, and communities can collectively foster a vibrant sports culture. Thus, the following hypothesis is put forth:

  • H5: The impact of community environment on adolescent physical exercise behavior, with school environment playing a moderating role.

Under the guidance of relevant theories, this paper constructs a theoretical model of adolescent physical exercise behavior (Fig. 2).

Fig. 2
figure 2

Theoretical model diagram of adolescent physical fitness behavior.

Full size image

Research design

Data sources

This study utilizes data from the "China Education Panel Survey (CEPS)" as its foundation. The data project is meticulously designed and implemented by the China Institute for Educational Finance Research at Renmin University of China. To date, two rounds of data have been made publicly available, encompassing information on 19,487 and 10,750 students, respectively. This paper selects the first round of data as the baseline and efficiently matches the two rounds of data through School IDs and Student IDs, focusing on respondents who participated in both surveys. Additionally, this paper merges parental data from the second survey with the matched student data to obtain more comprehensive information. During the data processing, this paper excludes samples that could not be successfully revisited, newly added samples, and data with missing values, thereby constructing a sample set that aligns with the research objectives. The final student (family) data is 9449, and the school data is 112.

Variable selection

Dependent Variable: Dependent Variable: In this study, the dependent variable is physical exercise behavior, which refers to the pattern of individuals’ engagement in various physical activities, encompassing multiple dimensions such as the frequency, intensity, duration, and type of activities. The CEPS data surveys the weekly and daily physical exercise time of students. Based on the relevant information on physical activity behavior in the data, this paper multiplies the number of exercise days per week by the minutes of exercise per day and then takes the logarithm of the result to derive a variable measuring students’ physical exercise behavior. In this research, the dependent variable is the individual’s physical exercise behavior, a variable that includes the behavioral patterns of individuals’ engagement in sports activities, including the frequency, intensity, duration, and type of activities. Utilizing data from the China Education Panel Survey (CEPS), this study specifically examines the weekly and daily physical exercise duration of students. According to the information on physical activity behavior in the data, this paper constructs a variable to measure students’ physical exercise behavior by multiplying the number of exercise days per week by the minutes of exercise per day and then taking the logarithm of the product. This approach aims to more accurately reflect the overall level of students’ physical exercise behavior. The method has been widely adopted in many studies. In this study, Kendalls coefficient of concordance reached 0.92, indicating a high degree of consistency among raters. This suggests that the research findings are highly credible.

Independent Variables: In this study, the independent variables span two levels: the individual (family) and the school. At the individual level, the focus is on two variables, “family environment” and "community environment." The family environment encompasses aspects of the family’s economic status, social standing, and the relationships, behaviors, and emotions among family members. Previous research often considers parents’ education levels, family economic status, and social standing as key indicators of the family environment. In this study, we concentrate on the relationship between family dynamics, parents’ education levels, their educational expectations for their children, and adolescents’ participation in physical exercise. To this end, we have selected the following three measurement indicators: parents’ highest educational level, educational expectations, and parent–child relationship. Specifically, parents’ highest educational level refers to the higher of the highest educational levels achieved by the student’s father and mother (1 = No education, 2 = Primary school, 3 = Junior middle school, 4 = Technical secondary school/Vocational school, 5 = Vocational high school, 6 = Senior high school, 7 = Junior college, 8 = Undergraduate, 9 = Postgraduate and above). Parents’ educational expectations represent the highest level of education that parents expect for their children, obtained from parent questionnaires, and divided into three levels: 1 = High school and below, 2 = College (including junior college and undergraduate), 3 = Postgraduate and above. The parent–child relationship is measured by the questions "How is your relationship with your father?" and "How is your relationship with your mother?" in the student questionnaire, with data obtained by calculating the average score of the two questions, divided into three levels: 1 = Not close, 2 = Average, 3 = Very close. Upon our examination, the Cronbachs alpha coefficient α was found to be 0.681, indicating that the data exhibits acceptable reliability. The community environment is constructed from questions about the community in the parent questionnaire, including indicators such as the social class of community neighbors, community security, and community cultural environment. Due to the inability to track detailed data on the communities in which students reside, this study treats the community as an individual-level construct, with this indicator primarily measured through the subjective perceptions of parents, scored according to the following criteria: 1 = Very poor, 2 = Fairly poor, 3 = Fairly good, 4 = Very good. Finally, by calculating the average of these indicators, a comprehensive evaluation of the community environment is obtained. In this study, the Cronbachs alpha coefficient for the community environment indicators is 0.727, indicating a high degree of reliability for the data.

At the school level, the independent variable is school environment, which holistically reflects the overall context in which students learn and engage in activities within the campus. This variable encompasses several dimensions, including the administrative level of the school’s ___location, home–school relations, school atmosphere, and sports facilities. Specifically, the administrative level of the school’s ___location is a categorical variable, classified into four categories based on the principal’s questionnaire data: 1 for municipalities directly under the central government, 2 for provincial capital city districts, 3 for prefecture-level city districts, and 4 for counties or county-level cities. Home–school relations are measured by the frequency of interactions and communications between the school and families, reflected by the number of home–school activities held in the previous term. The school questionnaire includes questions such as "Number of parent–teacher meetings", "Number of times written reports on student performance were sent to parents," "Number of times parents were invited to observe classes," "Number of times parents were invited to have discussions with teachers," "Number of times parents were invited to watch performances or participate in extracurricular activities," and "Number of life guidance lectures held for students." These questions are scored on a scale of 1 (none) to 4 (more than 5 times). The average score of all questions provides a quantitative measure of home–school relationships. School atmosphere is assessed by the frequency of uncivil behaviors such as fighting within the school, scored as 1 for never, 2 for 1–4 times, 3 for 5–10 times, and 4 for more than 10 times. This study selected eight relevant variables and calculated their average to reflect the overall state of school climate. The assessment of sports facilities includes the quantity and quality of sports facilities available at the school, graded as 1 for none, 2 for present but in need of improvement, and 3 for present and in good condition. By calculating the average of these variables, we obtained a quantitative indicator for sports facilities. In this study, the internal consistency coefficient α for the school environment indicators was found to be 0.643, indicating a relatively high level of reliability for the data. Additionally, to minimize the impact of gender and ensure the results are universally applicable, the researchers included gender (coded as 0 for female and 1 for male) as a control variable. The statistical analysis showed that the mean variance inflation factor for all variables was less than 10, indicating no multicollinearity issues among the variables. To evaluate the validity of the questions, ten experts in the field were asked to review each question. The results indicated that the vast majority of these experts—nine out of ten—confirmed the strong relevance of the questions, suggesting that the content validity of the assessment is high. Detailed descriptive statistics are presented in Table 1.

Table 1 The variables describe the statistics table.
Full size table

Model results and analysis

School differences in physical exercise behavior

This study initially employs the null model of HLM to examine whether the physical exercise behavior of adolescents varies across different schools, and to estimate the proportion of total error that is attributable to inter-group differences. The inter-group difference ICC value is calculated as a basis for further analysis of other models. The HLM null model formula is as follows:

$$\text{Level}-1: {\text{Y}}_{\text{ij}}={\upbeta }_{0\text{j}}+{r}_{ij}$$
(1)
$$\text{Level}-2: {\beta }_{0j}={\gamma }_{00}+{\mu }_{0j}$$
(2)
$$\text{Full}: {\text{Y}}_{\text{ij}}={\gamma }_{00}+{\mu }_{0j}+{r}_{ij}$$
(3)

In this context, \({\upbeta }_{0\text{j}}\) represents the first-level intercept, indicating the average level of each school, and \({\text{r}}_{\text{ij}}\) represents the random error, indicating the within-group variation. \({\upgamma }_{00}\) represents the fixed effect of the first-level intercept at the second level, indicating the overall average of physical exercise behavior among students in all schools. \({\upmu }_{0\text{j}}\) represents the random error at the second level, indicating the between-group variation.

According to Tables 2 and 3, intra group variation \({\sigma }^{2}\hspace{0.17em}\)= 0.127, inter group variation \({\tau }_{00}\)=0.019 (\({x}^{2}=\)1187.502, \(\text{p}=0.000\)), therefore, ICC(1) = 0.131, when \(\text{ICC}(1)>0.059\), means that there is a moderate intra group correlation, 13.1% of the variable physical exercise behavior is caused by different schools, indicating that the inter group differences of physical exercise behavior in different schools cannot be ignored, and a cross-level analysis is required48. In addition, about 86.9% of [0.127/0.127 + 0.019] is caused by intra group differences. Therefore, the embedded variable of school environment is introduced to explain its variation.

Table 2 Results of the analysis of school differences in physical exercise behavior.
Full size table
Table 3 Results of the analysis of school differences in physical exercise behavior.
Full size table

Randomization test of family environment, community environment and physical exercise behavior

In order to test the relationship between family environment, community environment and physical exercise behavior, it is necessary to further test the randomness of the coefficient of the first level variable, add the family environment and community environment variables into the first level model of the multi-level model, and do not introduce explanatory variables into the second level model temporarily, that is, only the first level has independent variables, and the second level has zero mode. This analysis mainly tests whether the intercept and slope of the first level regression model exist. The correlation analysis model is as follows:

$$\begin{aligned} {\text{Level}} - 1:{\text{Y}}_{{{\text{ij}}}} & = \upbeta _{{0{\text{j}}}} + \upbeta _{{1{\text{j}}}} \left( {gender} \right) + \beta _{{2j}} \left( {parentaleducation} \right) + \upbeta _{{3{\text{j}}}} \left( {educationalexpectation} \right) \\ & \quad + \upbeta _{{4{\text{j}}}} \left( {parent - childrelationship} \right) + \beta _{{5j}} \left( {communityenvironment} \right) + r_{{ij}} \\ \end{aligned}$$
(4)
$$\text{Level}-2: {\beta }_{0j}={\gamma }_{00}+{\mu }_{0j}$$
(5)
$${\upbeta }_{1\text{j}}={\gamma }_{10}+{\mu }_{1j}$$
(6)
$${\upbeta }_{2\text{j}}={\gamma }_{20}+{\mu }_{2j}$$
(7)
$${\upbeta }_{3\text{j}}={\gamma }_{30}+{\mu }_{3j}$$
(8)
$${\upbeta }_{4\text{j}}={\gamma }_{40}+{\mu }_{4j}$$
(9)
$${\upbeta }_{5\text{j}}={\gamma }_{50}+{\mu }_{5j}$$
(10)

\({\gamma }_{00}\) represents the mean number of the intercept, \({\gamma }_{10}, {\gamma }_{20}\) represents the mean number of the slope, \({r}_{ij}\) represents the number of residual variants of Level 1 (within groups), \({\mu }_{0j}\) is the number of variants of the intercept, and \({\mu }_{1j},{\mu }_{2j}\) are the number of variants of the slope.

According to Table 4, after adjusting for student gender, variables such as parents’ highest educational level (t = 3.198, p < 0.001), educational expectations (t = 5.600, p < 0.001), parent–child relationship (t = 2.068, p < 0.05), and community environment (t = 4.382, p < 0.001) were found to significantly influence adolescent physical exercise behavior. These findings suggest that both family and community environments positively affect engagement in physical activity, indicating that adolescents exercise behaviors vary based on their family and community contexts. Consequently, hypotheses H1 and H2 are confirmed.

Table 4 Results of the random coefficient regression model analysis.
Full size table

Effect of the school environment on adolescent physical exercise behavior

In order to further understand the main effects of the school environment, by adding the school environment to the intercept model, the intercept prediction model (intercepts-as-outcomes model) was analyzed to understand the influence of the school environment on physical exercise behavior.

$$\begin{aligned} {\text{Level}} - 1:{\text{Y}}_{{{\text{ij}}}} & = \upbeta _{{0{\text{j}}}} + \upbeta _{{1{\text{j}}}} \left( {gender} \right) + \beta _{{2j}} (parentaleducation) + \upbeta _{{3{\text{j}}}} \left( {educationalexpectation} \right) \\ & \quad + \upbeta _{{4{\text{j}}}} \left( {parent - childrelationship} \right) + \upbeta _{{5{\text{j}}}} (communityenvironment) + r_{{ij}} \\ \end{aligned}$$
(11)
$$\begin{aligned} {\text{Level}} - 2:\beta _{{0j}} & = \gamma _{{00}} + \gamma _{{01}} (home - schoolrelationship) + \gamma _{{02}} (theadministrativelevel) \\ & \quad + \gamma _{{03}} (schoolatmosphere) + \gamma _{{04}} (sportsfacilities) + \mu _{{0j}} \\ \end{aligned}$$
(12)
$${\beta }_{1j}={\gamma }_{10}+{\mu }_{1j}$$
(13)
$${\upbeta }_{2\text{j}}={\gamma }_{20}+{\mu }_{2j}$$
(14)
$${\upbeta }_{3\text{j}}={\gamma }_{30}+{\mu }_{3j}$$
(15)
$${\upbeta }_{4\text{j}}={\gamma }_{40}+{\mu }_{4j}$$
(16)
$${\upbeta }_{5\text{j}}={\gamma }_{50}+{\mu }_{5j}$$
(17)

\({\gamma }_{00}\) is the intercept of Level-2, \({\gamma }_{01}\), \({\gamma }_{02}\), \({\gamma }_{03}\), \({\gamma }_{04}\) is the slope of Level-2, and \({\gamma }_{10}\), \({\gamma }_{20}\), \({\gamma }_{30}\), \({\gamma }_{40}\), \({\gamma }_{50}\) is the mean number of the slope. The regression results were obtained according to the formula (Table 5).

Table 5 Regression model of the school environment on adolescent physical exercise behavior.
Full size table

As shown in Table 5, after the researchers controlled for gender, the home–school relationship (t = 2.348, p < 0.01) was found to have a significant positive effect on the dependent variable. This indicates that a stronger connection between home and school is correlated with increased levels of physical exercise behavior in adolescents, thus providing support for Hypothesis H3a in this study. The administrative level of the school (t = − 5.964, p < 0.001) and the school atmosphere (t = − 2.336, p < 0.05) negatively affected physical exercise behavior. In this study, the school administrative level variable is coded as integers from 1 to 4, with 1 representing municipalities directly under the central government, and higher numbers indicating lower administrative levels. The school atmosphere variable is also coded as integers from 1 to 4, with 1 representing schools that have never experienced fighting or other uncivil behaviors, and higher numbers indicating a poorer school climate, i.e., a higher frequency of uncivil behaviors. This suggests that higher administrative levels and better school environments are associated with increased levels of physical exercise among teenagers, supporting research hypotheses H3b and H3c. However, the impact of sports facilities (t = 1.039, p > 0.05) on physical exercise behavior was not significant, indicating that the availability of school stadiums and other facilities does not predict physical exercise behavior among adolescents. It is evident that specific aspects of the school environment can influence the physical exercise behavior of adolescents. This finding supports the study hypothesis that the home–school relationship, administrative level, and school atmosphere impact adolescents physical exercise behavior.

The regulation effect of different school environment on the family and community environment

Based on the above analysis, in order to understand the relationship between family environment, community environment, and physical exercise behavior in different school environments, this study employs a Hierarchical Linear Model (HLM) to examine the differences in the slope relationships between school environment, family environment, and community environment on physical exercise behavior. Level 1 includes physical fitness behavior as the dependent variable, incorporates family environment and community environment, and adds four elements of school environment to the intercept and slope models in Level 2 to test the cross-level moderating effect of school environment on physical exercise behavior and the slope relationships between family environment and community environment. The HLM model is as follows:

$$\begin{aligned} {\text{Level}} - 1:{\text{Y}}_{{{\text{ij}}}} & = \upbeta _{{0{\text{j}}}} + \upbeta _{{1{\text{j}}}} \left( {gender} \right) + \beta _{{2j}} \left( {parental\;education} \right) + \upbeta _{{3{\text{j}}}} \left( {educational\;expectation} \right) \\ & \quad + \upbeta _{{4{\text{j}}}} \left( {parent - childrelationship} \right) + \beta _{{5j}} \left( {community\;environment} \right) + r_{{ij}} \\ \end{aligned}$$
(18)
$$\text{Level}-2: {\beta }_{0j}={\gamma }_{00}+{\gamma }_{01}\left(school environment\right)+{\mu }_{0j}$$
(19)
$${\upbeta }_{1\text{j}}={\gamma }_{10}$$
(20)
$${\upbeta }_{2\text{j}}={\gamma }_{20}+{\mu }_{2j}$$
(21)
$${\upbeta }_{3\text{j}}={\upgamma }_{30}+{\text{r}}_{31}(school environment)+{\mu }_{3j}$$
(22)
$${\upbeta }_{4\text{j}}={\gamma }_{40}+{r}_{41}(school environment)+{\mu }_{4j}$$
(23)
$${\upbeta }_{5\text{j}}={\gamma }_{50}+{r}_{51}\left(school environment\right)+{\mu }_{5j}$$
(24)

Merge Level-1 and Level-2 to obtain the full model:

$$\begin{aligned} {\text{Full Model}}:{\text{Y}}_{{{\text{ij}}}} & = \gamma _{{00}} + \gamma _{{10}} \left( {gender} \right) + \gamma _{{20}} \left( {parental\;education} \right) + \gamma _{{30}} \left( {e{\text{ducational expectations}}} \right) \\ & \quad + \gamma _{{40}} \left( {p{\text{arent}} - {\text{child relationship}}} \right) + \gamma _{{50}} \left( {community\;environment} \right) + \gamma _{{01}} \left( {school\;environment} \right) \\ & \quad + r_{{31}} \left( {e{\text{ducational expectations}}} \right)\left( {school\;environment} \right) + r_{{41}} \left( {Parent - childrelationship} \right)\left( {school\;environment} \right) \\ & \quad + r_{{51}} \left( {community\;environment} \right)\left( {school\;environment} \right) + \mu _{{0j}} + \mu _{{1j}} \left( {family\;environment} \right) \\ & \quad + \mu _{{2j}} \left( {family\;environment} \right) + \mu _{{3j}} \left( {family\;environment} \right) + \mu _{{4j}} \left( {family\;environment} \right) \\ & \quad + \mu _{{5j}} \left( {community\;environment} \right) + r_{{ij}} \\ \end{aligned}$$
(25)

Specifically, gender was the control variable, and only its fixed effects were considered. Results are obtained according to the formula (Table 6).

Table 6 Regression model of school environment.
Full size table

Level one describes how the influence of family and community environments on adolescent physical exercise behavior varies based on the different organizational levels of school environments. This result must be explained by variables nested within the independent variables of level one. Table 6 presents a slope prediction model that controls for gender variables, examining the moderating effect of school environments on the relationship between family and community environments, as well as adolescent physical exercise behavior. The results show that under level one educational expectations, the relationship between home and school has a positive moderating effect on physical exercise behavior, while other factors have no significant moderating effect. This suggests that parents’ educational expectations play a moderating role in enhancing adolescent physical exercise behavior. Adolescents with harmonious home–school relationships may engage more in physical exercise behavior, indicating that home–school relationships are moderating variables for adolescent physical exercise behavior. This partially verifies the hypothesis H4 of this study, which posits that school environments have a moderating effect on the relationship between family environments and physical exercise behavior. However, the impact of community environments on physical exercise behavior does not vary with different school environments, suggesting that the moderating effect of school environments in this relationship does not exist.

From the random effects, it is evident that after incorporating variables related to family, community, and school environment, the random effects of the intercept terms, parental education level, and educational expectations remain significant. Consequently, further exploration of other explanatory variables is necessary. Nevertheless, for parent–child relationships and community environment, their variations have already been well explained, obviating the need for additional explanatory.

Conclusion and discussion

Environment is an important factor affecting the physical exercise behavior of adolescents. While assisting and guiding the ongoing refinement of beliefs, it also serves as a catalyst for promoting healthy actions among adolescents. Studies have found that the external environment plays a promoting role in the commitment to and persistence in exercise behavior49. Therefore, children and adolescents can obtain opportunities to participate in sports activities continuously in environments such as families, communities, and schools. Based on the analytical framework of “ecosystem theory” on the relationship between individual development and environment, utilizing data from the China Education Panel Survey, this study focuses on examining the comprehensive impact mechanism of family, community, and school environments on the physical exercise behavior of adolescents within the micro-system of the adolescent environment, providing a basis for enhancing adolescents’ physical exercise behavior from an environmental perspective.

Firstly, the study verified that family environment and community environment are the influencing mechanisms of adolescent physical exercise behavior, and parents’ education level, educational expectations, parent–child relationship, and community environment have a positive impact on adolescent physical exercise behavior. As the main body of children and adolescents’ sports environment, the environment created and constructed by families and communities is an important field for the formation of adolescent sports behavior50. Support from family and community environments has a protective effect on adolescent physical exercise behavior. The behavior of family members is mutually influenced. For families with long-term and closest relationships, family environmental factors are the fundamental reason for affecting children’s and adolescents’ sports behavior.

There is a direct correlation between the educational level of parents and their cognitive abilities. A one-sided understanding of the essence of education and the value of sports can lead to one-sidedness, biases, and absences in the planning, guidance, and other aspects of their children’s education. A lack of proper sports guidance and companionship during childhood can result in diminished enthusiasm for and participation in sports51. Another study also points out that the higher the educational level, the stronger the desire to participate in sports activities52. Families with higher educational attainment have relatively strong health awareness, which further affects the development direction of adolescent sports behavior, enabling their family-supported adolescents to engage more actively in sports53. The trust and emotional expression established between parents and children during their joint participation in sports activities help to establish good parent–child relationships and trust relationships54, which contribute to the development of healthy and positive emotions. This aids adolescents integrate into groups quickly and further fostering the consolidation of their physical exercise behaviors8. Therefore, good parent–child relationships have a dual positive impact on promoting adolescent physical exercise behavior. Parental educational expectations are an exogenous motivation for the development of adolescent behavior, and their mechanism and effect have a “Pygmalion effect”, that is, children tend to develop in the direction expected by their parents55. Parental educational expectations have an important influence on their children’s future development and goals56.

Some scholars explain this mechanism using the "Rosenthal effect," which suggests that based on their belief and admiration for their parents, adolescents’ behaviors and thoughts are subtly influenced by parental suggestions and impacts, leading them to develop in the direction expected by their parents and manifest in their actions57. Therefore, educational expectations affecting the future development of adolescents are an effective way to motivate self-actualization58. Parental educational expectations not only provide goals and motivation for striving and progress, but also have an equally important impact on the development of learning and other behavioral habits, such as positively influencing adolescents’ physical exercise behavior. However, behavioral interaction theory argues that the reasons for changes in individual attitudes and behaviors cannot be simply attributed to individual or environmental factors, but should also consider the interaction between the two59. Therefore, parental educational expectations are only one factor affecting adolescents’ physical exercise behavior, and future research should also focus on the interaction of multiple factors. From the perspective of the impact mechanism, parental educational expectations and parent–child relationships play a key role in adolescents’ physical exercise behavior, which further validates Kellman’s theory of family social capital. Therefore, parents should respect the educational wishes of adolescents, moderately adjust their educational expectations, coordinate parent–child relationships, and provide attention and support through various aspects such as more companionship and emotional investment, so that adolescents can feel the good environmental support from their families. At the same time, the popularization, standardization, and professionalization of community microenvironment support will be a powerful support for improving adolescents’ physical fitness behavior. As the most basic level of governance and living space, the physical and humanistic environments of communities can affect adolescents’ physical exercise behavior60. Existing research believes that a good community environment can stimulate adolescents’ exercise intentions and improve their level of physical activity during non-school periods61, which is consistent with the results of this study. As revealed by social adaptation theory, a good community atmosphere enhances individual health behavior62.

Secondly, Within the school microsystem, this study confirms that the school environment positively influences adolescents physical exercise behavior. It also verifies that the effect of educational expectations on physical exercise behavior is moderated by the quality of home–school relationships. Specifically, the impact of educational expectations differs depending on the level of engagement between home and school. When home–school relationships are strong, they can enhance adolescents physical activity levels by positively influencing their educational expectations. The strength of the home–school relationship is directly proportional to the influence of educational expectations on adolescents exercise behavior. Moreover, robust home–school connections not only serve as a positive predictor of academic performance but also significantly contribute to the development of adolescents behavior. These relationships are instrumental in fostering both academic success and desirable behavior in young people. This mechanism can be explained by the Overlapping Spheres of Influence Theory. According to this theory, the influences of schools and families on students are interconnected and mutually reinforcing63. For adolescents to achieve their goals to the fullest extent, it is essential that there is cooperation between these two spheres. A substantial body of research has confirmed that a positive home–school relationship not only has a beneficial effect on adolescents academic performance but also exerts a significant deterrent influence on externalizing problem behaviors64. Building on this foundation, the current study suggests that a cooperative co-education model, grounded in a robust home–school relationship, is also a crucial factor in fostering the development of adolescent physical exercise behaviors. Effective communication serves as the cornerstone in establishing home–school relationships. By fostering effective communication, the domains of school and home are interconnected, thereby integrating educational efforts across both time and space. This not only promotes the self-awareness and self-reflection of educational agents from both home and school, but also drives the renewal of educational concepts and practices, and contributes to fostering a harmonious educational environment. Throughout this collaborative process, both parties have the opportunity to learn from each other, refine their individual educational roles, and address each others shortcomings and limitations, collectively working towards a win–win outcome in education65. Therefore, when exploring the impact of the relationship between home and school on the behavioral development of adolescents, it is necessary to consider not only the direct effect of this environmental variable but, more importantly, the indirect effect on adolescents’ psychology and behavior. Namely, how the school environment can internalize individual motivation into behavioral outcomes through variables such as educational expectations and family environment66. Elucidating this mechanism remains a challenge for future research. In addition, due to variations in regional economic structures, schools differ in terms of institutional management, infrastructure, and overall school atmosphere. This is evident in the overall higher standards of basic education and sports participation in economically developed eastern regions compared to central and western regions. Among them, the accumulation of resources, including the administrative level of the school, school atmosphere, and sports facilities, all have an impact on the level of physical exercise among adolescents. The underlying issue stems from the unequal distribution of high-quality educational resources. Schools located in eastern regions pay more attention to the development of comprehensive quality, the improvement of physical education and health literacy, coupled with high-quality sports resources and more sufficient activity time67, which promotes students’ participation in sports and increases the possibility of students participating in physical exercise68, thus affecting adolescents’ physical exercise behavior. It can be seen that due to the scarcity and limited nature of resources, the development of adolescents’ physical exercise behavior has regional priority and hierarchical characteristics, leading to health inequality among adolescents. Environmental perception theory suggests that perceived environmental information can stimulate or alter the motivation and decision-making of individuals, thereby influencing the development of related behaviors. The role of peers as role models acts as a catalyst, can facilitate behavior transmission, and enhance the school’s exercise atmosphere, thus affecting the acquisition and occurrence of exercise behaviors among young people69. A positive school atmosphere positively influences the physical exercise behaviors of adolescents, reducing negative behaviors among peers70. Furthermore, the quality of a school’s sports facilities is directly proportional to the accessibility and practicality of sports activities, which is conducive to promoting the level of youth sports participation, aligning with existing research findings71.

The above findings hold significant policy implications for enhancing teenagers’ physical activity and promoting a healthy China. Firstly, at the school level, we should effectively enhance support by improving the school environment. This involves not only addressing shortcomings in facilities and equipment, enhancing educational conditions, mitigating resource disparities due to regional development imbalances, and achieving inter-school educational equality. We should also strive to reshape the school’s sports culture and enhance the cultural and humanistic ambiance. Furthermore, it is essential to extend school education to include families, create an effective platform for home–school interaction, encourage parental involvement in school governance, explore diverse models of home–school collaboration, and strengthen the nurturing role of schools towards families. Secondly, for communities, constructing and enhancing infrastructure is crucial. Actively fostering the development of community sports, minimizing disparities in equipment resources due to class differences within the community, and optimizing the utilization and accessibility of existing resources are key strategies. Additionally, shaping the community by fostering a humanistic environment and organizing a wide range of community activities can promote interactions among residents and enhance community cohesion and collective identity. Finally, at the family level, parents should adopt a sound view of health education, pay attention to and care for the growth and needs of adolescents, place scientific and reasonable educational expectations on their children, and actively cultivate a family atmosphere conducive to fostering a positive parent–child relationship. Regarding uncontrollable factors, educators should focus on the following two key areas: (1) Personalized Interventions: Provide interventions tailored to the specific circumstances and needs of each student. For those significantly affected by immutable factors, educators should offer extra support and resources. (2) Collaboration and Coordination: Work with families, community organizations, and other relevant institutions to develop and implement effective intervention strategies. Ensure that resources and efforts are coordinated at different levels to maintain consistency and enhance the effectiveness of these interventions.

In summary, based on ecosystem theory, this paper verifies the significant impact of family, community, and school environments on adolescents’ physical exercise behavior, and explores the mechanisms by which these environments influence their physical activity. The conclusions drawn contribute to a deeper understanding of the mechanisms affecting adolescents’ physical exercise behavior, serve as a reminder to recognize the importance of the environment on physical activity, and facilitate a more academic understanding of the influence of the environment on the family, school, and society. This helps to rationally plan for the coordinated and balanced development of family, community, and school environments, and to build an environmental infrastructure that is family-based, campus-centered, and community-supported, providing a reliable foundation for the physical health of adolescents.