Mapping the physical condition, peer and parental support, and MVPA: a network analysis in schoolchildren from the ACTIBESE project

Abstract

Background

Active behaviour denotes any type of movement that enhances overall physical well-being, a trend that has seen a gradual decline among schoolchildren globally. This study aimed to examine the complex interplay between physical condition, peer and parental support, and moderate and vigorous physical activity, both inside and outside school, using network analysis in a sample of Chilean children and adolescents.

Methods

The sample includes 655 schoolchildren aged 8 to 16. Instruments used include a questionnaire for sociodemographic and physical activity (PA) variables, assessing factors like gender, age, socioeconomic level, PA levels, self-perceived physical condition, and parental and peer support for PA. Descriptive statistics were used alongside linear regression to establish associations. Network Analysis, utilising the Fruchterman-Reingold algorithm, was employed to visualise relationships between variables. Centrality indices were calculated to determine variables' influence within the network, while sparsity values indicated network density.

Results

In children, recess-related activity strongly correlates with moderate-to-vigorous PA (MVPA), whereas in adolescents, MVPA is associated with activities during recess, lunchtime, and physical education classes. Parental support plays a significant role, particularly in adolescents. Outside school PA is well connected among peer support nodes in both children and adolescents, while adolescents show stronger connections with parental support, socioeconomic variables, and physical condition nodes. Notably, friends' encouragement and parental support were significantly associated with PA behaviours in both age groups.

Conclusion

Outside school PA, parental and peer support have a strong association with MVPA in children and adolescents.

Introduction

Active behaviour

Active behaviour refers to any form of movement contributing to overall physical well-being, a trend that has progressively declined among schoolchildren worldwide [1]. Current global estimates indicate a stagnation in activity levels, with no significant improvement over the last decade [2]. Numerous factors negatively impact active behaviour, including schoolchildren’s social and educational context and the overuse of sedentary technologies [3]. Conversely, PA has been linked to numerous health benefits, particularly in youth [4, 5], impacting physical, psychosocial, emotional, cognitive, and academic aspects [6]. Achieving these benefits necessitates adherence to recommended PA levels [7]. However, despite global efforts to promote movement, there is a pervasive “inactivity pandemic”; and less participation in physical education (PE) classes [8]. However, a few children and adolescents (20%) meet the daily recommendations of 60 min of moderate-to-vigorous PA (MVPA) worldwide [2, 9]. This widespread non-compliance suggests that individual-level interventions are insufficient, pointing to the need for a deeper exploration of the interpersonal and environmental drivers within the socio-ecological framework [10].

Given that children spend significant time in school, especially in less affluent nations like Chile, where educational settings often represent the primary opportunity for structured movement [11], declining PA and PE participation rates are concerning [12–14]. In this context, schools in emerging economies act as critical hubs where the interplay between institutional resources and social support systems determines whether a child reaches the recommended activity thresholds.

Interpersonal support

Bronfenbrenner’s socioecological theory emphasises the combined influence of individuals, families, schools, and communities on the development of schoolchildren [15]. It recognises personal, interpersonal, and environmental factors affecting PA levels [16]. Within this framework, interpersonal influences, specifically from parents and peers, act as critical microsystems that can either facilitate or hinder the transition from sedentary to active behaviours.

The relationship between parents and their offspring plays a crucial role in shaping PA behaviours. Beyond being primary role models who shape attitudes and preferences towards PA [17], parents provide a multifaceted support system. This involvement can take various forms, including logistical support (such as transportation to sports facilities or providing equipment), participating in PA together, and fostering a supportive environment at home [17, 18]. Crucially, this influence remains dynamic from early childhood through adolescence, where the parental shift from direct supervision to emotional encouragement becomes vital [16, 19].

While much of the traditional literature focuses on the general health benefits of PA, such as reduced obesity or improved self-esteem [16], the socioecological perspective suggests that the mechanism of influence is primarily social. For instance, parental enjoyment and perceived importance of exercise are directly mirrored in their offspring’s activity levels [17]. Numerous studies underscore this significance; for example, a meta-analysis by Edwardson et al. [17] revealed a positive association between parental support and children’s engagement in PA, highlighting that the domestic environment is the primary setting for early physical literacy. However, as schoolchildren transition into adolescence, the interpersonal landscape shifts. Family support remains significant, with mothers often playing a pivotal role in organising and encouraging PA [20, 21], but peer support emerges as an equally critical determinant.

On the other hand, peers impact active behaviour by fostering motivation, social identity, and enjoyment of PA [22, 23]. Understanding the synergy between these two interpersonal pillars, parents and peers, is vital for moving beyond individual-centred approaches and designing effective interventions that leverage the existing social networks of schoolchildren [24].

Inside and outside school PA

The social and educational context of schoolchildren provides a unique dual-environment opportunity to increase MVPA: the structured setting of the school and the autonomous space of the home and community [25]. Distinguishing between these two domains is critical because the socio-ecological barriers and facilitators differ significantly; while in-school PA is often governed by institutional resources and peer dynamics, out-of-school PA is heavily dependent on parental support and neighbourhood safety [26].

Regarding the above, children spend significant time in school, especially in less affluent nations like Chile, where a profound socioecological gap exists. In this context, school often fills gaps in educational and social development that are not addressed in the home environment due to socioeconomic constraints [10]. Declining PA and PE participation rates are even more concerning when considering that for many Chilean children, the school is the only safe space for physical movement [12–14]. Recent reports on PA in Chilean youth indicate suboptimal levels, both in-school and outside school, necessitating interventions that address these specific environmental disparities [27–29].

However, although educational curricular goals related to PA are sufficient, especially concerning active play and active transportation, there is a disconnect in how family and peer influences interact with these goals across different settings [28, 29]. Thus, there’s a need for a deeper exploration of the context-specific reasons for low PA among Chilean schoolchildren. Understanding how the school-home transition affects active behaviours is vital to bridging the current research gap [30].

Providing combined sessions that included opportunities for both free play and structured adult-led games led to higher levels of MVPA when compared to sessions solely focused on organisation [29, 31]. This evidence suggests that optimising in-school time through structured autonomy can compensate for the lack of PA opportunities outside school [32]. This has significant implications for public health initiatives to enhance children’s PA, as providing opportunities for free play emerges as a straightforward and potentially impactful approach to optimising the time dedicated to PA within the diverse Chilean socio-ecological landscape. “Thus, this study aimed to examine the complex interplay between physical condition, peer and parental support, and moderate and vigorous physical activity, both inside and outside school, using network analysis in a sample of Chilean children and adolescents from the ACTIBESE Project.

Methods

Study design and sample

The ACTIBESE (Active Behaviour in School Education) project is a government-funded research initiative in early education aimed at enhancing knowledge within the school system and refining its practices [33]. Principals of selected schools were invited to participate in ACTIBESE, followed by a comprehensive meeting involving administrative representatives and teachers to outline project objectives and stages. Subsequent informational sessions were conducted with parents and students, categorised by age: children in grades 3–6 (8 to 11 years old) and adolescents from grades 7 to grade 2 of secondary school (12 to 16 years old). The project entails both a cross-sectional study and an intervention study. The data analysed in this article correspond to the cross-sectional study.

Participants in the ACTIBESE project are drawn from two cities in the Valparaíso Region (Viña del Mar and Los Andes) and one city in the Araucanía Region (Temuco), from Chile and selected for logistical convenience. Public, subsidised, and private administration schools were included to reflect the diversity of the Chilean educational landscape. The sample encompasses students from rural and urban areas across various social contexts.

A total of 152 schools were identified, comprising 84 in Viña del Mar (n = 25,176), 19 in Los Andes (n = 7,401), and 49 in Temuco (n = 21,647). The sample size for the cross-sectional study was determined with 95% confidence, 5% error, and 50% heterogeneity. Thus, a representative sample of n = 382 schoolchildren was obtained by clusters, with n = 177 in Viña del Mar, n = 52 in Los Andes, and n = 153 in Temuco. Ultimately, n = 655 schoolchildren participated, including n = 233 children (10.7 ± 0.52) and n = 432 adolescents (13.8 ± 1.39). Further details can be found in Table 1.

Table 1.

Sociodemographic characteristics of children and adolescents

			Children
	Boys		Girls		Total
	N	(%)	N	(%)	N	(%)
Gender	112	(50.2)	111	(49.8)	223	(100)
Age (Mean ± SD)	10.6	± 0.53	10.6	± 0.51	10.7	± 0.52
Socio-economic level
Low	39	(34.8)	33	(29.7)	72	(32.3)
Medium	50	(44.6)	53	(45.7)	103	(46.2)
High	23	(20.5)	25	(26.5)	48	(21.5)
School
Public	65	(58.0)	55	(49.5)	120	(53.8)
Subsidized	26	(23.2)	27	(24.3)	53	(23.8)
Private	21	(18.8)	29	(26.1)	50	(22.4)

			Adolescents
	Boys		Girls		Total
	N	(%)	N	(%)	N	(%)
Gender	233	(53.9)	199	(46.1)	432	(100)
Age (Mean ± SD)	13.7	± 1.35	13.9	± 1.43	13.8	± 1.39
Socio-economic level
Low	70	(31.6)	51	(25.6)	121	(28.0)
Medium	108	(45.8)	91	(45.7)	199	(46.1)
High	55	(22.6)	57	(28.6)	112	(26.0)
School
Public	87	(44.1)	50	(25.1)	137	(31.7)
Subsidized	53	(22.9)	69	(34.7)	122	(28.2)
Private	93	(39.9)	80	(40.2)	173	(40.0)

Instruments

A multicomponent questionnaire was used to obtain sociodemographic and PA variables (Table 2). In sociodemographic data, the following were considered: gender, age, course, and socioeconomic level (SEL) through the Family Affluence Scale (FAS III), with questions such as the number of vehicles, household appliances, electronic devices, bathrooms, whether they have their own room, and if they travel abroad for vacation [34]. According to the score obtained, it is classified as low level (1–4 points), medium level (5–8 points), and high level (9–12 points).

Table 2.

Groups of nodes, nodes and their definition are used in the network analysis

Variables for network analysis
Groups of nodes	Acronymous	Definition of nodes
Socioeconomic level	SEL1	Does your family have a car or another motorized vehicle?
	SEL2	Do you have your own room?
	SEL3	How many computers bought by your family do you have?”
	SEL4	How many bathrooms are there in your house?”
	SEL5	Does your family have a dishwasher?
	SEL6	How many times has your family travelled outside the country for holidays/vacations in the last year?”
Physical activity in school	PA1	Physical activity during physical education classes
	PA2	Physical activity during recess
	PA3	Physical activity during the lunch break
Physical activity outside school	PA4	Physical activity after your school hours until to 18:00
	PA5	Physical activity in the afternoons (between 18:00 and 22:00 h.)
	PA6	Physical activity in weekend
Physical condition	PC1	General physical condition
	PC2	Cardiorespiratory fitness
	PC3	Muscle strength
	PC4	Speed/agility
	PC5	Flexibility
Parental support	SPa1	To see how you practice physical activity or sport?
	SPa2	Do you feel encouraged to do physical activity or sports?
	SPa3	Transport for physical activity or sports?
	SPa4	Do physical activity or sports with you?
	SPa5	To speak about importance of physical activity?
	SPa6	Do you have equipment at home to do physical activity or sports?
Peer support	SPe1	Friends encourage you to do physical activity?
	SPe2	Friends do physical activity or sports with you?
	SPe3	Friends or classmates mock for practice physical activity or sports?
	SPe4	Friends tell you that you’re doing well when you practice physical activity or sports?

To assess PA levels, we utilised the Youth Activity Profile (YAP) questionnaire, a validated self-report tool known for its simplicity and cost-effectiveness in estimating children’s moderate-to-vigorous PA (MVPA) and sedentary behaviour on a group level [35]. A Spanish (Chilean) adaptation of the YAP questionnaire was developed through a rigorous back-translation process, ensuring its feasibility and reliability for Spanish-speaking children [36]. Comprising nineteen items, including four general inquiries and fifteen specific ones, the YAP questionnaire is structured into three sections: (1) Activity at School, (2) Outside School Activity, and (3) Sedentary Habits. The data analysis involved computing average scores ranging from 1 to 5 (with 1 representing the lowest and 5 the highest) to gauge PA at school and outside of school. Furthermore, to quantify PA among children and adolescents, YAP scores will be converted into minutes per day of MVPA using the Fairclough equations [37]. These computations will yield separate estimates for MVPA duration during school hours, outside of school, and over the weekend, disaggregated by age group and gender.

The tool for assessing self-perceived physical condition among schoolchildren was the International Fitness Scale (IFIS) questionnaire, a validated assessment instrument designed to evaluate overall physical condition [38]. This questionnaire comprises five items, each corresponding to different aspects of general physical condition, namely cardiorespiratory fitness, muscular strength, speed/agility, and flexibility. Responses are recorded using a Likert-type scale ranging from 1 to 5, where 1 indicates low PA and 5 indicates high PA.

Parental support, both from mothers and fathers, was evaluated independently through a 5-item scale structured around the query: “How often does your mother/father engage in these activities?”. Participants rated responses on a 4-point Likert scale, ranging from “very frequently” to “never”. This assessment pertained to the level of encouragement for PA, the provision of logistical support (such as transportation), observation, and verbal affirmation of PA participation. Additionally, it encompassed the involvement of parents in physical activities alongside their children [39].

Peer support was assessed through two criteria: encouragement for PA and engagement in physical activities with peers. Participants were queried with the prompt: “How often do your friends participate in these activities?” Responses were gauged on a 4-point Likert Scale, ranging from “very frequently” to “never”. Additionally, peer socialisation was evaluated by quantifying the time spent with friends both “after school” and “in the evenings” on school days. Participants were asked: “How many days a week do you typically spend time with friends after school?” and “How many evenings during the school week do you usually spend time with friends?” [40]. Parental and peer questionnaires have been reliable for the Chilean context [41].

Procedures

Once the authorisations from the schools and parents were obtained, the research team attended the schools to administer the questionnaires. The instruments were administered in physical education or ethics orientation classes.

The self-report questionnaires were administered in pencil and paper format and lasted a maximum of 30 min. The evaluators hired specifically for the ACTIBESE project were trained physical education teachers who answered the children’s questions.

Statistics

Before the analysis, a Q-Q plot (quantile-quantile plot) will be used to visually verify normality. In addition, an interaction between sex and age was explored.

Descriptive statistics for study variables were obtained using the mean and standard deviation for continuous variables and frequency and percentages for categorical variables. The results were analysed according to age (children and adolescents) and gender (girls and boys). For this statistical analysis, SPSS version 29 (SPSS Inc., Chicago, IL, USA) was used.

The relationships between MVPA inside and outside the school, and all the assessed correlates, were calculated using a “Machine Learning” technique entitled Network Analysis. This technique establishes relationships through multiple interactions between variables from graphical representations [42]. Concerning the cross-sectional nature of this study, an undirected weighted network analysis was used to estimate the correlation between nodes (variables) from a correlation matrix, which, when transformed, are represented by positive or negative edges, which are the relationships between the different nodes, but with no arrowheads to indicate the direction of effect. A general network for children and adolescents, comprising all the samples, and two other different networks for adolescents, according to sex, were created.

The “Fruchterman-Reingold” algorithm was applied so that data were presented in the relative space where variables with stronger associations remained together, and the less strongly associated variables were repelled from each other [43]. The pairwise Markov random field model was used to improve the accuracy of the partial correlation network, which was estimated from L1 regularised neighbourhood regression. The least absolute contraction and selection operator was used to obtain regularisation and to make the model less sparse [44]. The Extended Bayesian Information Criterium (EBIC) parameter was adjusted to 0.5 to create a network with greater parsimony and specificity [45].

Regularised algorithms of the selection operator and minimum absolute reduction (LASSO) were used to obtain the precision matrix, which, when standardised, represents the associations between network variables. The thickness and colour intensity of the lines represent the magnitude of the associations. The blue lines represent positive associations, and the red lines represent negative ones. The plots were made with a cross-platform software program named JASP statistical package V 0.18.3 [46], a free software interface Version 0.18.3 (Nieuwe Achtergracht 129B, Amsterdam, The Netherlands).

The centrality index “Expected Influence”, which indicates the importance of a node for the structure and functioning of the network, was calculated. This centrality measure consists of the sum of all possible edge weights that connect one node to another [42]. It was used to assess the nature and strength of a variable’s cumulative influence within the network and, thus, the role it may be expected to play in the activation, persistence, and remission of the network [47]. A positive expected influence means that the influence of that specific node in the network tends to increase for the acquisition of an adequate network pattern.

The sparsity of the network, which indicates whether the network is dense or sparse, based on whether the number of edges is quadratic or linear in the number of nodes, was also calculated. This value shows the number of edges (connections) obtained in each emerged network. For the current study, sparsity was calculated based on non-zero relationships between nodes, considering that each node was connected to all other nodes in the network.

Ethics aspects

Once the respective schools have authorised the project, parents must sign an informed consent form for participation. Likewise, the students signed an informed consent form that described the study’s characteristics, aims, and activities. The project was developed following the Declaration of Helsinki and with the approval of the Bioethics Committee of the Pontificia Universidad Católica de Valparaíso (Code: BIOEPUCV-H 638–2023).

Results

Table 1 shows the general results of the study participants, stratified by sex and age group. The total sample consisted of 655 participants, divided into children (n = 223) and adolescents (n = 432). In the children’s group, the distribution by gender was balanced (50.2% boys, 49.8% girls), with a mean age of 10.7 ± 0.52 years. Regarding socioeconomic level, nearly half of the sample (46.2%) belonged to the medium level, followed by the low level (32.3%). More than half of the children (53.8%) attended public schools, while subsidised and private schools accounted for 23.8% and 22.4%, respectively. In the adolescents’ group, there was a slightly higher proportion of boys (53.9%) compared to girls (46.1%), with a mean age of 13.8 ± 1.39 years. Similar to the children, the medium socioeconomic level was the most prevalent (46.1%). However, the distribution by school type differed significantly from the younger group, with a higher percentage of adolescents attending private schools (40.0%), followed by public (31.7%) and subsidised schools (28.2%).

Figure 1A depicts the network structure of associations between PA, physical condition (PC), and socioecological factors in the school-aged population. Edge weights represent direct associations between variables after controlling for the rest of the network, with coefficients generally ranging from small to moderate magnitude (≈ 0.01–0.16). PC indicators showed greater internal cohesion, with comparatively stronger associations (up to ≈ 0.13), whereas physical activity variables, including school-based MVPA, displayed multiple but weaker links with interpersonal and contextual factors. SPa and SPe were consistently connected across the network, suggesting a bridging role between individual behaviours and contextual influences, while socioeconomic status showed mostly weak associations, indicating a predominantly indirect effect. Overall, the pattern of associations supports a socioecological framework in which PA and PC are shaped by the interplay of multiple factors with cumulative, rather than isolated, effects.

Fig. 1.

Partial correlations between different personal variables associated with moderate-vigorous physical activity (MVPA) in school in children and adolescents. MVPA in school (PA): orange nodes. Physical condition (PC): light blue nodes. Parental Support (SPa): green nodes. Peer support (SPe): yellow nodes. Socioeconomic level (SEL): pink nodes. Kind of school (SCHOOL): red node. MVPA in school (MVPA Sch): white nodes. The blue edges express positive correlations, while the red ones have negative correlations. The thickness of the graph indicates the weight of the ratio. Panel A for children and Panel B for adolescents

Figure 1-B shows that the most associations remained small to moderate, reinforcing the multifactorial nature of the system; however, several comparatively stronger links emerged, particularly within the SPe and SEL domains (up to ≈ 0.24–0.33), indicating tighter clustering among these contextual factors. PC variables again showed moderate internal connectivity (≈ 0.12–0.15) and consistent links with SPa. A small number of negative associations ( ≈ − 0.11) were observed, suggesting potential trade-offs or inverse relationships between specific contextual and behavioural indicators.

Figure 2 shows the centrality measures as Betweenness (frequency of interaction), Closeness (less distance), Strength (correlations) and Expected influence. In this figure, it is possible to observe a similarity between the different nodes compared to children and adolescents. The moderate-vigorous PA in school (MVPA Sch) has been used as a reference, where it stands out the parental support for transport to do PA or sport (SPa3) and PA in lunch (PA3) stand out with high Betweenness in children. According to Closeness, children show higher values in peer support concerning encouragement, to do PA, accompaniment to PA and speaking about PA (SPe1, SPe2, Spe3, Spe4, respectively) to MVPA compared to adolescents. Concerning Strength and Expected Influence, to do PA, accompaniment to PA (SPe2, SPa3), and PA after school (PA5) were higher than those of adolescents.

Fig. 2.

Centrality plot for each variable between children and adolescents in school

Figure 3-A presents that PC indicators show moderate internal connectivity, particularly between PC1, PC3, and PC4 (up to ≈ 0.16), suggesting coherence within this domain. PA variables are connected both among themselves and with school-based MVPA, although these associations are generally modest. SPa and SPe variables form a relatively dense cluster, with several moderate positive associations (≈ 0.12–0.18), highlighting their central role in the network. In contrast, socioeconomic indicators show weaker and more diffuse links, indicating primarily indirect influences. A small number of negative associations were observed, suggesting potential inverse relationships between specific behavioural and contextual variables.

Fig. 3.

Partial correlations between different personal variables associated with moderate-vigorous physical activity (MVPA) outside of school in children and adolescents. MVPA in school (PA): orange nodes. Physical condition (PC): light blue nodes. Parental support (SPa): green nodes. Peer Support (SPe): yellow nodes. Socioeconomic level (SEL): pink nodes. Kind of school (SCHOOL): red node. MVPA in school (MVPA Sch): white nodes. The blue edges express positive correlations, while the red ones have negative correlations. The thickness of the graph indicates the weight of the ratio. Panel A for children and Panel B for adolescents

In Fig. 3-B, edge weights indicate partial associations, with coefficients ranging from small to moderate magnitude (approximately − 0.09 to 0.25), reflecting the complex and interrelated nature of the system. PC variables show moderate internal connectivity, particularly among PC1, PC3, and PC4 (≈ 0.12–0.16), suggesting coherence within this domain. PA indicators are linked both among themselves and with school-based MVPA, with several moderate associations (≈ 0.15–0.19), highlighting the relevance of the school context. SPa and SPe variables form a dense cluster with some of the strongest associations in the network (up to ≈ 0.25), underscoring their central role as interpersonal determinants. In contrast, socioeconomic indicators exhibit weaker and more diffuse connections, indicating primarily indirect influences.

In Fig. 4, we can appreciate that the node, your friends encourage you to do sports (SPe2), has the highest betweenness and closeness in children within the network. In the case of adolescents, general physical condition (PC1), PA during recess (PA2), and PA in the afternoons (PA5) present the highest values of betweenness and closeness.

Fig. 4.

Centrality plot for each variable between children and adolescents outside school

Regarding the expected influence, children show the highest values in “friends do PA with you” (SPe2) and parental support to transport to do sports (SPa3). In the case of adolescents, the main nodes present the highest values of expected influence, which are general physical condition (PC1), “your friends encourage you to do sports” (SPe2), and parents encouraging you to do sports (SPa2).

More details for centrality and clustering measures per variable in MVPA inside and outside school are available in the “supplementary file”.

Discussion

The main results identified three primary opportunities for PA: physical education classes, recess, and lunchtime. Notably, adolescents engage in a greater amount of PA within the school than children. In adolescents, two critical variables were associated with school-based PA within the school: general physical condition level (individual level) and encouragement from parents (interpersonal level).

It was also observed that for children, there is a significant relationship between PA during recess and overall MVPA. Furthermore, children were more physically active outside of school compared to adolescents, especially in the afternoons.

In the school setting, various socio-ecological factors contribute to PA engagement. Research by Smith et al. [48] highlights the significance of recess as a critical microsystem when children engage in moderate to vigorous PA. This underscores the importance of environmental policies incorporating adequate recess periods into the school day to promote active lifestyles among children. Findings from Johnson and Jones [49] indicate that adolescents exhibit increased PA levels during specific structured and unstructured times, suggesting that these opportunities are positively associated with overall PA levels.

Comparatively, children tend to engage in more PA outside of school hours than adolescents [50]. Notably, Brown et al. [51] found that children demonstrate significantly higher levels of PA in the afternoons. This suggests a temporal window where the interaction between the individual and their neighbourhood environment is most active. This finding emphasises the importance of considering the timing and context of PA interventions targeting children outside school hours.

Regarding interpersonal dynamics, parental support emerges as a crucial factor influencing children’s and adolescents’ engagement in PA. In a Chinese study conducted on schoolchildren between 9 and 12 years, significant correlations were found between specific parental behaviours and MVPA, such as parental encouragement, accompaniment, financial backing, involvement, and role modelling in PA [51]. Previously, a study identified that parental logistic support is effective in promoting PA, whereas low logistic support acts as a barrier [52]. In our study, children identified “transportation” (logistic support) as a key factor. This highlights a mesosystem interaction: the parent’s ability to bridge the gap between the home and the sports facility. According to evidence, this logistic support is a significant determinant [53], potentially surpassing the influence of an active parental model alone [54].

However, in adolescents, the complex synergy between higher physical condition and parental encouragement is what primarily drives engagement in PA outside of the school environment [55]. Our findings reveal a pivotal shift in interpersonal dynamics across the life course: while children remain heavily dependent on adult-led logistic support, adolescents begin to prioritise autonomy, responding more effectively to verbal encouragement and the social reinforcement of their peer group. In contrast, children demonstrate higher engagement when active with peers in structured play [56]. This peer-level microsystem plays a vital role in socialisation, as peers provide not only social support but also the necessary motivation to sustain long-term active behaviours [57]. This transition suggests that as children mature, the source of social influence migrates from the family core toward broader social circles.

Finally, we contend that parental influence must be understood as a multidimensional construct. This association is not merely defined by the presence of an active parent model [58], but rather by a sophisticated interplay of direct strategies, such as providing transportation and financial resources, and indirect strategies, including behavioural modelling and emotional encouragement [59]. Current evidence suggests that direct strategies often yield more immediate effects on PA levels. This distinction is vital for the socio-ecological framework, as the relative effectiveness of these parental interventions varies significantly according to the child’s developmental stage [54]. For instance, while logistical support is a prerequisite for children, adolescents may perceive it as a secondary factor compared to the perceived social norms of their environment.

In conclusion, the socio-ecological model illustrates that PA is not the result of isolated variables, but of the continuous and reciprocal interaction between the individual’s internal characteristics and their shifting social and physical environments. Therefore, public health interventions must transcend individual-focused approaches and address these overlapping layers of influence. Specifically, our findings suggest that school-based programs should optimise Physical Education (PE) curricula to transition from mere instruction to fostering peer-support networks, while family-centred interventions should be tailored to the child’s age, prioritising logistical support for younger children and motivational, autonomy-supportive strategies for adolescents. By addressing these distinct entry points, institutional policies can create more sustainable and developmentally appropriate behavioural changes [3].

Strengths and Limitations

This study presents several notable strengths, beginning with the application of network analysis as a novel methodological approach. Unlike traditional linear regressions, this technique allows for a more sophisticated visualisation and examination of the complex interplay between physical condition, social support, and MVPA. Another significant strength is the contextual granularity provided by differentiating between MVPA levels inside and outside the school environment, which identifies specific settings where parental and peer support are most influential. Furthermore, the study utilised a diverse and representative sample of 655 Chilean schoolchildren from various school types and regions, enhancing the relevance of the findings for the national context. The use of validated and culturally adapted instruments, such as the YAP and the IFIS, further ensures the robustness of the data collected within the framework of the ACTIBESE Project.

Despite these strengths, some limitations must be acknowledged. First, the cross-sectional nature of the data precludes the establishment of causal relationships; therefore, the identified associations within the network do not imply directionality. Second, the reliance on self-reported measures for physical activity and fitness may introduce recall or social desirability bias, as objective measures like accelerometry were not utilised. Additionally, while the sample is representative of specific urban areas in Chile, the results may not be fully generalizable to rural populations or other Latin American regions with different socio-environmental characteristics. Finally, although key variables were included, the analysis did not account for other potential confounders such as school infrastructure or neighbourhood safety, which could further explain the dynamics of the identified network.

Conclusion

Promoting PA among children and adolescents requires a multifaceted approach considering both school and outside school environments. Strategies aimed at increasing PA levels should leverage the support of parents and peers while also addressing the unique needs and preferences of each age group. By fostering supportive environments both within and outside of the school setting, policymakers, educators, and parents can collectively contribute to the promotion of active lifestyles among children and adolescents.

Abbreviations

EBIC

Extended Bayesian Information Criterium

FAS III

Family Affluence Scale

IFIS

International Fitness Scale

MVPA

Moderate-vigorous physical activity

PA

Physical activity

PC

Physical condition

SEL

Socioeconomic level

Spa

Parental support

Spe

Peer support

YAP

Youth Activity Profile

Authors’ contributions

The study concept and design were conceived by FR-R assisted in refining the study questionnaires and study design by CM. FRR and NA-F handle data collection. Analyses will be conducted by FR-R, CM, JM and NA-F. FR-R and CM prepared the first draft of the manuscript. JM and NA-F critically and deeply revised the manuscript. All authors revised and approved the submitted version of the manuscript.

Funding

This project has been financed by the Fondo Nacional de Investigación para el Desarrollo Científico y Tecnológico, FONDECYT of the Agencia Nacional de Investigación y Desarrollo (ANID) of the Chilean Government. FONDECYT Regular project 2023, Nº 1230801.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Also, other materials are published on www.actibese.cl.

Declarations

Ethics approval and consent to participate

Once the respective schools have authorised the project, teachers and parents must sign an informed consent form to participate. Likewise, the students signed an informed consent form that described the study’s characteristics, aims, and activities. At the same time, parents’ children signed an informed consent to authorise the participation of their offspring. This project has been approved by the Bioethics Committee of the Pontificia Universidad Católica de Valparaíso, code BIOEPUCV-H 638–2023.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.