The association between social media content production and adolescent physical health promotion: the chain mediating roles of online social support and positive psychological capital

Abstract

Background

This study aimed to examine the association between social media content production and adolescent physical health promotion, as well as the potential mediating and sequential mediating roles of online social support and positive psychological capital.

Methods

A cross-sectional survey was conducted using self-report questionnaires from 3,442 adolescents recruited through a multi-stage sampling procedure across seven major geographic regions in China. Structural equation modeling and bootstrap analysis (5,000 resamples, 95% confidence intervals) were performed.

Results

Social media content production, online social support, and positive psychological capital were all significantly and positively associated with adolescent physical health promotion. Online social support and positive psychological capital showed independent and sequential mediating roles. The indirect associations via online social support (β = 0.231, 95% CI [0.204, 0.263]), positive psychological capital (β = 0.117, 95% CI [0.090, 0.145]), and their sequential association (β = 0.101, 95% CI [0.084, 0.118]) were all statistically significant.

Conclusion

This study indicates that social media content production is associated with adolescent physical health promotion through psychosocial factors. However, causal inference cannot be established due to the cross-sectional design. It provides a theoretical basis and practical implications for understanding the psychological and social determinants of adolescent health behaviors and for effectively leveraging social media to support adolescent physical health promotion in the digital era.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40359-026-05144-7.

Introduction

Adolescent physical health has become a major concern in global public health. According to the World Health Organization (WHO), more than 80% of adolescents worldwide do not meet the recommended levels of physical activity, and the risks of obesity and chronic diseases associated with prolonged sedentary behavior are increasingly evident at younger ages [1]. To address this challenge, the Chinese government has introduced a series of policies, including the Healthy China 2030 Plan, which emphasizes the development of a diversified system for adolescent physical health promotion [2]. However, in the context of rapid digitalization, traditional forms of physical activity are increasingly constrained by limitations of time and space, as well as limited interactivity, leading to a gradual decline in their reach and engagement [3]. With the advancement of mobile internet technologies, social media has become deeply embedded in adolescents’ daily lives. It has long been regarded as a contributing factor to adverse outcomes such as vision decline, increased sedentary behavior, and heightened psychological anxiety among adolescents. Nevertheless, social media also provides efficient channels for social connection and information dissemination, thereby offering opportunities for positive health interventions. Against this backdrop, examining how social media is associated with adolescent physical health promotion is of considerable practical significance.

At present, the relationship between social media content production and adolescent physical health promotion remains inconclusive. Existing studies have primarily focused on the negative consequences of passive exposure to social media content, such as internet addiction and social anxiety [4, 5]. However, other studies have indicated that individuals’ active content production behaviors, including posting exercise updates and sharing workout experiences, may be associated with health-related behaviors such as physical activity [6, 7]. According to social cognitive theory [8], individuals’ active interaction with their environment plays a central role in behavior change. In this context, social media content production represents not only a digital manifestation of individuals’ physical activity experiences but also an active form of health management behavior. In addition, engagement with professionally generated content may serve as an important entry point for participation, as exposure to such content can stimulate interest, facilitate interaction, and encourage subsequent content sharing and social engagement. Therefore, in the present study, social media content production refers to participatory engagement with sports-related social media content, including active content creation, interaction, and exposure to professionally generated content that may stimulate subsequent engagement, rather than being limited to content creation in the narrow sense. These findings suggest that the relationship between social media and adolescent physical health promotion is inherently complex. The possible explanatory associations have yet to be systematically examined, and these associations may vary according to additional psychosocial variables.

Online social support and positive psychological capital, as important psychosocial factors, are closely associated with physical health behaviors. As a key medium of social interaction, social media inherently facilitates frequent social engagement through content production behaviors [9]. According to social support theory, when adolescents share sports-related content on social media platforms, the likes, comments, and reposts they receive constitute typical forms of online social support [10]. Empirical studies have demonstrated that such affirmation and encouragement from virtual communities provide emotional support and informational resources and may be linked to greater willingness to participate in physical activity and greater adoption of health-related behaviors [11]. Furthermore, exposure to professionally generated sports content may also contribute to perceived online social support when it leads to interaction, feedback, or participation in online communities, thereby extending the pathways through which social media engagement is associated with social support. Positive psychological capital refers to an individual’s overall positive psychological state, typically encompassing self-efficacy, hope, optimism, and resilience [12]. In addition, the triadic reciprocal determinism proposed in social cognitive theory posits that human functioning is shaped by the interaction among environment, individual, and behavior and that external environmental support serves as an important source of positive psychological states [13]. Exposure to professionally generated sports content may also provide opportunities for observational learning and role modeling, which are associated with the development of self-efficacy, hope, and other components of positive psychological capital. Both online social support and positive psychological capital may therefore serve as mediators in the relationship between social media content production and adolescent physical health promotion. Accordingly, this study proposes a chain mediation model to examine the possible mediating associations linking social media content production with adolescent physical health promotion, aiming to provide a theoretical basis and practical implications for effectively leveraging social media resources to support adolescent physical health promotion in the digital era.

Theoretical framework and hypothesis development

Social media content production and adolescent physical health promotion

Social media platforms have become key spaces for adolescents to exchange information, establish social connections, and express themselves. Participation in sport-related social media content includes sharing health-related information, documenting physical activity, and engaging in interactive feedback, as well as actively using professionally generated sports content and guidance. According to social cognitive theory and health communication theory, individuals’ participatory engagement with online health content through creation, interaction, observation, and the use of professionally generated guidance may be associated with better health cognition, stronger behavioral intentions, and greater adoption of health behaviors. Empirical studies have shown that social media content production is not only related to the dissemination of sports knowledge and skills [14, 15] but is also associated with higher levels of adolescents’ daily physical activity through practices such as exercise check-ins and community interactions [16] and may be linked to more positive lifestyles and better physical and mental health, as well as social adaptability [17]. Research on the WeChat platform has demonstrated that health education delivered via official accounts has been associated with better learning effectiveness and exercise behaviors among college students [18, 19]. Similarly, the use of fitness applications has been found to be associated with higher motivation and persistence in physical activity [20]. Previous empirical studies have suggested that sport-related social media participation may be associated with health-related cognition, motivation, and behavioral tendencies. For example, activities such as sharing exercise records, participating in online exercise check-ins, exchanging sport-related information, and using fitness applications have been linked to physical activity motivation, exercise persistence, and healthier lifestyles [21]. These findings indicate that social media content production should not be understood merely as online self-expression but also as a form of health-oriented digital participation that may be associated with adolescents’ self-reported physical health promotion.

Taken together, social media content production may be positively associated with adolescent physical health promotion. Based on this, the study proposes the following hypothesis:

• H1: Social media content production is significantly and positively associated with adolescent physical health promotion

Mediating role of online social support

Online social support refers to multidimensional supportive resources obtained in online environments, including informational, emotional, companion, and instrumental support. It represents an important extension and complement to real-world social support [22, 23]. Research has shown that adolescents can gain a sense of recognition and belonging from peers and online communities through active self-disclosure and interactive participation on social media, thereby alleviating feelings of isolation and stress in offline social interactions. Peris highlighted that healthy online interpersonal relationships can positively interact with real-world social support, jointly promoting individual adaptation and social integration [24]. In addition, online social support has been shown to significantly and positively predict adolescents’ psychological well-being and social adaptability and to serve as a mechanism for promoting health-related behaviors [25]. According to a report by the American Academy of Pediatrics, social media allows adolescent users to strengthen existing friendships and establish new connections online, which reduces social isolation and loneliness and indirectly enhances mental health [26].

Adolescents, as “digital natives,” are proficient in using social media to access and share health-related information, and their health behaviors are often strongly influenced by online social support. For example, sharing exercise updates among peers, along with receiving likes and comments, can provide emotional encouragement and social recognition, thereby enhancing individuals’ motivation to engage in physical activity. Online exercise communities established through social media can also offer instrumental support and monitoring functions, promoting adherence to regular exercise [27]. Furthermore, online social support can indirectly facilitate the adoption and maintenance of health behaviors by buffering real-world stress and enhancing psychological resources [28]. Accordingly, the association between social media content production and online social support may arise from adolescents’ participatory engagement with sport-related social media content, including both peer interaction around user-generated content and engagement with professionally generated content that facilitates feedback, information exchange, or participation in online communities. When adolescents share sport-related posts, exercise records, or health-related experiences on social media, likes, comments, reposts, and peer feedback may provide emotional encouragement, informational resources, companionship, and practical suggestions. These responses make health-related behaviors visible and socially reinforced, thereby linking sport-related content production with perceived online social support. Based on this rationale, the present study proposes the following hypothesis:

• H2a: Social media content production is significantly and positively associated with online social support.

• H2b: Online social support is significantly and positively associated with adolescent physical health promotion.

• H2c: Online social support plays a mediating role in the association between social media content production and adolescent physical health promotion.

Mediating role of positive psychological capital

Positive psychological capital refers to an individual’s positive psychological state developed during growth, encompassing dimensions such as self-efficacy, hope, optimism, and resilience [29]. Research indicates that adolescents’ positive psychological capital is not only closely associated with their psychological well-being but also effectively predicts their engagement in adolescent physical health promotion behaviors [30]. Studies on internet use and behavior have further shown that positive psychological capital is significantly related to factors such as cyberbullying, online interpersonal trust, and internet addiction and may serve as a mediator between online activity and health-related behaviors. For instance, adolescents with higher self-efficacy are more likely to actively explore and utilize online resources to support health behaviors, while resilience facilitates the maintenance of exercise habits when facing challenges [31].

Social media content production provides adolescents with opportunities for self-expression, social feedback, observational learning, and access to professional exercise guidance, which may be associated with the development of a healthier self-image and positive psychological resources. Engagement in the creation, interaction, and use of sport-related social media content may represent an active process of self-construction and capability validation, which contributes to enhancing their self-efficacy and hope. Empirical research has demonstrated that positive psychological interventions delivered via the WeChat platform can effectively improve college students’ hope and optimism [18]. These findings indicate that active participation and content creation on social media may foster the development and maintenance of health behaviors by strengthening individuals’ positive psychological capital. Furthermore, a growing body of research has shown that positive psychological capital is closely related to health-related behaviors. Adolescents with higher self-efficacy may be more confident in setting and maintaining exercise goals; those with higher hope may be more likely to identify pathways for achieving health goals; optimism may support positive expectations toward health behavior change; and resilience may help adolescents maintain exercise habits when facing barriers or setbacks. Therefore, positive psychological capital provides an important internal psychological resource for understanding adolescents’ physical health promotion [32]. Accordingly, the present study proposes the following hypothesis:

• H3a: Social media content production is significantly and positively associated with positive psychological capital.

• H3b: Positive psychological capital is significantly and positively associated with adolescent physical health promotion.

• H3c: Positive psychological capital plays a mediating role in the association between social media content production and adolescent physical health promotion.

The sequential mediating roles of online social support and positive psychological capital

There is a close relationship between online social support and positive psychological capital. On one hand, online social support is associated with adolescents’ higher self-efficacy and sense of belonging by providing informational, emotional, and companion support and may be linked to the development of their positive psychological capital [33]. On the other hand, individuals with higher levels of positive psychological capital are more capable of actively seeking and effectively utilizing social support resources, forming a positive feedback loop between support acquisition and psychological strengths [34]. Previous research has demonstrated that social support, including online support, serves as an important psychological buffer under stress and can be indirectly associated with health behaviors by enhancing levels of psychological capital [35, 36]. Related research has indicated that social support mediates the influence of online social support on individuals’ subjective well-being, suggesting that online and offline support are jointly related to psychological resources and behavioral outcomes [28].

Accordingly, it can be inferred that social media content production, including both active content production and content-related engagement, may first be associated with higher perceived online social support, which is subsequently associated with higher positive psychological capital and is further related to the adoption and maintenance of physical health behaviors. This sequential association provides a possible explanatory pattern, progressing from the acquisition of external resources to internal psychological construction and, finally, to behavioral expression. This ordering is primarily theory-driven. Drawing on social cognitive theory and social support theory, this study conceptualizes online social support as an external social resource and positive psychological capital as an internal psychological resource. However, given the cross-sectional nature of the data, the proposed sequence should be understood as a theoretically informed statistical model rather than an empirically confirmed temporal order. Based on this rationale, the present study proposes the following hypothesis:

• H4a: Online social support is significantly and positively associated with positive psychological capital.

• H4b: Online social support and positive psychological capital play sequential mediating roles in the association between social media content production and adolescent physical health promotion.

In summary, to ensure consistency between the hypothesis system and the structural model, building on previous research, this study introduces online social support and positive psychological capital as mediating factors between social media content production and adolescent physical health promotion and constructs a sequential mediation model (see Fig. 1).

Fig. 1.

Hypothetical model

Participants and methods

Research participants

This study employed a cross-sectional survey design, and the minimum sample size was calculated using the population proportion formula [37]. With a 95% confidence level, a margin of error of 0.03, and an expected prevalence of 0.5, the theoretical minimum sample size was 1,067 participants. However, because this study further involved structural equation modeling and mediation analysis, the adequacy of sample size was not evaluated solely on the basis of the population proportion formula. The need for stable parameter estimation, bootstrap mediation testing, and the potential influence of invalid questionnaires and missing data were also considered. Therefore, the target sample size was set at over 1,500 participants [38]. The final analytic sample included 3,442 valid participants, which exceeded both the minimum sample size estimated for the cross-sectional survey and the sample size requirement for conducting structural equation modeling with stable parameter estimation.

To improve geographic coverage of the sample, this study adopted the commonly used seven-region framework of China and used the population distribution from the Seventh National Population Census as a reference for regional quotas [39]. The sampling procedure was a multi-stage design combining regional quota sampling, school-level convenience sampling, and class-level cluster sampling. Specifically, the survey covered seven major geographic regions: East China, Central China, Southwest China, South China, North China, Northwest China, and Northeast China. One secondary school was selected from each region based on accessibility, willingness to participate, and administrative feasibility. The selected sites were Dezhou No. 9 Middle School in Shandong Province for East China; Zhengzhou No. 4 Middle School in Henan Province for Central China; Pidu No. 1 Middle School in Chengdu, Sichuan Province, for Southwest China; Jiahe Xindu School in Guangzhou, Guangdong Province, for South China; the High School Affiliated to Renmin University of China Beijing Economic-Technological Development Area School in Beijing for North China; Baotou No. 6 Middle School in Inner Mongolia for Northwest China; and Dongfeng No. 3 Middle School in Jilin Province for Northeast China.

The number of questionnaires distributed in each region was determined with reference to the population proportion of the corresponding region in the Seventh National Population Census. A total of 3,996 questionnaires were distributed: 1,240 in East China, 620 in Central China, 568 in Southwest China, 540 in South China, 468 in North China, 288 in Northwest China, and 272 in Northeast China. Within each participating school, intact classes were surveyed as clusters where administrative arrangements allowed. A total of 3,897 questionnaires were returned, and 3,442 valid questionnaires were retained after excluding invalid responses, yielding an overall valid response rate of 86.1% based on the number distributed. Questionnaires were excluded if they contained substantial missing responses, patterned or repetitive answers, clearly inconsistent responses, or responses from participants outside the target age range. Participants were aged 13–19 years, with a mean age of 16.3 years (SD = 1.6). Detailed information on the regional distribution of the sample, including provinces or municipalities, participating schools, and the number of questionnaires distributed and retained, is provided in Supplementary Material 1.

The study was approved by the Ethics Committee of Qufu Normal University (Approval No.: 2021098) and was conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from the schools, students, and their parents prior to data collection.

Research instruments

All questionnaires were administered in Chinese. Responses were scored on a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree, with higher scores indicating higher levels of the corresponding construct. Reverse-coded items, where applicable, were recoded before calculating the total or dimension scores. The full items, dimensions, scoring procedures, and reverse-coded items of the questionnaires are provided in Supplementary Material 1. The original sources of the scales were cited, and all measurement instruments were used or adapted for non-commercial academic research in accordance with their original publications.

Adolescent social media content production questionnaire

The “Adolescent Social Media Content Production Questionnaire” [40] was employed to assess adolescents’ engagement with sports-related content on social media. This questionnaire was developed in the authors’ previous work and was designed specifically for adolescent sport-related social media contexts. It comprises nine items across two dimensions: user-generated content (UGC; five items) and professionally generated content (PGC; four items). The UGC dimension reflects active content production and interaction, such as sharing personal exercise records and commenting on peers’ exercise-related posts. The PGC dimension reflects engagement with professionally generated sport-related content, such as receiving professional sports advice, following sport bloggers, or using fitness apps. All items were rated on a five-point Likert scale, with higher scores indicating higher levels of sport-related social media content production. No reverse-coded items were included in this questionnaire.

In this study, the questionnaire demonstrated good internal consistency (Cronbach’s α = 0.883). Confirmatory factor analysis indicated a good model fit (χ²/df = 4.352, GFI = 0.993, CFI = 0.994, TLI = 0.992, RMSEA = 0.031), with standardized factor loadings ranging from 0.707 to 0.829. Composite reliability values exceeded 0.85, and average variance extracted values were greater than 0.59, demonstrating satisfactory reliability and construct validity. The three-week test-retest reliability was acceptable (r = 0.792) and was retained only as supplementary evidence of measurement stability.

Adolescent physical health promotion questionnaire

The “Adolescent Physical Health Promotion Questionnaire” [41] was employed to assess adolescents’ self-reported physical health promotion, including cognition, affect, and behaviors related to health-promoting activities. In this study, adolescent physical health promotion refers to a self-reported construct reflecting health-related cognition, affect, and behavioral tendencies, rather than objective physical health status, physical fitness, or clinical health indicators. The questionnaire consists of 28 items across three dimensions: physical health cognition (9 items), physical health affect (9 items), and physical health behavior (10 items). All items were rated on a five-point Likert scale, with higher scores indicating higher levels of adolescent physical health promotion. Reverse-coded items, where applicable, were recoded before score calculation.

In this study, the questionnaire demonstrated excellent internal consistency, with an overall Cronbach’s α of 0.958 and α coefficients for the three dimensions ranging from 0.926 to 0.945. Confirmatory factor analysis indicated a good model fit (χ²/df = 3.159, GFI = 0.978, CFI = 0.989, TLI = 0.988, RMSEA = 0.025), with standardized factor loadings ranging from 0.641 to 0.923. Composite reliability values exceeded 0.92, and average variance extracted values were all above 0.58, indicating that the questionnaire exhibited strong reliability and construct validity. The three-week test-retest reliability was acceptable (r = 0.832) and was retained only as supplementary evidence of measurement stability.

Adolescent online social support questionnaire

The “Adolescent Online Social Support Questionnaire” [23] was employed to measure adolescents’ perceived social support obtained from online environments. The original questionnaire comprised four dimensions: informational support (5 items), companion support (8 items), emotional support (6 items), and instrumental support (4 items), totaling 23 items. In the present study, the original items were retained, and one sport-related item (“The social support I receive online is helpful for my physical exercise”) was added to improve the contextual relevance of the scale for sport and physical health promotion research, resulting in a revised version with 24 items.

This added item was classified under the instrumental support dimension because it reflects the perceived practical usefulness of online support for physical exercise. The rationale for adding this item was that the present study focused on the association between sport-related social media engagement and adolescent physical health promotion; therefore, it was necessary to capture whether online social support was perceived as practically relevant to physical exercise. The revised questionnaire was reviewed during the instrument development process and pilot tested among middle school students before the formal survey. Follow-up interviews were also conducted with a subsample of students to examine item clarity and contextual appropriateness. All items were rated on a five-point Likert scale, with higher scores indicating higher levels of perceived online social support.

In this study, the revised questionnaire demonstrated excellent internal consistency, with an overall Cronbach’s α of 0.96 and α coefficients for each dimension ranging from 0.892 to 0.922. Confirmatory factor analysis indicated a good model fit (χ²/df = 4.940, GFI = 0.970, CFI = 0.981, TLI = 0.979, RMSEA = 0.034), with standardized factor loadings ranging from 0.628 to 0.910. Composite reliability values exceeded 0.87, and average variance extracted values were all above 0.58, demonstrating that the revised questionnaire had strong reliability and construct validity in this sample. The three-week test-retest reliability was satisfactory (r = 0.832) and was retained only as supplementary evidence of measurement stability.

Adolescent positive psychological capital questionnaire

The “Positive Psychological Capital Scale” [42] was employed to measure adolescents’ levels of positive psychological capital. The scale consists of four dimensions: self-efficacy (7 items), resilience (7 items), hope (6 items), and optimism (6 items), with a total of 26 items. Responses are scored on a five-point Likert scale, with higher total scores indicating higher levels of positive psychological capital. Reverse-coded items were recoded before calculating the total and dimension scores.

In the present study, the scale demonstrated excellent internal consistency, with an overall Cronbach’s α of 0.96 and α coefficients for each dimension ranging from 0.848 to 0.917. Confirmatory factor analysis indicated a good model fit (χ²/df = 4.697, GFI = 0.969, CFI = 0.980, TLI = 0.978, RMSEA = 0.033), with standardized factor loadings ranging from 0.627 to 0.912. Composite reliability values exceeded 0.85, and average variance extracted values were all above 0.49. Although the average variance extracted value for the hope dimension was 0.493, slightly below the conventional threshold of 0.50, it was considered acceptable because the corresponding composite reliability value met the recommended standard. These results indicate that the scale exhibited strong reliability and construct validity in this sample.

Quality control

Prior to formal data collection, the research team provided standardized training for all survey administrators, covering the survey procedures, administration protocols, and standardized explanations of key questionnaire variables to ensure consistency and rigor in implementation. During the survey, classroom order was maintained with the assistance of homeroom or subject teachers. Survey administrators were responsible for on-site supervision, addressing participants’ questions, collecting the questionnaires, and verifying the completeness of each response. The study employed an anonymous questionnaire method, with questionnaires distributed, completed, and collected on-site, and the completion time was limited to 30 min. In addition, participants were informed that their responses would be used solely for research purposes and that there were no right or wrong answers, which may help reduce social desirability bias and potential common method bias.

Statistical analysis

The raw data were compiled using EpiData 3.1, and all entries were performed using a double-entry procedure to ensure accuracy. The dataset was subsequently processed in SPSS 25.0 and AMOS 23.0 for data cleaning, descriptive statistics, correlation analysis, structural equation modeling, and bootstrap analyses. Mediating roles were assessed using the Bootstrap method with 5,000 resamples, and statistical significance was set at p < 0.05.

In the structural equation model, the latent variables were represented by theoretically defined dimension-level indicators rather than by all individual questionnaire items. This approach was adopted to retain the conceptual structure of each scale while reducing model complexity and improving estimation stability. Specifically, social media content production was represented by user-generated content and professionally generated content; online social support was represented by informational support, companion support, emotional support, and instrumental support; positive psychological capital was represented by self-efficacy, resilience, hope, and optimism; and adolescent physical health promotion was represented by physical health cognition, physical health affect, and physical health behavior. Although dimension-level averaged indicators were used, measurement error was not ignored, because each dimension-level indicator was modeled as an observed indicator of its corresponding latent construct with an associated residual term.

Results and analysis

Test for common method bias

All questionnaire data in this study were collected from adolescents, which may introduce common method bias. Therefore, Harman’s single-factor test was conducted to assess the presence of common method bias. The results indicated that the first principal factor with an eigenvalue greater than 1 accounted for 31.424% of the total variance, which is below the critical threshold of 40%. This finding suggests that common method bias may not be dominant in the present study.

Descriptive statistics of the sample

The basic demographic information of the surveyed adolescents, including their gender, age, and geographic region, is presented in Table 1.

Table 1.

Descriptive analysis of basic information of adolescents

Attribute	Category	Number (n)	Percentage (%)
Gender	Male	1785	51.9
	Female	1657	48.1
Age (years)	13	263	7.6
	14	314	9.1
	15	389	11.3
	16	801	23.3
	17	874	25.4
	18	514	14.9
	19	287	8.3
Geographic Region	North China	409	11.9
	Northeast China	227	6.6
	East China	1065	30.9
	Central China	554	16.1
	South China	484	14.1
	Southwest China	483	14.0
	Northwest China	220	6.4

Correlation analysis

Pearson correlation analyses were conducted to examine the relationships among social media content production, online social support, positive psychological capital, and physical health promotion. To further assess discriminant validity within the integrated construct framework, the Fornell-Larcker criterion was applied by comparing the square root of the average variance extracted for each latent construct with its correlations with other constructs [43]. As shown in Table 2, all variables were significantly and positively correlated, indicating meaningful associations among the constructs. More importantly, the square root of the average variance extracted for each latent construct exceeded its correlations with the other constructs. This result suggests that the four latent constructs were empirically distinguishable from one another and supports the appropriateness of including them simultaneously in the subsequent structural equation model.

Table 2.

Results of correlation analysis and discriminant validity assessment

Variables	Social Media Content Production	Online Social Support	Positive Psychological Capital	Physical Health Promotion
Social Media Content Production	0.769
Online Social Support	0.481**	0.781
Positive Psychological Capital	0.397**	0.480**	0.807
Physical Health Promotion	0.517**	0.635**	0.637**	0.778

**p < 0.01

Structural equation modeling analysis

Model evaluation

Structural equation modeling offers a robust method for examining both direct associations and mediating roles among variables. To investigate the possible mediating associations linking social media content production with adolescents’ physical health promotion, this study developed a mediation-based structural equation model grounded in the theoretical framework. The model included online social support and positive psychological capital as potential mediators to test whether they jointly mediate the relationship between social media content production and adolescents’ physical health promotion, while also retaining the direct path from social media content production to physical health promotion. The initial structural model was estimated using AMOS 23.0, as shown in Fig. 2.

Fig. 2.

Initial structural equation model

As shown in Table 3, the CMIN/DF value was 5.33 (> 5), exceeding the acceptable range and indicating a poor model fit. Therefore, modifications were required to improve the model’s correspondence with the observed data.

Table 3.

Initial structural model fit

Model Fit Indicators	Optimal Threshold Value	Observed Values	Fit Assessment
CMIN	——	314.477	——
DF	——	59.000	——
CMIN/DF	< 5.00	5.330	Poor
RMR	< 0.08	0.017	Good
GFI	> 0.90	0.986	Good
AGFI	> 0.90	0.979	Good
NFI	> 0.90	0.985	Good
IFI	> 0.90	0.987	Good
TLI	> 0.90	0.983	Good
CFI	> 0.90	0.987	Good
RMSEA	< 0.08	0.035	Good

Model modification

Model modification was conducted cautiously to avoid purely data-driven overfitting. Modification indices were used only as diagnostic references, and residual covariances were considered only when they had substantive justification. The following principles were applied: (1) residual covariances between indicators belonging to different latent variables were not added without clear theoretical justification; (2) residual terms were not allowed to correlate with their corresponding latent variables; and (3) residual covariances were considered only between indicators within the same latent construct or with closely related substantive content.

Based on these principles, only one residual covariance was added to the final revised model, namely the covariance between e3 and e6. In the revised model, e3 corresponded to instrumental support and e6 corresponded to informational support, both of which were observed dimensions of online social support rather than individual questionnaire items. Allowing their residuals to correlate was considered substantively reasonable because instrumental support and informational support both reflect online supportive resources related to practical guidance, information acquisition, and exercise-related advice. This shared dimension-specific variance may not be fully captured by the general online social support factor, as shown in Fig. 3.

Fig. 3.

Revised structural equation model

The modification was made sequentially. After the residual covariance between e3 and e6 was added, the model was re-estimated, and no additional residual covariance was introduced into the final model.

Model fit indices

As shown in Table 4, the CMIN/DF value was 4.472, which falls below the recommended threshold of 5. Furthermore, AGFI, GFI, NFI, TLI, IFI, and CFI all exceeded the criterion of 0.90. The RMR value was 0.016, which is below 0.08, and the RMSEA value was 0.032, also below 0.08. Overall, after the addition of the residual covariance between e3 and e6, the revised model showed improved fit indices and met commonly accepted criteria, indicating an acceptable fit to the data.

Table 4.

Revised structural model fit

Model Fit Indicators	Optimal Threshold Value	Observed Values	Fit Assessment
CMIN	——	259.366	——
DF	——	58.000	——
CMIN/DF	< 5.00	4.472	Good
RMR	< 0.08	0.016	Good
GFI	> 0.90	0.989	Good
AGFI	> 0.90	0.982	Good
NFI	> 0.90	0.987	Good
IFI	> 0.90	0.990	Good
TLI	> 0.90	0.987	Good
CFI	> 0.90	0.990	Good
RMSEA	< 0.08	0.032	Good

Structural equation modeling of path coefficients

As shown in Table 5, social media content production was significantly and positively associated with online social support (β = 0.628, p < 0.001) and positive psychological capital (β = 0.278, p < 0.001). Online social support was also significantly and positively associated with positive psychological capital (β = 0.379, p < 0.001).

Table 5.

Structural path coefficients

Associations	Standardized Coefficients	Unstandardized Coefficients	S.E.	C.R.	p	Hypotheses
Social Media Content Production → Online Social Support	0.628	0.582	0.025	23.564	***	Supported
Social Media Content Production → Positive Psychological Capital	0.278	0.306	0.032	9.461	***	Supported
Online Social Support → Positive Psychological Capital	0.379	0.450	0.032	13.961	***	Supported
Social Media Content Production → Adolescent Physical Health Promotion	0.255	0.265	0.026	10.290	***	Supported
Online Social Support → Adolescent Physical Health Promotion	0.368	0.413	0.026	15.978	***	Supported
Positive Psychological Capital → Adolescent Physical Health Promotion	0.422	0.398	0.019	20.890	***	Supported

***p < 0.001

In addition, social media content production showed a significant positive association with adolescent physical health promotion (β = 0.255, p < 0.001). Online social support (β = 0.368, p < 0.001) and positive psychological capital (β = 0.422, p < 0.001) were also significantly and positively associated with adolescent physical health promotion.

Taken together, these results indicate that the proposed model demonstrates a coherent pattern of associations, with both online social support and positive psychological capital being significantly related to adolescent physical health promotion.

As shown in Table 6, the bootstrap analysis further clarified the direct, total, and indirect associations in the structural model. The total association between social media content production and adolescent physical health promotion was significant (β = 0.704, 95% CI [0.661, 0.745]). After online social support and positive psychological capital were included as mediators, the direct association remained significant (β = 0.255, 95% CI [0.197, 0.309]), indicating a partial mediation pattern.

Table 6.

Bootstrap results for direct and indirect associations

Associations	Standardized Effect	Proportion (%)	95% CI (Lower)	95% CI (Upper)
Total association	0.704	100	0.661	0.745
Direct association	0.255	36.222	0.197	0.309
Total indirect association	0.449	63.778	——	——
Social Media Content Production → Online Social Support → Adolescent Physical Health Promotion	0.231	32.812	0.204	0.263
Social Media Content Production →Positive Psychological Capital → Adolescent Physical Health Promotion	0.117	16.619	0.090	0.145
Social Media Content Production → Online Social Support → Positive Psychological Capital → Adolescent Physical Health Promotion	0.101	14.347	0.084	0.118

Bootstrap confidence intervals were based on 5,000 resamples. An indirect effect was considered significant when the 95% confidence interval did not include zero. Confidence intervals are reported for the specific indirect effects

The specific indirect associations were then examined separately. First, the indirect association through online social support was significant (β = 0.231, 95% CI [0.204, 0.263]), accounting for 32.812% of the total association. Second, the indirect association through positive psychological capital was significant (β = 0.117, 95% CI [0.090, 0.145]), accounting for 16.619% of the total association. Third, the sequential indirect association through online social support and positive psychological capital was also significant (β = 0.101, 95% CI [0.084, 0.118]), accounting for 14.347% of the total association. Because the bootstrap 95% confidence intervals for all three specific indirect associations did not include zero, these indirect associations were statistically significant.

The total indirect association was 0.449, accounting for 63.778% of the total association. This proportion refers to the combined indirect association through all mediating variables rather than the sequential indirect association alone. These findings indicate that online social support and positive psychological capital function as both independent and sequential mediators in the association between social media content production and adolescent physical health promotion.

Discussion

This study examined the association between social media content production and adolescent physical health promotion by integrating online social support and positive psychological capital into a theory-driven statistical mediation model. The findings extend previous research by shifting attention from general social media use to sport-related content production and by highlighting the joint relevance of external online support and internal positive psychological resources in adolescent physical health promotion. In the digital media context, these results provide a more focused understanding of how sport-related social media engagement is associated with adolescents’ self-reported health-promoting tendencies.

It should be noted that the operationalization of social media content production in this study reflects a participatory perspective, encompassing both active content creation and engagement with professionally generated content. Although browsing behavior is traditionally considered a form of content consumption, it may function as a precursor to interaction and content production by stimulating interest, facilitating social feedback, and promoting participation in online communities. This integrated approach provides a broader understanding of how different forms of social media engagement are associated with psychosocial resources and health-related behaviors. It also helps justify the inclusion of both UGC and PGC dimensions in the measurement framework.

Social media content production is associated with adolescent physical health promotion

The results indicate that social media content production was positively associated with adolescent physical health promotion. This finding further supports the value of distinguishing sport-related content production from general social media use or passive exposure. In this study, content production included not only sharing exercise records and interacting with peers but also engagement with professionally generated sports content, which may provide adolescents with exercise knowledge, social feedback, and opportunities for self-expression.

This finding aligns with previous studies [44]. As social media and digital platforms become increasingly embedded in adolescents’ daily lives, they have become important contexts associated with adolescents’ health perceptions, participation in physical activity, and social interactions. Social media has transformed traditional information dissemination, enabling adolescents not only to receive information but also to produce, share, and interact with content. In the context of physical health, this participatory and interactive engagement may be related to the spread of health knowledge, the sharing of exercise experiences, and the reinforcement of physical activity behaviors [45]. At the same time, the effects of social media are dual in nature. Positive studies suggest that participatory engagement with sport-related social media content may be linked to adolescents’ self-identity, sense of belonging, and social adaptation through self-expression, experience sharing, and interactive feedback [46] and may be linked to better psychological well-being and health behaviors. However, other studies indicate that social media use is not universally beneficial; excessive use may be linked to sedentary behavior, addiction, social comparison, anxiety, and depression [47]. For example, research has found a significant positive association between social networking site addiction and depression, with rumination mediating this relationship and self-esteem further moderating the process [48].

Therefore, the positive association between social media content production and adolescent physical health should be interpreted dialectically. Unlike general social media use, participatory engagement with sport-related social media content in the present study is health-oriented and action-focused. It may provide adolescents with opportunities to access exercise information, accumulate supportive resources, and reinforce positive experiences within digital contexts, which may be associated with more stable tendencies toward health promotion [45]. In this regard, social media may function as both a medium for disseminating adolescent physical health information and a social context for supporting health behaviors and positive psychological resources.

Online social support as a mediator in the association between social media content production

This study found that social media content production was positively associated with online social support and that online social support was further associated with adolescent physical health promotion. This suggests that online social support may be an important social resource linking sport-related content production with adolescents’ health-promoting tendencies.

In the context of social media, adolescents’ participatory engagement with sport-related social media content, including production, sharing, interaction, and active engagement with professionally generated guidance, may provide opportunities to obtain informational, emotional, companion, and instrumental support. Therefore, social media functions not only as a platform for disseminating sports information but also as a key digital context for adolescents to access social support. Previous studies have highlighted a bidirectional relationship between adolescents’ social media posting behaviors and their perceived social support, with content sharing and interactive feedback continuously reinforcing their sense of being supported and socially connected [49]. Moreover, the Internet and social media have reshaped adolescents’ social interactions, allowing social support to transcend temporal and spatial boundaries and exhibit greater immediacy, visibility, and interactivity [50].

More broadly, online communication can extend adolescents’ social networks, broaden the scope of interpersonal interaction, and help maintain real-life social ties [51, 52]. For adolescents, who are in a critical period of identity formation and peer relationship development, online environments may meet their needs for information acquisition, emotional expression, peer recognition, and social connection. Therefore, participatory engagement with sport-related social media content should not be understood merely as an individual act of self-expression but also as a social practice embedded in interaction and support exchange.

In the context of physical health promotion, online social support may be particularly relevant because participatory engagement with sport-related social media content can make exercise experiences more visible, shareable, and socially reinforced, while also facilitating access to professional advice and exercise-related information. For example, sharing exercise records, participating in online exercise check-ins, receiving likes and comments, and exchanging sport-related information may provide encouragement, recognition, practical advice, and a sense of companionship. At the same time, online social support should be interpreted in a balanced manner. Although it can serve as a valuable extension of real-life support, the quality of online content and peer advice may vary. Adolescents may also be exposed to misinformation, commercialized recommendations, or uncritical reliance on peer suggestions [53]. These forms of support may be associated with adolescents’ stronger sense of engagement, belonging, and sustained motivation for physical activity when the digital environment is appropriately guided. This interpretation is consistent with research showing that social support is associated with health behaviors, psychological adaptation, exercise adherence, and overall quality of life [54] and that peer support is particularly important for adolescents’ persistence in physical activity, potentially through self-efficacy and self-regulation [55].

Taken together, online social support may help connect the digital context of social media content production with adolescents’ physical health promotion. From the perspective of social cognitive theory, individual behavior is associated with the interaction among environmental, personal, and behavioral factors; in mass communication contexts, symbolic transmission and social reinforcement may be related to cognition, emotion, and behavior [56]. In physical activity research, social support, self-efficacy, and outcome expectations are widely recognized as important psychosocial factors associated with exercise behavior [57]. These findings suggest that school-based and digital health promotion programs should not only provide sport and health information but also create structured online support environments, such as supervised exercise check-ins, peer feedback, and guided exchange of reliable sport-related knowledge.

Positive psychological capital as a mediator in the association between social media content production and adolescent physical health promotion

The results suggest that positive psychological capital is an important psychological resource involved in the association between social media content production and adolescent physical health promotion. Unlike studies that mainly focus on the direct relationship between social media use and adolescent health outcomes, the present study highlights a possible psychological process by which participatory engagement with sport-related social media content is associated with adolescents’ health-promoting tendencies. Positive psychological capital generally consists of self-efficacy, hope, optimism, and resilience, which may help adolescents maintain constructive cognitive appraisals and adaptive action tendencies when facing stress, setbacks, or difficulties in sustaining physical activity.

Social media content production may be associated with positive psychological capital through self-presentation and social feedback related to user-generated content, as well as observational learning and professional guidance derived from engagement with professionally generated content. When adolescents share exercise experiences, athletic achievements, or healthy lifestyles online and receive likes, comments, or encouragement, these experiences may be linked to more positive exercise-related self-perceptions, perceived competence, and optimistic expectations. This interpretation is consistent with social cognitive theory, which suggests that individuals can develop positive self-evaluations and behavioral beliefs through observational learning, social persuasion, and mastery experiences [58]. At the same time, the dual nature of social media-related associations should be recognized. Active participation and proactive content production may be associated with self-affirmation, social connectedness, and positive emotions, whereas passive browsing, excessive social comparison, or negative interactions may increase anxiety, envy, or the risk of depression [59]. Therefore, the association between social media use and adolescent positive psychological capital may depend on patterns of use, the quality of interactions, and individual psychological characteristics.

Positive psychological capital may also be closely related to adolescent physical health promotion because it reflects developable psychological resources that support goal persistence, stress coping, and problem-solving. Previous studies have shown that positive psychological capital is associated with adolescent subjective well-being and mental health and may help translate positive cognitions into sustained actions in learning, daily life, and health-related contexts [60]. Within physical health promotion, self-efficacy may support confidence in exercise participation, hope may help adolescents maintain goal-oriented expectations, optimism may encourage positive appraisals of exercise experiences, and resilience may support persistence when adolescents encounter barriers to physical activity.

From the perspective of social cognitive theory, individual behavior is associated with the interaction among environmental, cognitive, and behavioral factors. As a form of symbolic communication and interactive practice, social media content production may be related to adolescents’ cognition, affect, and behavior through observational learning, outcome expectations, social reinforcement, and self-evaluation [56]. In physical activity research, social cognitive theory has been widely used to explain exercise-related behavior through constructs such as self-efficacy, outcome expectations, social support, and goal setting [57]. Accordingly, positive psychological capital may serve as an internal psychological resource through which sports-related social media engagement is associated with more stable health-promoting tendencies.

Chain mediation roles of online social support and positive psychological capital in adolescents

The results suggest that online social support was positively associated with positive psychological capital, indicating that adolescents who perceive more comprehensive online support may also report higher levels of internal psychological resources. This finding is consistent with the Youth Physical Activity Promotion Model, which classifies determinants of adolescent physical activity into predisposing, reinforcing, and enabling factors [61]. In the present study, social media content production can be understood as an enabling factor, online social support as a reinforcing factor, and positive psychological capital as a predisposing factor. This provides a useful framework for understanding the sequential association from digital engagement to social support, psychological resources, and physical health promotion.

Social cognitive theory provides a complementary explanation for this sequential association. According to this theory, individual behavior is associated with the dynamic interaction among environmental, cognitive, and behavioral factors. In mass communication contexts, symbolic transmission may be related to adolescents’ cognition, affect, and behavior through observational learning, outcome expectations, self-efficacy, and social reinforcement [56]. From this perspective, online social support represents an external social resource generated through digital interaction, whereas positive psychological capital represents an internal psychological resource that may support goal pursuit, optimism, resilience, and perceived competence in physical activity contexts. Therefore, the two factors may be associated with adolescent physical health promotion in complementary ways rather than operating separately.

Further analysis based on the hypothesized model showed that online social support and positive psychological capital were associated with the relationship between social media content production and adolescent physical health promotion through both independent and sequential forms. The total indirect association accounted for 63.778% of the total association, whereas the sequential association involving online social support and positive psychological capital accounted for 14.347%. This distinction suggests that the overall indirect association was not limited to the sequential form but reflected the combined contributions of multiple forms within the model. Previous research has shown that social support is not only positively associated with physical activity behaviors but may also be indirectly associated with exercise participation and adherence through positive psychological resources such as psychological capital [62]. In particular, peer support and emotional support are strongly associated with adolescent physical activity through greater participation willingness, persistence, and behavioral maintenance [63]. Taken together, online social support and positive psychological capital may work together in the context of sport-related social media engagement and may be jointly associated with higher levels of adolescent physical health promotion.

These findings may offer several practical implications for school-based and digital health promotion, although they should be interpreted as implications derived from observed associations rather than evidence from an intervention. In school settings, physical education teachers and health educators could guide adolescents to create positive sport-related content, share exercise experiences, participate in supervised exercise check-ins, provide constructive peer feedback, and critically engage with reliable professionally generated sport and health information. Such activities may help connect daily physical activity with supportive social interaction and positive psychological experiences. At the platform level, social media services could be encouraged to promote reliable sport and health information, support constructive feedback, and reduce harmful comparison-based interactions. These strategies may help create a more supportive digital environment for adolescent physical health promotion.

Study limitations

This study has several limitations. First, this study employed a cross-sectional survey design, which limits the ability to establish causal relationships and temporal ordering. Although the proposed chain mediation model was theory-driven, its directional sequence cannot be empirically confirmed. Alternative models are also plausible, including reversed or parallel mediation structures. Future research could use longitudinal tracking, experimental designs, multi-group analyses, and cross-cultural comparative studies to further investigate the dynamic processes, group differences, cultural applicability, and possible causal directions through which social media content production influences adolescent physical health promotion. Second, although the main findings were based on a theory-driven structural model, residual confounding from demographic, family, school, and behavioral factors cannot be fully ruled out. Future studies should collect more comprehensive background information and further examine these associations using covariate-adjusted, stratified, or multilevel models. Third, the revised model was not cross-validated in an independent sample, and a fully integrated CFA including all latent variables in the structural model was not further reported. Although the use of dimension-level averaged indicators was theoretically justified and helped reduce model complexity, it may still simplify the item-level measurement structure. Future studies should further test the robustness of the measurement and structural models using integrated CFA, alternative model comparison, HTMT analysis, split-sample validation, independent samples, or preregistered model specifications. Fourth, although the study improved geographic coverage by including adolescents from seven major geographic regions in China, the use of convenience and cluster sampling components and the reliance on population proportion-based sample size estimation may limit generalizability and methodological rigor. Future studies should adopt more rigorous probability sampling strategies and sample size estimation approaches tailored to structural equation modeling and mediation analysis. Fifth, all variables were measured using self-report questionnaires from the same respondents at a single time point, which may introduce common method bias. Although Harman’s single-factor test was conducted, this approach cannot fully rule out its presence. Future studies are encouraged to adopt more rigorous designs, such as collecting data from multiple sources or at multiple time points, to further address this issue. Additionally, the operationalization of social media content production includes both active production and passive engagement components, which may introduce conceptual overlap. Future studies are encouraged to distinguish more clearly between different types of social media use behaviors.

Conclusion

This study examined the associations among social media content production, online social support, positive psychological capital, and adolescent physical health promotion using a cross-sectional chain mediation model. The main conclusions are as follows:

1. Social media content production was significantly and positively associated with adolescent online social support, positive psychological capital, and physical health promotion. These findings indicate that adolescents who report greater participatory engagement with sport-related social media content, including active creation, interaction, and engagement with professionally generated content, also tend to report higher levels of perceived online social support, positive psychological capital, and physical health promotion.

2. Online social support and positive psychological capital jointly played significant sequential mediating roles in the association between social media content production and adolescent physical health promotion, with the indirect association reaching statistical significance. This suggests that social media content production may be linked to adolescent physical health promotion through higher levels of online social support and positive psychological capital.

3. The study identifies a possible explanatory pattern of “social media content production → online social support → positive psychological capital → adolescent physical health promotion.” This pattern suggests that participatory engagement with sport-related social media content may be associated with stronger social connections and the accumulation and transformation of positive psychological resources, which may support the formation and maintenance of self-reported health-promoting behavioral tendencies.

However, because this study was based on cross-sectional survey data, the findings cannot establish policy effectiveness, intervention effects, or long-term improvements in adolescent physical health. Therefore, the practical implications should be interpreted as preliminary and association-based rather than as evidence of sustained health outcomes.

Authors' contributions

Minmin Du: writing-original draft, data curation, software, methodology. Xinchun Wang: writing-reviewing and editing, conceptualisation, methodology, supervision. Jintian Zhang: writing-reviewing and editing, formal analysis. Minmin Du and Xinchun Wang contribute equally to the work.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of Qufu Normal University (Approval No.: 2021098) and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from parents or legal guardians for participants under 18 years of age, along with assent from the minors. Participants aged 18 years or older provided their own informed consent. Participation was voluntary and anonymous, and all data were collected without personal identifiers. During data collection, teachers were present only to maintain classroom order and did not participate in the administration or supervision of the survey.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.