Family Sport Involvement and Adolescent Prosocial Behavior: A Latent Profile Analysis

Abstract

Background

Family sport involvement is critical for adolescent prosocial development, yet how distinct patterns of parental engagement relate to prosocial behavior and through what psychological processes remains unclear. Integrating Self-Determination Theory with Bourdieu’s sociological concepts of habitus and capital, this study aimed to examine the distinct profiles of family sport involvement and their mechanisms in predicting adolescents’ prosocial behavior.

Methods

Data were drawn from a large-scale national survey of Chinese youth (N= 14,271). Latent profile analysis (LPA) was utilized to identify familial sport involvement profiles. Subsequently, multiple regression and Karlson-Holm-Breen (KHB) mediation analyses were conducted to evaluate the direct and indirect associations, as well as the moderating role of family socioeconomic status (SES).

Results

LPA identified four distinct profiles of family sport involvement: disengaged (24.80%), moderate involvement (31.36%), modeling-only (19.61%), highly supportive (24.22%). Compared to the disengaged profile, the moderate involvement, modeling-only, and highly supportive profiles were significantly associated with adolescents’ prosocial behavior. Mechanistically, these effects were mediated by autonomous motivation, self-efficacy, and extracurricular physical exercise. Self-efficacy emerged as the primary mediator, accounting for 75–83% of the indirect effects. Crucially, family SES moderated these pathways: higher SES amplified the developmental benefits, though this moderation was uniquely significant only within the highly supportive profile.

Conclusion

This study theoretically bridges the gap between psychological agency and structural realities in youth development. Practically, while objective capital (SES) maximizes developmental benefits, cultivating an autonomy-supportive sport habitus is the core psychological driver of positive youth development. Promoting highly supportive family sport involvement offers a vital pathway to mitigate socio-structural disparities through targeted parenting interventions.

Introduction

Prosocial behavior significantly facilitates adolescents’ psychosocial adjustment,¹ and psychological well-being.² Furthermore, empirical evidence highlights that prosocial tendencies exert a strong positive impact on subsequent academic achievement and social preferences.^3,4 However, during early adolescence, sport participation,⁵ predominantly unfolds within family and school contexts. Because the interactive nature of sports provides an ideal environment for adolescents to learn cooperation, empathy, and rule adherence,⁶ family sport involvement encompassing parental support, co-participation, and role modeling may serve as a foundational catalyst for their positive socialization.

The salience of family sport involvement has been profoundly amplified by China’s recent macro-level policy shifts, most notably the 2021 “Double Reduction” policy. By restricting off-campus academic tutoring and reducing school assigned homework, this policy has prompted families to redirect significant time and resources toward non-academic pursuits, particularly youth sports.¹ However, a critical practical problem persists: despite this rapid expansion in adolescent sport participation, there is a notable lack of empirical studies utilizing large-scale survey data to systematically classify the distinct patterns of family sport involvement in China. Furthermore, the specific psychological and behavioral mechanisms through which family sport involvement fosters prosocial behavior remain largely underexplored.

Moreover, a significant theoretical gap exists in current literature. while Self-Determination Theory (SDT) elucidates the micro-psychological mechanisms driving adolescents’ prosocial behavior, it often overlooks the macro-structural inequalities that dictate how families redirect their resources in this new policy era. To address this gap, the present study integrates Bourdieu’s concepts of capital,⁷ habitus⁸ with SDT within the Chinese cultural context. Specifically, we conceptualize varying patterns of family sport involvement as deeply ingrained sport habitus, bounded by the objective capital of family socioeconomic status (SES).

Guided by this dual-lens framework and utilizing a large-scale national dataset, the direction of this research is to identify distinct family sport involvement profiles within the Chinese context and uncover these complex underlying mechanisms. Specifically, we pursue three objectives. First, via Latent Profile Analysis (LPA), we delineate empirical typologies of family sport involvement to quantitatively capture familial sport habitus. Second, drawing upon SDT, we elucidate the mechanistic model linking these habitus profiles to adolescents’ prosocial behavior. Finally, we investigate the boundary conditions of family SES within these pathways. Ultimately, by identifying specific subgroups of family sport involvement, this study aims to inform tailored, socioeconomically-sensitive interventions for adolescents’ psychosocial well-being.

Theoretical Analysis and Research Hypothesis

Prosocial behavior, encompassing voluntary interpersonal actions such as sharing, helping, and cooperating, is a critical component of healthy youth development.⁹ While families and schools serve as primary socialization contexts for internalizing these values,¹⁰ the sporting context is increasingly recognized as an optimal arena for cultivating prosociality. Specifically, sports provide a structured and interactive environment, thereby directly fostering observable prosocial behaviors.⁶ Importantly, the intersection of family and sport plays a pivotal role, adolescents raised in sports-oriented family environments are significantly more likely to engage in organized athletic activities,¹¹ thereby enhancing these developmental benefits.

From Bourdieu’s perspective, sport serves as a critical social practice that mirrors and perpetuates existing societal structures. Moving beyond a economic definition, Bourdieu (1986) expanded the concept of capital to encompass cultural, social, and symbolic forms,⁷ all of which play a significant role in structuring social behaviours, including sports participation.⁶ The concept of habitus refers to the internalized dispositions that shape individuals’ perceptions, thoughts, and actions, including their preferences and behaviours in sports.⁸ By integrating Bourdieu’s sociological concepts into the context of youth development emphasizes the family as a primary socialization agent. Family capital within the sports field refers to the tangible resources parents invest, which we operationalized through measured indicators of structural support (eg, providing financial resources and assistance).¹² Family sport habitus encompasses the internalized frameworks, perspectives, values, and practices shared within a household,¹³ that foster ideas about sport as a natural part of life. Because habitus is embodied and materially expressed through everyday practices, we operationalized the transmission of this sport-oriented habitus through behavioral indicators, specifically parental co-participation, parents’ own training habits (role modeling), and active encouragement. By integrating the two concepts, we establish a clear correspondence between sociological theory and our psychological assessment. Crucially, this active family environment serves as more than just a site for athletic skill acquisition or class reproduction;¹² rather, it constitutes a vital developmental context for cultivating adolescents’ prosocial behavior.

However, the psychological impacts of parental sports involvement are complex and non-linear. While active parental support generally fosters positive youth development, the qualitative nature of this involvement dictates the developmental outcomes. For instance, Stein and Raedeke (1999) demonstrated that youth psychological experiences fluctuate depending on whether parental involvement is perceived as insufficient, adequate, or excessive.¹⁴ Furthermore, specific parental socialization values can yield unintended consequences; notably, a maternal overemphasis on performance and status within sports has been linked to an increased frequency of adolescent antisocial behaviors.¹⁵ These findings underscore that parental involvement is not a construct, but rather a complex interplay of distinct parenting dimensions.

To adequately capture this psychological and behavioral complexity, analytical approaches must evolve. Historically, research on family sport socialization has relied on variable-centered methods, operationalizing parental involvement along isolated, continuous dimensions. While conceptually informative, this approach risks obscuring the unobserved heterogeneity and naturally occurring combinations of parenting practices within individual families. Consequently, adopting a person-centered approach is increasingly advocated in developmental sport psychology. By identifying distinct, multidimensional profiles of family sport involvement, LPA offers a robust methodological alignment with the concept of family habitus, enabling a more precise examination of how specific configurations of parental practices differentially shape youth prosocial behavior.

Psychological and Behavioral Mechanisms

The developmental impacts of family sport involvement on youth prosocial behavior are rarely direct. Rather, they operate through distinct psychological and behavioral pathways. This study posits a moderated mediation model, we first elucidate these indirect effects through three primary mechanisms.

First, at the psychological level, both self-efficacy and autonomous motivation function as vital internal mechanisms. Drawing upon self-determination theory, a supportive family sport habitus—one that emphasizes choice and minimizes coercion—effectively satisfies youths’ basic psychological needs for autonomy, competence, and relatedness.¹⁶ This need satisfaction promotes the internalization of sport values, fostering robust autonomous motivation and social self-efficacy. Such psychological assets are fundamental to prosocial development; adolescents who feel competent and autonomously driven are significantly better positioned to engage in cooperative and empathetic peer interactions. Conversely, when youth are deprived of this autonomy-supportive context,¹⁷ the frustration of psychological needs may impede the cultivation of prosocial dispositions.

Second, at the behavioral level, actual extracurricular physical exercise acts as the essential behavioral translation of family sport involvement. Although parental involvement provides the psychological foundation, yielding prosocial benefits requires the conversion of this disposition into sustained physical activity. However, the efficacy of this behavioral pathway is contingent upon macro-institutional frameworks.¹⁸ In contrast to egalitarian Nordic welfare regimes,¹⁹ or the neoliberal system in the United States,²⁰ China presents a unique structural reality. The historically entrenched exam-oriented culture structurally constrains this conversion process, frequently demanding that family capital be channeled into academic shadow education rather than physical investments.²¹ This institutional crowding-out of physical activity not only poses widespread health risks but also significantly diminishes opportunities for peer socialization. Therefore, in the Chinese context, actual extracurricular exercise functions as a culturally bound behavioral mediator. It represents a critical breakthrough against institutional constraints, facilitating the peer-based social interactions vital for adolescents’ prosocial development.

Despite their theoretical importance, existing studies have predominantly examined self-efficacy, autonomous motivation, and exercise behavior in isolation. However, in genuine psychosocial processes, cognitive appraisals, motivational drives, and behavioral habits function concurrently. To address this gap, the current study integrates these three mechanisms into a unified multiple mediation framework. This concurrent modeling allows us to compare the relative magnitude of each pathway, revealing how they synergistically translate family sport habitus into youth prosociality.

The Moderating Role of Family Socioeconomic Status

While psychological needs and behavioral habits act as essential mediators, these developmental processes do not occur in a socioeconomic vacuum.²² Scholars have long debated how family socioeconomic status structurally constrains or facilitates prosocial development, a debate primarily divided between the resource perspective and the risk management perspective.

From a resource and family investment standpoint, financial affluence dictates parents’ capacity to intervene in their children’s socialization pathways.²³ Aligned with Lareau’s (2003) concept of concerted cultivation, high-SES parents tend to deliberately structure their children’s leisure time—often through organized sports—to foster interpersonal rules and teamwork.²⁴ In this context, youth from privileged backgrounds perceive robust social support, and this sense of abundance facilitates altruistic behaviors. Conversely, the economic pressure prevalent in low-SES households can severely hinder the cultivation of adolescents’ prosocial behaviors.^25,26

Paradoxically, the risk management perspective offers a contrasting prediction. Proponents argue that individuals from lower social classes, facing highly uncertain environments, exhibit greater generosity, compassion, and trust.²⁷ Relying on tightly knit social bonds and mutual aid for survival, lower-SES individuals often demonstrate a “having less, giving more” paradigm driven by heightened empathy.¹⁷ However, a recent comprehensive meta-analysis by Wu et al (2025) synthesized these conflicting views, concluding that the relatively high psychological and material costs of prosociality, combined with the experience of scarcity, ultimately lead individuals from lower-class backgrounds to exhibit lower overall levels of prosocial behavior.²⁸

Integrating these sociological and psychological insights, we posit that family SES acts as a crucial moderator. Specifically, family SES is likely to moderate the strength of the indirect pathways (ie, motivation and exercise behavior) linking family sport habitus to prosocial behavior. For high-SES families, a supportive sport habitus can be seamlessly translated into structured extracurricular exercise, thereby amplifying the positive effects on prosociality. In contrast, for low-SES youth, the profound experience of material scarcity and the structural lack of access to organized sports may act as a developmental bottleneck. Consequently, even if low-SES parents successfully cultivate their children’s autonomous motivation, the translation of this psychological asset into actual cooperative behaviors may be truncated by resource limitations.

Theoretical Framework: Integrating SDT and Bourdieu’s Sociological Concept

Previous research has primarily explored the psychological mechanisms of adolescent prosocial behavior in isolation, or focused independently on broad parenting styles, neglecting their potential integration through the critical lens of family sport habitus. Furthermore, it remains largely underexplored how specific profiles of family sport involvement operate through interrelated psychological and behavioral mechanisms, and whether family SES acts as a boundary condition for these effects.

To address these gaps, the present study integrates the macro-structural lens of Bourdieu’s concepts of habitus and capital with the micro-psychological perspective of SDT to systematically model the pathways from family sport involvement to prosociality. By doing so, we propose that family sport involvement operates as a manifestation of family habitus, which fosters youth prosocial behavior through specific psychological and behavioral mechanisms derived from SDT. Furthermore, family SES, representing structural capital, is posited as a boundary condition that moderates these pathways, illustrating how objective resources constrain or facilitate the effects of family sport habitus. Specifically, we aim to test the following hypotheses (Figure 1):

Hypothesis 1: Family sports involvement positively impacts prosocial behavior in adolescents.

Hypothesis 2: Self-efficacy, autonomous motivation, and extracurricular physical exercise mediate the impact of family sports involvement on prosocial behavior in adolescents.

Hypothesis 3: Family SES moderates the positive direct effect of family sports involvement on prosocial behavior in adolescents.

Figure 1.

Theoretical Model and Hypotheses.

Study Design

Data

Data for this study were drawn from the China Adolescent Health Theme Database (DYH). This longitudinal project monitors multidimensional adolescent well-being, including physical activity and psychological health. The sampling framework encompassed 186 secondary schools across 17 municipalities in Shandong Province, China. Accounting for approximately 7% of China’s population and economic output, Shandong presents a highly diverse socio-economic landscape, spanning developed coastal cities to developing inland areas, coupled with a strong cultural emphasis on youth sports participation.²⁹ These specific demographic and cultural characteristics make it an ideal context for examining the intersection of family socioeconomic status, adolescent sports involvement, and youth development. However, because the data are regionally specific, caution is warranted before generalizing the findings to a strictly national level.¹²

The study focused on a cohort of junior high school students. From an initial pool of 14,647 participants, 376 responses were excluded due to missing data or statistical outliers. The final analytic sample comprised 14,271 respondents (7598 boys [53.2%] and 6673 girls [46.8%]; mean age=13.42 years, SD=21.6).

Measures

Adolescents’ Prosocial Behavior

Adolescents’prosocial behavior was measured using a 13-item scale originally developed by Gan and Zhang.³⁰ The assessment focused on social involvement, specifically assessing their collaborative efforts, proactive assistance to peers, empathy toward others’ challenges, and the breadth of their social circles. An example item is “able to actively help classmates”. Respondents rated items on a 5-point Likert scale (ranging from 1=never to 5=frequently), where higher scores represented a higher incidence of such behaviors. The internal consistency of this scale was excellent (Cronbach’s α=0.96). A final composite score was derived by averaging all item responses.

Family Sport Involvement

Family sport involvement profiles were identified using five indicators capturing three dimensions. Parental modeling and family co-participation were assessed by adolescents’ reports of their father’s, mother’s, and joint family’s weekly exercise frequencies. An example question is “How many times did your father exercise in the past week?” All rated on 5-point scales ranging from 1 (0 times) to 5 (4 or more times). Resource facilitation was captured by two items evaluating extracurricular sports involvement: the quantity of sports programs (1=0 programs to 4=3 or more programs) and the duration of sustained participation (1=<3 months to 5≤25 months). To facilitate an intuitive interpretation of the results, the raw scores of these five indicators were retained for conducting the LPA. Subsequently, all variables were standardized to ensure equal weighting in profile estimation. Furthermore, because this measure is a composite index of different item formats, traditional Cronbach’s α is not applicable.

Family Socioeconomic Status

Drawing upon the widely recognized tripartite model of socioeconomic status,³¹ the SES index comprised five items covering three distinct dimensions: education, occupation, and family wealth. First, paternal and maternal educational levels were each measured on a 9-point ordinal scale (ranging from 1=no formal education to 9=postgraduate degree). Second, paternal and maternal occupational status were evaluated using a 13-point categorization system, covering various professions. For statistical analysis, these 13 categories were rank-ordered (from 1 to 13) based on occupational prestige and socio-economic standing, where higher values indicate higher occupational status. Third, given that adolescents rarely know their exact household income, this study drew upon the classic construction method of the index of Economic, Social and Cultural Status (ESCS) originally developed for the Programme for International Student Assessment (PISA),³² family wealth was assessed via self-perceived economic status (“How would you rate your family’s economic condition?”), rated on a 5-point Likert scale (from 1=poorest to 5=wealthiest). Because these five items utilize different measurement scales, a principal component analysis (PCA) was conducted to generate a standardized composite SES score, a standard methodological approach for constructing wealth indices.³³ The Kaiser-Meyer-Olkin (KMO) measure was 0.793, indicating the data were highly suitable for PCA. The first principal component accounted for 55.92% of the total variance. Following the PCA, factor scores were extracted to form a continuous SES index, where higher scores reflect a higher level of family SES.

Self-Efficacy

It was assessed using a 6-item measure adapted from an established instrument designed to evaluate the cognitive, emotional, and behavioral resilience of adolescents aged 10–18 years in adverse situations, as well as their capacity to withstand adversity and achieve positive adaptation.³⁴ This measure consists of six items, with example items including “I have many things to be proud of” and “I am able to do things as well as most other people”. Responses were rated on a 5-point Likert scale (ranging from 1=never to 5=usually). The self-efficacy score was computed by averaging the responses across the six items, with higher scores denoting greater self-efficacy (Cronbach’s α=0.85).

Autonomous Motivation

To assess autonomous exercise motivation, a 9-item instrument was developed by Gan and Zhang,³⁰ including items assessing students’ desire to engage in athletic activities and their perceived ease of participating in various sports. An example item is “Can you easily participate in track and field and ball games”. Each response was recorded on a 4-point Likert scale, with values spanning from 1=never to 4=very easily. The autonomous motivation score was derived by averaging the nine items, where a higher score reflects greater autonomous motivation (Cronbach’s α=0.89).

Extracurricular Physical Exercise

Extracurricular physical exercise was assessed using a single-item behavioral measure. This item served as participants’ self-assessment of extracurricular physical exercise during the past week. The specific question was: “Which of the following best describes your physical activity in the past 7 days?” Responses were rated on a 5-point Likert scale, ranging from 1 (all or most of my free time was spent doing things that involved little physical effort) to 5 (very often [7 or more times] did physical activities in my free time). Higher scores indicated a higher frequency of extracurricular physical exercise participation.

Covariates

Based on previous research^12,35 and data availability, we included gender, residential area, only-child status, and parental educational expectations (parents’ expectations for both the adolescent’s academic performance and educational attainment) as covariates. The descriptive statistics are shown in Table 1.

Table 1.

Descriptive Statistics

Variables	Measurement	Mean/%	SD
Gender	Boy	53.24
	Girl	46.76
Residence area	Urban Center	37.05
	Suburb	6.97
	Township	25.59
	Other County	2.75
	Rural	27.64
Only child	Yes	27.43
	No	72.57
Parental academic expectations		2.10	0.93
Parental educational expectations		4.44	1.35
Family socioeconomic status		−0.01	1.01
Autonomous motivation		2.68	0.58
Self-efficacy		3.58	0.67
Extracurricular physical exercise		2.37	1.05
Prosocial behavior		3.69	0.88

Statistical Analysis

All data preparation and statistical analyses were conducted using Stata version 17.0. The analytical strategy was executed systematically in four primary phases. First, LPA was performed to derive distinct profiles of family sport involvement based on the observed indicators. Second, ordinary least squares regression models were estimated to examine the main associations between the identified profiles and prosocial behavior. Third, to formally evaluate the hypothesized mediation mechanisms, the KHB method was applied. This specific approach was utilized to decompose the total effects into direct and indirect components, enabling robust estimation of the mediation pathways while adjusting for relevant covariates. Finally, moderation analyses were conducted by introducing interaction terms (family sport involvement profiles×Family SES) into OLS models to test whether Family SES buffered or amplified the main effects. Upon detecting significant interactions, post-estimation marginal analyses were conducted using Stata’s margins command to calculate predictive margins, elucidating profile effects across varying levels of Family SES.

Latent Profile Analysis (LPA) was selected as the primary analytical method due to several distinct methodological advantages. First, traditional variable-centered analytical paradigms are fundamentally limited in their capacity to classify individuals based on complex behavioral combinations, making it particularly challenging to identify high-risk subgroups.³⁶ Specifically in this study, family sport involvement is highly multidimensional, encompassing diverse combinations of parental modeling, family co-participation, and resource facilitation. To address this complexity, LPA serves as a robust person-centered mixture modeling technique. It effectively captures the underlying distributional characteristics of these observed variables, revealing how they co-occur to form distinct family profiles. Furthermore, compared to traditional distance-based clustering (eg, K-means), LPA offers rigorous, model-based statistical criteria (eg, AIC, BIC, and Entropy) to objectively determine the optimal number of latent classes. Ultimately, by identifying these specific typologies of family sport involvement and evaluating their corresponding health effects, this approach provides a strong empirical foundation for developing precisely targeted intervention strategies.^37,38

Mediation analysis examined temporal relationships using Karlson-Holm-Breen (KHB) method.^39,40 Originally developed to ensure valid comparisons of coefficients across nested models, the KHB method provides a highly robust statistical framework for effect decomposition. Specifically, this approach was utilized to decompose the total effects of family sport involvement profiles on the core outcomes into direct and indirect components. Compared to traditional mediation procedures, the KHB method is particularly advantageous as it accurately calculates the standard errors for the indirect pathways and provides unbiased estimates of mediation effects, even when adjusting for multiple confounding covariates.

Results

Latent Profile Analysis of Family Sport Involvement

Table 2 presents the model fit indices and model-estimated class proportions for a series of latent profile models, estimating one to four profiles, conducted to identify distinct unobserved subgroups of family sport involvement.

Table 2.

Fit Statistics for Latent Profile Models

Latent Profiles	N	AIC	BIC	Entropy	Estimated Class Proportions
Profile 1	14,271	203,014	203,042	—	1.00
Profile 2	14,271	189,314	189,435	0.7857	0.391/0.609
Profile 3	14,271	183,047	183,213	0.7714	0.271/0.484/0.245
Profile 4	14,271	180,747	180,959	0.8497	0.254/0.330/0.176/0.240

Notes: Higher entropy values indicate clearer class separation, entropy is not available for the one-profile model. Estimated class proportions represent model-based latent profile prevalence parameters estimated directly by the mixture model, as opposed to proportions based on modal class assignment (see Table 3).

Abbreviations: AIC, Akaike information criterion; BIC, Bayesian information criterion.

The optimal number of latent profiles was determined based on a combination of statistical fit indices and criteria for practical interpretability. Following the standardization of all indicator variables, a series of models ranging from two to four classes were estimated using Generalized Structural Equation Modeling (GSEM). Specifically, the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were utilized, with lower values indicating superior model fit and parsimony.⁴¹ Furthermore, entropy was used to evaluate classification quality and class separation. Entropy values range from 0 to 1, with a recommended threshold of ≥.80 indicating high classification accuracy.⁴² Finally, to ensure theoretical meaningfulness and prevent the over-extraction of spurious groups, a minimum class size criterion was applied, requiring that no identified latent class should comprise less than 5% of the total sample.⁴³

As shown in Table 2 and visually depicted in the newly added Elbow Plot (Figure S1 in the Supplementary Material), while the AIC and BIC values continued to decrease with the addition of more classes, the four-profile solution demonstrated the highest entropy (0.85) and retained adequate sample sizes across all subgroups (ranging from 17.6% to 33.0%). Most importantly, beyond these statistical metrics, the four-profile solution offered the clearest empirical and theoretical explanation of the data. It successfully captured distinct, non-overlapping typologies of family sport involvement), which perfectly align with theoretical perspectives on family socialization. Therefore, the four-profile model was selected as the most appropriate representation of family sport involvement.

LPA based on the five indicators of family sport involvement identified a four-profile model (Table 3). Unstandardized scores are displayed for intuitive interpretation. These variables were subsequently standardized for regression analyses.⁴⁴

Table 3.

Conditional Means and Sample Distribution of Family Sport Involvement Profiles

Indicator	Profile 1	Profile 2	Profile 3	Profile 4
	Disengaged	Moderate Involvement	Modeling-Only	Highly Supportive
Paternal exercise frequency	1.655	3.142	3.822	4.506
Maternal exercise frequency	1.243	2.801	4.579	4.683
Family co-participation frequency	1.387	2.661	2.021	4.300
Extracurricular sports (quantity)	1.356	1.716	1.461	2.020
Extracurricular sports (duration)	1.528	1.828	1.558	2.109
Sample Proportion	24.80%	31.36%	19.61%	24.22%
Sample Size (n)	3539	4476	2799	3457

Notes: Values represent the conditional means of each indicator given profile membership. Sample proportions and sample sizes (n) are based on participants’ most likely profile assignment (modal assignment) derived from posterior probabilities.

As shown in Table 3, Profile1 labeled as disengaged (24.80%), was characterized by the lowest mean scores across all behavioral indicators. Profile 2 designated as moderate-involvement, represented the largest subgroup (31.36%) and exhibited intermediate levels of physical activity across all assessed dimensions. Profile 3 termed modeling-only (19.61%), revealed a distinct in pattern: it demonstrated a high frequency of maternal weekly workouts but relatively low engagement in joint family workouts and organized sports classes. Profile 4 identified as highly supportive (24.22%), reported the highest levels of participation across all metrics, particularly regarding both parents’ individual exercise routines and family co-participation.

The latent profile configurations for family sport involvement are visualized in Figure 2, illustrating the conditional mean scores across the four profiles. Specifically, the disengaged was characterized by a consistently suppressed profile, with all indicators remaining substantially below the scale’s theoretical midpoint. In contrast, moderate involvement demonstrated a parallel trajectory to disengaged profile but at an intermediate elevation, with paternal workout frequency emerging as the primary behavioral driver.

Figure 2.

Characteristics of the four latent profiles based on family sport involvement.

A notably distinct pattern was observed in modeling-only profile, which exhibited significant behavioral asymmetry. This profile was defined by a prominent peak in maternal physical activity alongside disproportionately low levels of family workouts and interest sports class involvement. Viewed through Bourdieu’s theory, this profile captures a highly individualized, rather than collective, sporting habitus. Despite parents possessing a strong internal disposition for exercise, there is a distinct failure to translate this into active intergenerational transmission. Finally, the highly supportive manifested a high-plateau configuration for family and interest exercise. This group represented the highest level of behavioral synergy, with robust participation across paternal, maternal, and joint family domains. Theoretically, this synergy exemplifies a robust, collective family habitus, where parental dispositions are successfully converted into shared practices and investments in children’s physical capital. The comparative analysis suggests that the most critical source of variance between the high-engagement group and other profiles lies in the frequency of collective family sport and structured interest training—essentially, the mechanisms of habitus transmission. It is precisely the identification of this nuanced modeling-only typology, which would be obscured in a simpler three-class model, that makes the four-profile solution the most empirically and theoretically sound representation of the data.

Table 4 presents the descriptive statistics comparing demographic and socioeconomic characteristics across the distinct profiles of Chinese family sport involvement. Results revealed that the four patterns differed significantly on all examined characteristics, with the sole exception of single-child status.

Table 4.

Sample Distribution and Predicted Probabilities of Family Sport Involvement Profiles

Variable	Mean/% (SD)	Profile 1: Disengaged (n=3539)	Profile 2: Moderate Involvement (n=4476)	Profile 3: Modeling-Only (n=2799)	Profile 4: Highly Supportive (n=3457)	P
Gender	Boys (%)	56.09	53.66	49.48	52.82	<0.001
	Girls (%)	43.91	46.34	50.50	47.17
Only child	Yes (%)	28.28	26.72	27.44	27.48	0.487
	No (%)	71.72	73.28	72.56	72.52
Residence area	Urban Center (%)	34.67	35.77	39.63	39.08	<0.001
	Suburban (%)	6.87	7.26	7.36	6.36
	Township (%)	25.74	26.97	23.51	25.34
	Other County (%)	3.08	2.79	2.61	2.46
	Rural (%)	29.64	27.21	26.90	27.76
Parental academic expectations	M (SD)	2.14 (0.97)	2.10 (0.91)	2.07 (0.92)	2.09 (0.91)	<0.001
Parental educational expectations	M (SD)	4.27 (1.45)	4.37 (1.38)	4.54 (1.22)	4.62 (1.25)	<0.001
Family socioeconomic status	M (SD)	−0.002 (1.07)	−0.031 (1.03)	−0.005 (0.97)	0.017 (0.97)	<0.001
Autonomous motivation	M (SD)	2.52 (0.64)	2.67 (0.57)	2.63 (0.49)	2.92 (0.52)	<0.001
Self-efficacy	M (SD)	3.39 (0.67)	3.54 (0.64)	3.62 (0.63)	3.82 (0.66)	0.004
Extracurricular physical exercise	M (SD)	2.07 (1.05)	2.39 (0.98)	2.27 (0.95)	2.75 (1.09)	<0.001
Prosocial behavior	M (SD)	3.40 (0.92)	3.62 (0.84)	3.75 (0.78)	4.03 (0.83)	<0.001

Note: P-values derived from Chi-square tests for categorical variables and ANOVA for continuous variables.

As detailed in Table 4, The results highlighted a distinct polarization between disengaged and highly supportive profiles. Specifically, families in Profile 4 generally resided in urban centers, possessed higher SES, and held higher educational expectations. Correspondingly, adolescents in Profile 4 exhibited the highest levels of autonomous motivation, self-efficacy, physical exercise, and prosocial behavior (p<0.01). In contrast, Profile 1 demonstrated the least favorable outcomes across all parental and adolescents behavioral indicators, while Profile 2 and 3 represented intermediate levels.

Family Sport Involvement Profiles and Prosocial Behavior

Table 5 presents the results of the multiple regression analyses examining the direct effects of family sport involvement profiles on prosocial behavior and the underlying mediating mechanisms.

Table 5.

Regression Analysis

	Model 1	Model 2	Model 3	Model 4	Model 5
Gender (Ref: Girls)	0.085***	−0.064***	0.010	−0.184***	0.099***
	(0.014)	(0.009)	(0.011)	(0.017)	(0.011)
Residence area (Ref: Urban Center)
Suburb	−0.222***	−0.192***	−0.194***	0.024	−0.038
	(0.028)	(0.019)	(0.022)	(0.035)	(0.022)
Township	−0.163***	−0.105***	−0.168***	0.029	−0.019
	(0.018)	(0.012)	(0.014)	(0.023)	(0.014)
Other County	−0.226***	−0.147***	−0.196***	0.079	−0.052
	(0.043)	(0.029)	(0.033)	(0.053)	(0.033)
Rural	−0.255***	−0.198***	−0.224***	0.005	−0.046**
	(0.018)	(0.012)	(0.014)	(0.023)	(0.014)
Only child	0.053**	0.043***	0.031*	−0.087***	0.023
	(0.016)	(0.011)	(0.012)	(0.020)	(0.013)
Parental academic expectations	−0.065***	0.035***	−0.060***	−0.021*	−0.030***
	(0.007)	(0.005)	(0.006)	(0.009)	(0.006)
Parental educational expectations	0.139***	0.069***	0.099***	0.007	0.051***
	(0.005)	(0.003)	(0.004)	(0.006)	(0.004)
Family socioeconomic status	0.064***	0.070***	0.042***	−0.007	0.017**
	(0.007)	(0.005)	(0.006)	(0.009)	(0.006)
Family sport involvement patterns (Ref: Disengaged)
Moderate involvement	0.202***	0.147***	0.131***	0.315***	0.063***
	(0.018)	(0.012)	(0.014)	(0.023)	(0.014)
Modeling-only	0.287***	0.095***	0.187***	0.202***	0.124***
	(0.021)	(0.014)	(0.016)	(0.026)	(0.016)
Highly supportive	0.565***	0.369***	0.377***	0.678***	0.184***
	(0.020)	(0.013)	(0.015)	(0.024)	(0.016)
Autonomous motivation					0.241***
					(0.010)
Self-efficacy					0.718***
					(0.009)
Extracurricular physical exercise					0.031***
					(0.005)
Constant	2.875***	2.271***	3.159***	2.493***	−0.020
	(0.045)	(0.030)	(0.035)	(0.056)	(0.047)
N	14271	14,271	14,271	14,271	14271
Adjusted R2	0.1431	0.1226	0.1269	0.0627	0.4805

Notes: Standard errors in parentheses; *p < 0.05, **p < 0.01, ***p < 0.001.

As shown in Table 5, all models controlled for covariates including gender, residence area, only-child status, parental expectations, and family socioeconomic status. First, to test hypothesis 1 regarding the distinct profiles of involvement, LPA revealed four distinct profiles. Building upon this classification, Model 1 was then estimated to examine the direct relationship between these identified profiles and prosocial behavior. Compared to the disengaged group, adolescents in the moderate involvement (B=0.202, p<0.001), modeling-only (B=0.287, p<0.001), and highly supportive (B=0.565, p<0.001) groups exhibited significantly higher levels of prosocial behavior. The results support hypothesis 1.

Second, Models 2 through 4 evaluated the predictive roles of family sport involvement profiles on the proposed mediators. Results indicated that higher levels of family sport involvement were consistently and positively associated with autonomous motivation, self-efficacy, and extracurricular physical exercise. Notably, the highly supportive profile demonstrated the strongest positive associations across all three mediators.

Finally, Model 5 revealed that autonomous motivation (B=0.241, p<0.001), self-efficacy (B=0.718, p<0.001), and extracurricular physical exercise (B=0.031, p<0.001) all positively and significantly predicted prosocial behavior. Importantly, after introducing these mediators into the model, the predictive coefficients of the family sport involvement profiles on prosocial behavior decreased but remained statistically significant (p<0.001). This indicates that autonomous motivation, self-efficacy, and extracurricular physical exercise partially mediated the relationship between family sport involvement profiles and prosocial behavior.

Mediation Analysis

We employed the KHB method to evaluate the multiple mediating mechanisms.^39,40 Prior to interpreting the independent mediation effects, we assessed potential multicollinearity among the mediators (extracurricular physical exercise, self-efficacy, and autonomous motivation). While the bivariate correlations (see Table S1 in the Supplementary Material) showed significant associations, collinearity diagnostics revealed that the Variance Inflation Factor (VIF) values for all predictors ranged from 1.04 to 1.35, well below the conservative threshold of 5. Furthermore, all tolerance values ranged from 0.74 to 0.96, remaining well above the acceptable minimum.

In support of hypothesis 2, the mediation results presented in Table 6 revealed that, with the disengaged profile as the reference, all three active engagement profiles demonstrated positive and highly significant total and direct effects on prosocial behavior (p<0.001).

Table 6.

Mediation Analysis

	Total Effect	Direct Effect	Total Indirect Effect	Proportion (Total)	Mediator	Specific Indirect Effect	Proportion of Indirect Effect
Family sport involvement profiles (Ref: Disengaged)
Moderate involvement	0.202*** (0.015)	0.063*** (0.015)	0.139*** (0.021)	68.97%	Autonomous motivation	0.024***	17.45%
					Self-efficacy	0.113***	81.01%
					Extracurricular physical exercise	0.002***	1.53%
Modeling-only	0.287*** (0.016)	0.124*** (0.016)	0.163*** (0.021)	56.91%	Autonomous motivation	0.035***	21.73%
					Self-efficacy	0.124***	75.73%
					Extracurricular physical exercise	0.004***	2.55%
Highly supportive	0.565*** (0.016)	0.184*** (0.017)	0.381*** (0.022)	67.49%	Autonomous motivation	0.052***	13.62%
					Self-efficacy	0.319***	83.56%
					Extracurricular physical exercise	0.011***	2.82%

Notes: Robust standard errors are in parentheses. Control variables are included in all models. *** p < 0.001.

Notably, a substantial proportion of total effects was mediated by the included psychosocial and behavioral variables. The total indirect effects accounted for 68.97%, 56.91%, and 67.49% effects for the moderate involvement, modeling-only, and highly supportive profiles, respectively.

Disentangling the indirect effects revealed that self-efficacy served as the predominant mediating pathway. Specifically, self-efficacy explained the vast majority of the indirect effects in the moderate involvement (81.01%), modeling-only (75.73%), and highly supportive (83.56%) profiles. Autonomous motivation also provided a meaningful, mediating contribution, ranging from 13.62% to 21.73%. In contrast, extracurricular physical exercise yielded a statistically significant indirect effect (p<0.001), its relative contribution to the overall mediation was minimal (1.53–2.82%). Collectively, both multiple regression models and KHB decomposition fully support hypothesis 2.

Heterogeneity Analysis

Addressing hypothesis 3 regarding the moderating role of Family SES, Table 7 presents the regression model incorporating interaction terms between the four profiles and family SES.

Table 7.

Moderating Analysis

Variables	Prosocial Behavior
Family sport involvement profiles (Ref: Disengaged)
Moderate involvement	0.052*
	(0.023)
Modeling-only	0.120***
	(0.025)
Highly supportive	0.140***
	(0.024)
Family socioeconomic status	0.017**
	(0.006)
Interaction (Ref: Disengaged × SES)
Moderate involvement × SES	0.017
	(0.029)
Modeling-only× SES	0.006
	(0.033)
Highly supportive × SES	0.071*
	(0.031)
Controlled	Yes
N	14271
Adj-R2	0.4814

Notes: Standard errors in parentheses; *p < 0.05, **p < 0.01, ***p < 0.001.

As shown in Table 7, both family SES and the family sport involvement profiles yielded significant main effects. Relative to the disengaged profile, adolescents in the other three profiles exhibited significantly higher levels of prosocial behavior. Regarding the moderating effects, only the interaction between the highly supportive profile and family SES was significant (B=0.071, p<0.05). This indicates that higher family SES significantly amplifies the positive impact of a highly supportive family environment on adolescent prosocial behavior, whereas the effects of the moderate involvement and modeling-only profiles remain stable regardless of family SES levels. Consequently, hypothesis 3 was fully supported.

To elucidate the nature of this significant interaction, predictive margins were plotted for high and low family SES (split at the mean for visualization purposes, see Figure 3).

Figure 3.

Interaction of Family Sport Involvement Profiles and Family SES.

Notes: Family SES was entered as a continuous variable in the main regression analysis. For visualization purposes only, the predictive margins are plotted for high family SES and low family SES, which correspond to values above and below the mean score, respectively.

Figure 3 presents the differences between high and low family SES were largely negligible across the disengaged, moderate involvement, and modeling-only profiles, with their respective trajectories remaining essentially parallel. However, a distinct divergence emerged at the highly supportive profile. Specifically, transitioning from a modeling-only to a highly supportive family environment yielded a sharp increase in prosocial behavior for adolescents from high family SES backgrounds. In contrast, this beneficial effect was notably attenuated for their low family SES peers. These findings confirm that higher family SES significantly amplifies the positive impact of an optimal (highly supportive) family sport environment on adolescent prosocial development.

Discussion

This study supported the proposed model regarding the development of adolescents’ prosocial behavior. The findings demonstrated that family sports involvement has a positive direct effect on adolescents’ prosocial behavior. Addressing recent calls for more person-centered approaches to capture the complexity of sport parenting,⁴⁵ we utilized LPA to identify four distinct profiles of family sport involvement. Moving beyond traditional methodologies that treat parental participation as a unidimensional or continuous variable,⁴⁶ our data indicate that the prosocial correlates of sport are fundamentally contingent upon the qualitative configuration of family sport involvement. While the “highly supportive profile” showed the strongest positive association with prosocial behavioral outcomes, critically, our findings reveal that it is the specific behavioral configuration, rather than mere involvement volume, that drives development. Viewed through a Bourdieusian lens, these quantitative profiles do not merely equate to parenting styles but reflect the behavioral manifestations of differentiated familial sport habitus.

Crucially, this study extends prior literature by identifying potential the specific psychological pathways through which these familial contexts relate to positive social outcomes. Although previous studies acknowledge a supportive environment as a vital foundation,⁴⁷ our results offer robust empirical evidence that its relationship is largely statistically mediated by the mobilization of internal psychological assets. Consistent with SDT,¹⁶ self-efficacy emerged as the mediating variable, operating in conjunction with autonomous motivation and active extracurricular exercise. This corroborates and extends recent evidence suggesting that internalized confidence and self-determined motivation in physical activity are closely linked to moral and prosocial functioning.⁴⁸

Furthermore, by identifying family SES as a significant moderator, our findings theoretically bridge the gap between sociological determinism and psychological agency. While previous literature has primarily focused on independently social class differences,³⁵ contextual factorsError: Reference source not found or SDT, Error: Reference source not found our integrated framework provides a dual perspective. A critical examination of the interaction plot reveals a sobering reality regarding macro-structural inequalities: the data confirm an amplifying role of objective capital rather than a compensatory one. Across the “disengaged”, “moderate involvement”, and “modeling-only” profiles, prosocial outcomes remain relatively indistinguishable between high- and low-SES families. However, a stark divergence emerges strictly within the “highly supportive” profile. The transition to a highly supportive familial sport habitus yields a significantly sharper increase in adolescents’ prosocial behavior for those from high-SES backgrounds. This highlights a critical structural inequality: while psychological agency (a supportive habitus) is beneficial, high-SES families possess the broader structural resources or cumulative capital—necessary to more effectively convert this familial support into tangible prosocial outcomes.

Taken together, these findings suggest that positive familial psychological interactions in sports can, to a significant extent, buffer against objective class disparities. This research thus answers contemporary calls to integrate micro-level psychological mechanisms with macro-level socio-ecological realities.⁴⁹ Critically, our data refute the idealistic notion that positive familial psychological interactions in sports can fully buffer against objective class disparities. While fostering an autonomy-supportive environment remains essential, the structural class gap actually widens at the highest levels of parental support. Consequently, from a practical behavioral management standpoint, future interventions must adopt a dual-pronged approach. Relying solely on parent-training programs to cultivate supportive home environments is insufficient, as low-SES families face a systemic ceiling effect in translating this support. Therefore, micro-level educational interventions for caregivers must be strictly coupled with macro-level structural supports, such as sustained financial subsidies and the equitable allocation of basic community public services,⁵⁰ to ensure that youths from socioeconomically disadvantaged backgrounds can reap the full developmental benefits of their families’ psychological investments.

Conclusion

This study significantly advances the sport psychology literature by elucidating the complex, multifaceted nature of family sport socialization through the integration of macro-level sociological concepts (Bourdieu) and micro-level psychological mechanisms (SDT). By moving beyond traditional variable-centered approaches, our findings highlight that the prosocial outcomes of youth sport stem not from parental participation frequency alone, but from the qualitative configuration of family involvement. From a macro-sociological perspective, this configuration reflects the structural quality of how sports are integrated into family life, which strongly aligns with Bourdieu’s notion of family cultural capital and habitus transmission. Crucially, at the micro-level. Crucially, we illuminate the underlying developmental mechanisms driven by SDT, revealing that a highly supportive family ecosystem acts as a critical facilitator. Specifically, this optimal family sport profile fosters prosociality by mobilizing adolescents’ internal psychological assets, particularly through the enhancement of exercise self-efficacy and the promotion of autonomous motivation. This confirms SDT’s central premise that supportive family sport environments are essential for internalizing positive values and sustaining self-determined behaviors.

Furthermore, integrating family SES into the model uncovers crucial socioeconomic disparities, demonstrating how the unequal opportunities afforded by economic conditions significantly amplify these developmental pathways leading to prosocial behavior. Practically, these insights highlight the need to transition away from generic “increase parental involvement” campaigns. Instead, evidence-based behavioral interventions must focus on targeted parent education that helps families transition toward the optimal family sport profile. Such programs should equip parents with specific strategies to cultivate autonomy-nurturing sporting climates, which our findings show are essential for directly enhancing adolescents’ exercise self-efficacy and autonomous motivation. Moreover, recognizing that lower-SES families may face structural constraints in independently providing such optimal environments, community-based public sport services and local initiatives must step in. These local initiatives should offer compensatory support specifically designed to boost self-efficacy and foster autonomous motivation among disadvantaged youth. to offer compensatory psychological and environmental support. Ultimately, optimizing the youth sport experience requires a synergistic approach, ensuring that sport serves as an equitable and empowering platform for cultivating prosociality across all social classes.

Limitations and Future Directions

While this study provides novel insights into family sport socialization, several methodological limitations warrant acknowledgment. First, the cross-sectional design precludes definitive causal inferences. Future studies employing longitudinal tracking or randomized controlled interventions are necessary to elucidate the temporal precedence and dynamic causal pathways of the observed psychosocial transfer effects. Second, relying exclusively on self-reported data may introduce common method variance and subjective reporting bias. Subsequent research would benefit from methodological triangulation, adopting multi-informant assessments alongside objective behavioral metrics, such as wearable accelerometry for physical activity tracking.

Beyond methodological constraints, future research must expand the demographic and contextual scope of this inquiry. The current sample’s representativeness is somewhat limited regarding the urban-rural divide and broader macro-regional disparities in China. Future investigations should prioritize sociodemographic inclusivity, specifically examining the family sport profiles of rural or socioeconomically vulnerable populations. Finally, this study did not account for the potential differential impacts of specific sport modalities (eg, team-based versus individual sports). Moving forward, researchers are encouraged to leverage Ecological Systems Theory to construct comprehensive, multi-level models that integrate family, peer, and school microsystems, thereby fully delineating the complex mechanisms underlying youth sport participation and positive socialization behavioral development.

Funding Statement

This study was supported by the Guangdong Philosophy and Social Sciences Planning (Grant: GD25YTY04) and The 2023 Fujian Provincial Social Science Fund Project (Grant: FJ2023B015).

Data Sharing Statement

The datasets analyzed in this study are archived in in Population Health Data Archive (https://doi.org/10.12213/11.A0031.202107.209.V1.0). The raw data are restricted from public access. However, researchers may obtain the data by submitting a formal request and a valid academic rationale through the repository’s official platform.

Ethics Approval

This study was approved by the Ethics Committee of Shandong University (approval number: 20180517). Written informed consent was obtained from the parents or legal guardians of all adolescent participants prior to data collection, and participants’ anonymity was preserved throughout the study. All data analyzed in the present study were de-identified and accessed via the National Population Health Science Data Center of China under official authorization. According to Article 32 of the Measures for Ethical Review of Life Science and Medical Research Involving Human Subjects (National Health Commission of China, February 18, 2023), ethics review is exempt for research involving: (1) utilizing legally obtained public data, or data generated through the observation of public behavior without interference; and (2) utilizing anonymized informational data. Because the present study meets these criteria, this secondary analysis was exempt from review by the Institutional Review Board of Guangzhou Sport University. All methods were performed following the relevant guidelines and regulations.

Authors’Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare no competing interests.