A multilevel mediation analysis of coaches’ life skills coaching,caring climate,and the development and transfer of life skills in student-athletes

Abstract

Sport provides a powerful platform for fostering life skills that can be applied out of sport contexts. Among the factors influencing this process, coaches’ life skills coaching and the perceived caring climate play a critical role. This study examined the multilevel relationships among coaches’ life skills coaching, student-athletes’ perceptions of a caring climate, life skills development, and transfer using a multilevel mediation model. Participants were 431 student-athletes from 28 Korean school sport teams (including a coach from each team). Coaches reported on their life skills coaching practices, while athletes assessed the caring climate, life skills development, and transfer. Multilevel mediation analysis was conducted to capture both within- and between-team effects. Results showed that coaches’ life skills coaching positively influenced athletes’ perceived caring climate and life skills development. Life skills development significantly predicted transfer and mediated the effect of coaching on transfer. However, the caring climate did not significantly mediate this relationship and showed a negative effect at the group level. These findings emphasize the indirect yet vital role of life skills coaching in promoting transfer and highlight the need for intentional, development-focused coaching strategies. Multilevel modeling offers a more precise understanding of how coaching influences athletes’ development within sport environments.

Keywords

Culture positive youth development transfer

Introduction

Life is like an invisible playing field in which sport unfolds. The process of complying with established rules and cooperating with teammates resembles the way athletes follow the rules of sport and work collaboratively with their peers.¹ Scholars have long sought to articulate the theoretical foundations underlying the positive values and experiences inherent in sport.^2,3 Among the various frameworks, the Positive Youth Development (PYD) approach has received widespread support for its emphasis on the skills and competencies cultivated through sport participation.⁴ These abilities and competencies are commonly conceptualized under the umbrella of life skills.

Life skills refer to behavioral, emotional, and cognitive strategies that enable individuals to actively cope with the diverse challenges and adversities encountered throughout life. Representative examples include communication skills for effective interpersonal interaction, goal-setting skills for pursuing life direction and purpose, and problem-solving skills for managing unforeseen obstacles. Importantly, the development of life skills through sport aims to facilitate their transfer across multiple life domains, including home, school, and everyday settings.⁵ This process is referred to as life skills transfer. For instance, a student-athlete who applies the goal-setting strategies used in sports to academic pursuits demonstrates such transfer.

The body of research on life skills has evolved from initial conceptual explorations^6,7 to the development of theoretical models,^3,5 and further to the application of these models in practice through program implementation.^2,8,9 More recent studies emphasize the importance of structured programs and the role of significant others in shaping life skills outcomes. In particular, life skills development and transfer are influenced by the sociocultural environment surrounding the athlete.¹⁰ A psychologically and physically safe setting that fosters a positive climate has been shown to enhance life skills outcomes.

One notable environmental factor is the presence of a caring climate, characterized by mutual warmth, respect, and support among athletes. This climate has been identified as a key predictor of positive development within youth sport contexts.¹¹ A meta-analysis by Holt et al.⁴ confirmed that a caring climate contributes to PYD, facilitates life skills transfer, and promotes the formation of positive life attitudes. Similarly, Gould, Flett, and Lauer¹² reported that both a caring and task-involved climate, especially when fostered by coaches, positively influence youth development. Cronin and Allen¹³ also found that positive coaching climates significantly impact life skills development among student-athletes. Collectively, these findings suggest that student-athletes’ perceptions of a caring climate can play a critical role in facilitating life skills development and transfer.

Coaches are consistently identified as central figures in the life skills development of student-athletes.¹⁴ The extent to which coaches focus on performance outcomes versus life skills instruction significantly influences athletes’ developmental trajectories.¹⁵ The concept of life skills coaching refers to a coach's deliberate use of both direct and indirect strategies, grounded in a developmental philosophy, to promote life skills acquisition and transfer.^16,17 According to Bean et al.,¹⁸ effective life skills coaching involves intentional instruction—offering athletes opportunities to practice skills and reflect on their applicability beyond sport. For example, coaches may guide athletes to use deep breathing or positive self-talk to manage pre-competition anxiety, and then prompt them to consider how these strategies might be applied in academic or social contexts. Numerous empirical studies have highlighted the significant influence of life skills coaching on student-athletes’ development and transfer of life skills. Pierce et al.¹⁹ identified life skills coaching as a core element in promoting such outcomes among youth sport participants. Likewise, Bae et al.²⁰ found that life skills coaching plays a key role in facilitating both the development and transfer of life skills. Complementary qualitative studies employing structured program interventions^17,21 have also confirmed that appropriate coaching strategies serve as catalysts for positive skill development.

Taken together, these findings suggest a causal relationship between coaches’ life skills coaching, athletes’ perceptions of a caring climate, and the development and transfer of life skills.⁶ This pattern is further supported by recent research targeting youth and adolescent athletes.^13,19,20 Nevertheless, additional empirical studies are needed to provide more robust evidence of the influence of life skills coaching on perceived life skills outcomes and their transfer. A key limitation of existing research lies in its reliance on cross-sectional quantitative designs that fail to account for nested data structures. That is, although athletes’ perceptions of the team climate and life skills may be influenced by the life skills coaching practices of their coaches, most studies have not distinguished between levels of sampling units.^22,23

To address this issue, multilevel modeling (MLM) provides a promising analytic approach.^24,25 Unlike traditional regression models, MLM accounts for the hierarchical structure of data, allowing for the simultaneous examination of within-group and between-group effects. This requires the differentiation of variables across two levels: Level 2 (e.g., coaches) and Level 1 (e.g., individual athletes). In the present study, coach-level data were modeled at Level 2, while athlete-level perceptions were analyzed at Level 1. This multilevel framework enables a more precise evaluation of both direct and indirect effects operating within and across levels. By doing so, this research aims to overcome limitations of prior cross-sectional designs, offering deeper insight into the actual impact of life skills coaching, perceived team climate, and the subsequent development and transfer of life skills. Furthermore, the use of MLM reduces the risk of ecological fallacies by accounting for both individual and group-level variations in sport team environments.

Present study

Research on life skills in sport has progressed considerably, with a growing body of work bridging theory and practice. As empirical evidence has accumulated, the field has evolved from an early focus on individual attributes toward greater recognition of the role of significant others in shaping athletes’ developmental outcomes. Among these influential figures, coaches have garnered particular attention due to their substantial impact on student-athletes’ experiences. However, many existing findings have been derived from analyses at the individual level, which fail to account for potential error variance stemming from group-level influences. This limitation highlights the need for analytic approaches that can capture the nested nature of sport environments. Accordingly, the present study aimed to examine the relationships among coaches’ perceptions of life skills coaching (Level 2), student-athletes’ perceptions of a caring climate (Level 1), and their life skills development and transfer (Level 1) using a MLM approach (see Figure 1). This analytic framework allows for the simultaneous investigation of coach- and athlete-level variables while addressing the hierarchical structure inherent in sport teams.

Figure 1.

Hypothesized research model (H, hypothesis).

Figure 2.

Direct and indirect estimates with research model (*p < 0.05, **p < 0.01, ***p < 0.001).

To achieve the research aim, the following hypotheses were formulated. Hypothesis 1: Student-athletes’ (Level 1) perceptions of a caring climate will positively predict their life skills transfer. Hypothesis 2: Student-athletes’ (Level 1) perceived life skills development will positively predict their life skills transfer. Hypothesis 3: Coaches’ (Level 2) life skills coaching will positively predict student-athletes’ (Level 1) perceptions of a caring climate. Hypothesis 4: Coaches’ (Level 2) life skills coaching will positively predict student-athletes’ (Level 1) perceived life skills development. Hypothesis 5: Coaches’ (Level 2) life skills coaching will positively predict student-athletes’ (Level 1) perceived life skills transfer. Hypothesis 6: The relationship between coaches’ (Level 2) life skills coaching and student-athletes’ (Level 1) perceived life skills transfer will be mediated by perceptions of a caring climate. Hypothesis 7: The relationship between coaches’ (Level 2) life skills coaching and student-athletes’ (Level 1) perceived life skills transfer will be mediated by perceived life skills development.

Methods

Participants

Participants in this study consisted of coaches and athletes affiliated with school sports teams in South Korea. These school sports teams represent a distinctive model of sport participation in the Korean context, designed to foster the development of student-athletes into elite-level competitors. In South Korea, student-athletes are students who, while enrolled in school, primarily engage in intensive training and competition within a state-led elite sport system that prioritizes athletic performance over academic education.²⁶ Data were collected from 35 school sports teams using a cluster sampling method. Each participating unit included one coach and the athletes under their guidance. Because the unit of data collection needed to include the group level, any teams for which incomplete or insincere athlete responses resulted in fewer than the minimum required number of athletes (i.e., fewer than five) were excluded from the final analysis, and both the coach and athlete data from those teams were removed. The reason for applying a minimum of five athletes per team is to obtain reliable group means for calculating team level variables. Simulation studies report that two level models can be estimated reliably when the average cluster size is five or more,²⁷ and that clusters smaller than five may overestimate group level variance. Therefore, considering the team size distribution in our data (5–27 athletes, mean = 15.4), teams with fewer than five athletes were excluded.

Coaches’ responses were treated as Level 2 data, while athletes’ responses were treated as Level 1 data. The 28 coaches were each affiliated with different teams and were responsible for training their respective athletes. The coaches represented the following sports: soccer (n = 10, 35.72%), taekwondo (n = 10, 35.72%), golf (n = 2, 7.14%), archery (n = 2, 7.14%), ice hockey (n = 2, 7.14%), and judo (n = 2, 7.14%). Their average age was 34.21 years (SD = 5.33), and their average coaching experience was 9.30 years (SD = 4.52). The athlete sample consisted of 431 students, including 211 males (48.96%) and 120 females (27.84%). The distribution by sport was as follows: soccer (n = 179, 46.14%), taekwondo (n = 153, 39.43%), golf (n = 10, 2.57%), archery (n = 9, 2.32%), ice hockey (n = 13, 3.35%), and judo (n = 24, 6.19%). The athletes’ average age was 15.35 years (SD = 1.57), and their average athletic career length was 4.53 years (SD = 2.51).

Measures

Life skills coaching

To assess coaches’ perceptions of life skills coaching (Level 2 data), the Korean version of the Coaching Life Skills in Sport Questionnaire (CLSS-Q),²⁸ originally developed by Camiré et al.¹ was used. The scale consists of 26 items across five subscales: structuring and facilitating a positive sport climate (9 items), discussing life skills (5 items), practicing life skills (4 items), discussing transfer (5 items), and practicing transfer (3 items). Participants responded using a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). To verify multilevel reliability, both Cronbach's α and McDonald's ω were calculated. The results indicated acceptable to excellent internal consistency: structuring and facilitating a positive sport climate (α = 0.70, ω = 0.75), discussing life skills (α = 0.79, ω = 0.75), practicing life skills (α = 0.78, ω = 0.82), discussing transfer (α = 0.88, ω = 0.87), and practicing transfer (α = 0.87, ω = 0.88). The overall reliability of the life skills coaching score was α = .89 and ω = .92.

Caring climate

To measure athletes’ perceptions of the caring climate (Level 1 data), the Korean version of the Caring Climate Scale (CCS),²⁹ developed by Newton et al.³⁰ was employed. This unidimensional scale includes 13 items. Participants responded using a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). Reliability analysis confirmed strong internal consistency, with Cronbach's α = 0.92 and McDonald's ω = 0.92.

Life skills development and transfer

To assess athletes’ perceptions of life skills and their transfer beyond sport, the Korean version of the Life Skills Scale for Sport – Transfer Scale (LSSS-TS),³¹ developed by Mossman et al.³² was used. This instrument concurrently evaluates both life skills and their transfer across life domains. The life skills component consists of 34 items across eight subscales: goal setting (5 items), time management (4 items), emotional skills (4 items), interpersonal communication (4 items), social skills (4 items), leadership (4 items), teamwork (5 items), and problem solving and decision-making (4 items). Internal consistency for each subscale was as follows: goal setting (α = 0.88, ω = 0.88), time management (α = 0.88, ω = 0.88), emotional skills (α = 0.85, ω = 0.86), interpersonal communication (α = 0.87, ω = 0.87), social skills (α = 0.89, ω = 0.89), leadership (α = 0.87, ω = 0.87), teamwork (α = 0.92, ω = 0.92), and problem solving and decision-making (α = 0.92, ω = 0.92). The overall reliability of the life skills development score was α = .92 and ω = .92.

The transfer component evaluates how life skills are applied in various life contexts (e.g., home, school, peer relationships). Each of the eight life skills domains includes four items, totaling 32 items: goal setting, time management, emotional skills, interpersonal communication, social skills, leadership, teamwork, and problem solving and decision-making. The internal consistency for each subscale was high: goal setting (α = 0.89, ω = 0.89), time management (α = 0.90, ω = 0.90), emotional skills (α = 0.93, ω = 0.93), interpersonal communication (α = 0.92, ω = 0.92), social skills (α = 0.94, ω = 0.94), leadership (α = 0.87, ω = 0.88), teamwork (α = 0.93, ω = 0.93), and problem solving and decision-making (α = 0.93, ω = 0.93). All items were measured using a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). The overall reliability of the life skills transfer score was α = .93 and ω = .93.

Research procedure

To recruit participants, the research team contacted stakeholders involved in school athletic teams across South Korea. After providing a detailed explanation of the study's purpose and procedures, data were collected from teams that voluntarily agreed to participate. Visits to each team were arranged in consultation with the coaches. Prior to data collection, both coaches and athletes were informed about the ethical considerations and objectives of the research, and only those who voluntarily consented participated in the study. For minor athletes, parental consent was obtained in accordance with ethical guidelines. Each participant was assigned a unique identification number to match the data of coaches and athletes from the same team. All collected data were carefully reviewed and cleaned to ensure consistency and accuracy prior to analysis.

Data analysis

All statistical analyses were performed using SPSS Version 23. Prior to the analyses, all datasets were screened for careless or insufficient responses, and such cases were removed during the initial data cleaning process. Missing data were minimal across variables (< 5%). Little's Missing Completely at Random (MCAR) test indicated that the missingness was completely at random; therefore, remaining missing values were handled using mean substitution. Outliers were examined using standardized z-scores (±3.29), and no values exceeded this threshold, so all valid cases were retained for analysis.

Descriptive statistics were calculated to examine the normality of the data, including mean, standard deviation, skewness, and kurtosis. Normality was assumed when the absolute values of skewness and kurtosis were below ±3 and ±8, respectively.³³ Pearson's product-moment correlation coefficients (r) were computed to assess bivariate relationships among study variables. Multicollinearity was evaluated using the same correlation matrix; correlation coefficients below .80 were considered acceptable.³⁴ To examine the nested structure of the data and assess between-group variability, ICC were calculated using a linear mixed model.³⁵ To test the study hypotheses, a multilevel mediation model was conducted using the MLMed macro for SPSS.²⁵ This macro simultaneously estimates both fixed effects (representing the consistency of relationships within and between groups) and random effects (reflecting individual or group-level variance). The random effects analysis helps identify natural variation in intercepts across individuals or teams.

All variables were standardized before entering into the model to facilitate interpretation and comparison. Group-level effects were estimated using group means, while within-group effects were estimated by centering variables at the group level.³⁶ For both indirect and contextual indirect effects, the MLMed macro generated 95% Monte Carlo confidence intervals based on 10,000 resamples. An indirect effect was considered statistically significant if the confidence interval did not include zero.

Results

Preliminary analyses

To examine the data structure of the present study (i.e., 28 teams with a minimum of five athletes per team), ICC and design effects were calculated for the main variables. The ICC were 0.12 for caring climate, 0.09 for life skills development, and 0.10 for life skills transfer. Based on these ICC, the corresponding design effects were 2.73, 2.30, and 2.44, respectively, indicating meaningful between-team variance and non-independence of observations.³⁷ To verify the normality and multicollinearity of the collected data, a series of analyses were conducted. The mean scores of life skills coaching, caring climate, life skills development, and transfer ranged from 3.99 to 4.53. Among them, the transfer variable had the lowest mean score, while the caring climate showed the highest mean score. The standard deviations ranged from 0.35 to 0.61. Skewness values were between −0.14 and −0.10, and kurtosis values ranged from −0.90 to −0.53. According to the criteria suggested by Kline,³³ these results support the assumption of normality.

Meanwhile, all measurement variables showed positive correlations. The strongest correlation was found between life skills development and transfer (r = 0.76, p < 0.01). In contrast, the lowest correlation was observed between the Level 2 variable (i.e., life skills coaching) and the Level 1 variables (i.e., caring climate and transfer), with the lowest being between caring climate and transfer (r = 0.21, p < 0.01). Most importantly, no correlations exceeded the threshold of 0.80, indicating that multicollinearity was not a concern (Tables 1).

Multilevel mediation model analysis

Intraclass correlation coefficient

The ICC indicates the extent to which individuals within the same group resemble one another. A high ICC value suggests a substantial similarity among individuals within groups, implying significant between-group differences. When the ICC is high, it means that a large portion of the total variance is attributable to between-group differences, thereby supporting the use of a multilevel model rather than a single-level model.

To calculate the ICC, a linear mixed model was conducted with transfer as the dependent variable. The residual variance was estimated at 0.33 (SE = 0.02, Wald Z = 13.35, p < 0.001, 95% CI [0.29, 0.39]), and the intercept variance was 0.04 (SE = 0.02, Wald Z = 1.99, p < 0.05, 95% CI [0.01, 0.10]). Based on these estimates, the ICC was calculated as 0.89, indicating that approximately 89.2% of the variance in transfer is explained by between-group differences. This result suggests a high degree of similarity among individuals within the same group and substantial variability between groups.

Testing hypothesis

Model fit indices indicated that the proposed multilevel mediation model adequately fit the data. The −2 log-likelihood (−2LL) value was 2893.71, with AIC = 2905.71 and BIC = 2936.01, supporting the model's overall explanatory power. Results from the fixed effects analysis showed that life skills development at the individual level (Level 1) had a significant positive effect on transfer (Estimate = 0.78, p < 0.001). In contrast, the direct effect of the caring climate on transfer was not significant at the individual level (Estimate = 0.03, p > 0.05) and was significantly negative at the between-group level (Estimate = −0.41, p < 0.05).

At the group level (Level 2), life skills coaching was found to have a significant positive effect on both caring climate (Estimate = 0.20, p < 0.01) and life skills development (Estimate = 0.31, p < 0.001). However, the direct effect of life skills coaching on transfer was not statistically significant (Estimate = 0.02, p > 0.05). The residual variances for the Level 1 variables were all statistically significant: caring climate (Estimate = 0.90, SE = 0.07, p < 0.001, Wald Z = 13.47), life skills development (Estimate = 0.87, SE = 0.07, p < 0.001, Wald Z = 13.41), and transfer (Estimate = 0.36, SE = 0.03, p < 0.001, Wald Z = 13.41). At Level 2, only the random intercept for transfer was significant (Estimate = 0.05, SE = 0.02, p < 0.05), indicating meaningful group-level variation in transfer and suggesting that both individual- and group-level factors may contribute to this outcome.

The indirect effect of life skills coaching on transfer through life skills development was statistically significant (Estimate = 0.29, p < 0.001), confirming that life skills coaching influences transfer indirectly via enhanced life skills development. In contrast, the indirect effect through caring climate was not statistically significant (Estimate = −0.08, p > 0.05). A comparison of the two indirect effects revealed that the indirect pathway through life skills development was not only significant but also stronger than the pathway through caring climate (Estimate = 0.37, 95% CI [0.18, 0.60]) (Tables 2, Figure 2).

Discussion

This study examined the relationships among coaches’ life skills coaching, athletes’ perceived caring climate, life skills development, and transfer using a multilevel mediation model. Such a research design makes a valuable theoretical and practical contribution by empirically validating the role of significant others—frequently emphasized in sport psychology and PYD literature—within a nested data structure. Importantly, by reflecting both within- and between-group differences, the study helps prevent ecological fallacies that can arise in single-level analyses, thereby offering a more refined understanding of the multilevel relationship between coaching and youth development.

The findings suggest that coaches’ life skills coaching significantly facilitated athletes’ life skills development, which in turn strongly predicted the transfer of those skills. Although the direct effect of life skills coaching on transfer was not significant, the indirect effect—mediated by life skills development—was both statistically significant and substantial. This suggests that life skills coaching does not directly lead to transfer; rather, it systematically develops athletes’ life skills, which then enable them to apply those skills in other life domains. These results support and extend prior research^15,18,19 by empirically confirming the efficacy of life skills coaching within a multilevel framework. Consequently, coaches can be seen not only as facilitators of athletic performance but also as educational agents who contribute to athletes’ holistic development and social adjustment.

Unexpectedly, the role of the caring climate differed from prior assumptions. At the individual level (Level 1), the direct effect of caring climate on transfer was not significant, and at the group level (Level 2), it was found to have a negative effect. This finding contrasts with previous literature,^11,13 which reported positive associations between a caring climate and athletes’ psychological well-being and development. Several factors may explain this discrepancy. First, there may have been a perceptual mismatch between the coach's intentions and the athletes’ interpretations. Even if coaches intended to create a positive atmosphere, athletes may have perceived these efforts as controlling or intrusive. This phenomenon suggests the possibility highlighted in previous research that when the coach–athlete relationship functions smoothly, coaching effectiveness can be enhanced.³⁸ Second, caring climate may not directly influence transfer; instead, it may promote emotional safety and psychological stability, which in turn support life skills development. Third, cultural factors unique to Korean sport—such as its competitive focus and hierarchical structure—may lead athletes to interpret warmth and empathy in less favorable terms. These contextual nuances highlight the need for future research to refine the conceptualization of caring climate and ensure cultural validity in its measurement.

A key methodological contribution of this study lies in its application of a multilevel model to empirically examine the structure at the group level. The ICC for life skills transfer was 0.89, indicating that approximately 89% of the total variance was explained by between-group differences. This suggests that athletes who were coached by the same individual shared similar experiences and team cultures, and that team-based experiences had a greater impact on transfer than individual differences. The ICC serves as a quantitative indicator of intra-group similarity or inter-group variability,²³ demonstrating the extent to which group members resemble one another. Given the exceptionally high ICC found in this study, it may be interpreted that athletes’ transfer of life skills is shaped more significantly by the team environment, such as coaching climate or team culture, than by individual characteristics. This suggests that learning and developmental experiences are commonly shared within teams, and that coaching strategies are disseminated and reinforced at the team level rather than tailored solely to individuals. Such findings underscore the importance of capturing group effects that are often overlooked in single-level analyses and provide a robust theoretical and methodological rationale for the use of multilevel modeling.

Moreover, the findings underscore the importance of team-based interventions, indicating that future life skills coaching programs should be designed with group-level changes in mind. From a practical standpoint, this study offers the following implications. First, coaches should view life skills not merely as tools to enhance athletic performance but as essential competencies that can be transferred to broader life domains. To achieve this, coaches must establish clear goals for life skills development, provide repeated practice opportunities, facilitate reflection, and offer feedback that encourages real-world application. For example, practicing deep breathing before a crucial game may serve as a strategy for managing stress in daily life—an embodiment of coaching with educational intent. Second, while a caring climate serves as an emotional foundation for maximizing coaching effectiveness, it is not sufficient on its own to guarantee transfer. Therefore, caring should be understood as a means rather than an end, and it should be combined with effective strategies for developing life skills.

This study also highlights how shared developmental experiences at the team level can shape athletes’ socio-emotional environments, emphasizing the impact of a coach's language, attitudes, and feedback style.³⁹ As noted in Bean et al.¹⁷ explicit and intentional coaching plays a pivotal role in facilitating the internalization and transfer of learned life skills. Consequently, coaching strategies should integrate both individualized approaches tailored to athletes’ personal needs and team-wide approaches that foster a supportive group culture.⁴⁰ Finally, the results of this study reinforce the idea that sport is not merely a venue for acquiring physical skills, but a vital educational context where young people can learn to live. Coaches must function as facilitators of PYD, and future program design should reflect this educational philosophy. Furthermore, from a policy perspective, these findings call for increased recognition of the value of life skills coaching and the systematic integration of such content into coach education and development programs.

Limitations and future direction

This study analyzed the relationships among life skills coaching, caring climate, life skills development, and transfer using a multilevel model, thereby deriving theoretical and practical implications. Nevertheless, several limitations exist, which provide directions for future research. First, although the ICCs and design effects of the main variables confirmed the appropriateness of the data structure, the sample size remains a limitation in terms of statistical optimality. Specifically, since the Level-2 sample (n = 28 teams) was smaller than the commonly recommended threshold of 30, Level-2 standard errors may have been underestimated, and variations in team sizes could have influenced the stability of the results. Therefore, caution is required when interpreting the findings, and future studies should include a larger number of teams across a wider range of sports to secure a more robust Level-2 sample.

Second, since this study focused on student-athletes in Korea, the generalizability of the findings is limited. As noted in the discussion, the unique cultural characteristics of Korean youth sport—such as its competitive orientation and hierarchical coach–athlete relationships—may have shaped athletes’ perceptions of caring climate and life skills transfer. Therefore, the generalizability of our findings may be limited, and future studies should examine whether these patterns hold in different cultural contexts. These findings also highlight the need for practitioners to consider cultural values and communication norms when designing life skills coaching strategies. Coaching behaviors that are perceived as supportive in one cultural context may be interpreted differently in another; thus, culturally responsive approaches should be integrated into coach education programs.

Third, although the use of a multilevel model is a strength in that it accounts for group-level structures and controls for ecological fallacy, focusing on team-level data also limits the ability to explain individual athletes’ developmental changes or personal differences. For instance, even under the same coach's guidance, athletes’ life skills development processes may vary depending on individual experiences, motivation, and psychological characteristics. Future studies should incorporate psychosocial factors at the individual level (e.g., self-efficacy, motivation types, coaching receptivity) within multilevel modeling designs.

Regarding measurement, while this study employed a well-defined conceptualization of life skills coaching, it did not conduct a detailed analysis of the specific stages of life skills coaching as outlined by Bean et al.¹⁸ Future research should differentiate the effects of each coaching stage on athlete development to clarify the specific strategies and effectiveness of life skills coaching. Lastly, despite applying a multilevel model, the causal pathways among caring climate, life skills development, and transfer lacked clear structural explanation. Although caring climate did not demonstrate a significant mediating effect on transfer in this study, it suggests that caring may operate indirectly or through intermediate variables. Therefore, expanded models incorporating mediators such as emotional stability, team cohesion, and social connectedness should be explored in future research.

Conclusion

This study analyzed the effects of coaches’ life skills coaching on student-athletes’ perceptions of caring climate, life skills development, and life skills transfer using a multilevel model. The results showed that life skills coaching positively influenced life skills development, and life skills development emerged as a key predictor of transfer. In particular, life skills development mediated the relationship between life skills coaching and transfer, suggesting that coaching can have a meaningful impact on athletes’ lives beyond sports. On the other hand, caring climate did not exhibit a significant mediating effect as expected, and at the group level, it showed a negative effect. This may indicate that the impact of caring climate operates indirectly through psychological and behavioral factors such as life skills development or reflects perceptual differences influenced by cultural context. These findings highlight the important role coaches should play as facilitators of youth development. Practically, deliberate coaching strategies and reflective learning experiences are necessary to design programs that promote positive life changes for athletes. Future research should employ longitudinal designs and incorporate diverse cultural contexts to build more generalizable evidence.

Table 1.

Descriptive statistics and correlation coefficients.

Factor	1	2	3	4
1. Life Skills Coaching	1
2. Caring Climate	0.22**	1
3. Life Skills Development	0.31**	0.57**	1
4. Life Skills Transfer	0.22**	0.42**	0.77**	1
Mean	4.20	4.53	4.23	4.00
Standard deviation	0.35	0.49	0.52	0.61
Skewness	−0.14	−1.11	−0.10	−0.13
Kurtosis	−0.90	0.55	−0.87	−0.54

Note. All questionnaires were assessed using a 5-point Likert scale.

**p < 0.01

Table 2.

Parameter estimates for multilevel mediation model.

Paths	Estimates	SE	t / Z	95% CI
Paths	Estimates	SE	t / Z	LL	UL
Within effects
Caring¹ → LST¹	0.03	0.04	^t 0.69	−0.05	0.11
LSD¹ → LST¹	0.78	0.04	^t 19.08***	0.70	0.86
Between effects
Caring¹ → LST¹	−0.41	0.18	^t −2.29*	−0.77	−0.04
LSD¹ → LST¹	0.93	0.19	^t 5.05***	0.55	1.32
LSC² → Caring¹	0.20	0.07	^t 2.04**	0.06	0.34
LSC² → LSD¹	0.31	0.06	^t 4.89***	0.18	0.44
LSC² → LST¹	0.02	0.07	^t 0.29	−0.13	0.17
Random effects (Level 1 residual)
Caring	0.05 (0.90)	0.04 (0.07)	^Z1.51 (13.47***)	0.01 (0.78)	0.19 (1.05)
LSD	0.04 (0.87)	0.03 (0.07)	^Z1.27 (13.41***)	0.01 (0.75)	0.20 (1.01)
LST	0.05 (0.36)	0.02 (0.03)	^Z2.21* (13.41***)	0.02 (0.31)	0.13 (0.41)
Between indirect effects
LSC → Caring → LST	−0.08	0.05	^Z−1.75	−0.18	−0.01
LSC → LSD → LST	0.29	0.08	^Z3.48***	0.14	0.47
Between Indirect Effect Contrasts	0.37			0.18	0.60
Model fit statistics
−2 log-likelihood (−2LL)	2893.71
Akaike Information Criterion (AIC)	2905.71
Bayesian Information Criterion (BIC)	2936.01

*p < 0.05, **p < 0.01, ***p < 0.001

Note. All confidence intervals at Monte-Carlo 10,000 samples; Estimator = REML; LST refers to life skills transfer; LSD refers to life skills development; LSC refers to life skills coaching; ¹ refers to level 1 variables; ² refers to level 2 variables; ^t refers to t scores; ^Z refers to Wald Z

Footnotes

Acknowledgements

The authors would like to express their sincere gratitude to all those who contributed to the successful completion of this study. This research was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF). We extend our heartfelt thanks to the NRF for their positive evaluation of our research proposal. We are also deeply grateful to our fellow researchers who provided valuable advice to enhance the quality of this work, as well as to the participants who generously devoted their time despite their busy schedules.

ORCID iDs

Junsu Bae

Seung-Hwan Woo

Ethical considerations

The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Institutional Review Board (IRB) of Yongin University (IRB No. 2-1040966-AB-N-01-2405-HR-340-1).

Consent to participate

All participants gave their informed consent for inclusion before they participated in the study.

Consent for publication

This study does not include any personally identifiable information such as names, images, or video recordings. Prior informed consent for the academic publication of the research findings was obtained from all participants. Therefore, this item is not applicable to the present study.

Authorship contributions

The authors confirm the following contributions to the paper: the first, second, and third authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by the fourth and fifth authors. The first draft of the manuscript was written by the first and fifth authors, and the fourth author provided comments on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2023S1A5A2A01081605).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy issues.

References

Camiré

Turgeon

Kramers

, et al. Development and initial validation of the coaching life skills in sport questionnaire. Psychol Sport Exerc 2021; 53: 101845.

Bae

Cho

Lim

. Examining the role of life skills in mediating the relationship between the basic psychological needs and subjective well-being of Taekwondo student-athletes. Int J Environ Res Public Health 2021; 18: 11538.

Jacobs

Wright

. Transfer of life skills in sport-based youth development programs: a conceptual framework bridging learning to application. Quest 2018; 70: 81–99.

Holt

Neely

Slater

, et al. A grounded theory of positive youth development through sport based on results from a qualitative meta-study. Int Rev Sport Exerc Psychol 2017; 10: 1–49.

Pierce

Gould

Camiré

. Definition and model of life skills transfer. Int Rev Sport Exerc Psychol 2017; 10: 186–211.

Hodge

Danish

Martin

. Developing a conceptual framework for life skills interventions. Couns Psychol 2013; 41: 1125–1152.

Holt

Deal

Smyth

. Future directions for positive youth development through sport. In: Positive youth development through sport. Rondon: Routledge, 2016: 45–47.

Hardcastle

Tye

Glassey

, et al. Exploring the perceived effectiveness of a life skills development program for high-performance athletes. Psychol Sport Exerc 2015; 16: 139–149.

Weiss

Bolter

Kipp

. Evaluation of The First Tee in promoting positive youth development: group comparisons and longitudinal trends. Res Q Exerc Sport 2016; 87: 271–283.

10.

Gano-Overway

Carson

. The Mapp way: success through combining the motivational and caring climates. J Appl Sport Psychol 2021; 33: 238–258.

11.

Gano-Overway

Guivernau

. Setting the SCENE: developing a caring youth sport environment. J Sport Psychol Action 2018; 9: 83–93.

12.

Gould

Flett

Lauer

. The relationship between psychosocial developmental and the sports climate experienced by underserved youth. Psychol Sport Exerc 2012; 13: 80–87.

13.

Cronin

Allen

. Examining the relationships among the coaching climate, life skills development and well-being in sport. Int J Sport sci Coach 2018; 13: 815–827.

14.

Kramers

Thrower

Steptoe

, et al. Parental strategies for supporting children’s psychosocial development within and beyond elite sport. J Appl Sport Psychol 2023; 35: 498–520.

15.

Camiré

Trudel

Forneris

. Coaching and transferring life skills: philosophies and strategies used by model high school coaches. Sport Psychol 2012; 26: 243–260.

16.

Turnnidge

Côté

Hancock

. Positive youth development from sport to life: explicit or implicit transfer? Quest 2014; 66: 203–217.

17.

Bean

Kendellen

Forneris

. Moving beyond the gym: exploring life skill transfer within a female physical activity-based life skills program. J Appl Sport Psychol 2016; 28: 274–290.

18.

Bean

Kramers

Forneris

, et al. The implicit/explicit continuum of life skills development and transfer. Quest 2018; 70: 456–470.

19.

Pierce

O’Neil

Camiré

, et al. Examining the relationship between perceived coaching approaches for life skills development and life skills outcomes for high school athletes. Int Sport Coach J 2024; 1: 1–12.

20.

Bae

Lim

O’Sullivan

. The role of sport coaches on student-athletes’ life skills development and transfer: using multilevel model analysis. Int J Sport Sci Coach 2024; 19: 944–955.

21.

Weiss

Kipp

Reichter

, et al. Evaluating girls on the run in promoting positive youth development: group comparisons on life skills transfer and social processes. Pediatr Exerc Sci 2020; 32: 172–182.

22.

Newland

Newton

Moore

EWG

, et al. Transformational leadership and positive youth development in basketball. Int Sport Coach J 2019; 6: 30–41.

23.

Raudenbush

Bryk

. Hierarchical linear models: Applications and data analysis methods (Vol. 1). USA: Sage, 2002.

24.

Hou

, et al. Entrepreneurial team knowledge diversity and creativity: a multilevel analysis of knowledge sharing, individual creativity, and team creativity. Front Psychol 2021; 12: 717756.

25.

Rockwood

. Advancing the formulation and testing of multilevel mediation and moderated mediation models. Master's thesis, The Ohio State University, 2017.

26.

Park

Lim

. A chronological review of the development of elite sport policy in South Korea. Asia-Pac J Sport Soc Sci 2015; 4: 198–210.

27.

Theall

Scribner

Broyles

, et al. Impact of small group size on neighbourhood influences in multilevel models. J Epidemiol Community Health 2011; 65: 688–695.

28.

Bae

Lim

TH.

The validation of Korean Coaching Life Skills in Sport Questionnaire (KCLSS-Q). Korean J Sport Sci 2022; 33: 382–395.

29.

Sim

. Validation of the Korean version of perceived caring climate scale and its influence on elementary school students’ attitudes toward physical education class. KSEPE 2012; 18: 133–142.

30.

Newton

Fry

Watson

, et al. Psychometric properties of the caring climate scale in a physical activity setting. Rev de Psicol Deporte 2007; 16: 67–84.

31.

Bae

Yang

Woo

, et al. Validation of the Korean version of the Life Skills Scale for Sport–Transfer Scale (KLSSS-TS). KCDC 2025; 27: 96–111.

32.

Mossman

Robertson

Williamson

, et al. Development and initial validation of the life skills scale for sport–transfer scale (LSSS-TS). Psychol Sport Exerc 2021; 54: 101906.

33.

Kline

. Principles and practice of structural equation modeling. New York: Guilford publications, 2023.

34.

Grewal

Côté

Baumgartner

. Multicollinearity and measurement error in structural equation models: implications for theory testing. Mark Sci 2004; 23: 519–529.

35.

Junça-Silva

Mosteo

Lopes

. The role of mindfulness on the relationship between daily micro-events and daily gratitude: a within-person analysis. Pers Individ Dif 2023; 200: 111891.

36.

Ikhide

Adedapo

Moin

. Cognitive diversity and initiative-taking in crisis: a multisource multilevel dual-path model. Small Group Res 2024; 56: 635–664.

37.

Maas

Hox

. Sufficient sample sizes for multilevel modeling. Methodology (Gott) 2005; 1: 86–92.

38.

Jowett

. Coaching effectiveness: the coach–athlete relationship at its heart. Curr Opin Psychol 2017; 16: 154–158.

39.

Côté

Bruner

Erickson

, et al. Athlete development and coaching. Sports coaching: Professionalisation and practice, 2010.

40.

MacDonald

Côté

Eys

, et al. The role of enjoyment and motivational climate in relation to the personal development of team sport athletes. Sport Psychol 2011; 25: 32–46.