This study conducted a web-based questionnaire survey among high-level young athletes. To collect research data, the researchers employed convenience sampling and snowball sampling methods. After obtaining approval from local authorities, the researchers reached out to athletes from four universities, youth training centers, and high-level clubs in different provinces of central and southern China. Using social media platforms such as QQ and WeChat, the researchers inquired about their interest in participating in the study and sent the questionnaire tool link, inviting them to participate sincerely. All participants were informed of the survey’s purpose and volunteered before completing the questionnaire. Additionally, based on the snowball method, athletes who completed the questionnaire were encouraged to invite their friends and teammates to participate in the survey. Participants who completed the survey were then given training equipment as a token of appreciation. By the end of July, we had distributed 600 questionnaires and successfully collected responses from 482 athletes. After eliminating invalid questionnaires (due to data anomalies, missing information, blank responses, etc.), a total of 403 valid questionnaires were obtained, resulting in a valid response rate of 62.7%.

Table 1 presents the demographic characteristics of the 403 student athletes participating in this survey. Among the respondents, (1) the male-to-female ratio was approximately two to one; (2) 69% of the participants were under the age of 23; (3) athletes participating in badminton, soccer, basketball, and track and field accounted for half of the total; (4) in the past year, about 70% of athletes participated in sports competitions only 1–3 times. The reason for this relatively high number is attributed to the official lifting of COVID-19 prevention and control measures in China on January 8, 2023, allowing sports competitions to resume. Prior to that date, sports competitions could not be held normally, and athletes were unable to participate in them.

The questionnaire comprised five sections. The first section requested respondents to provide their demographic information, including age, gender, the sports they were engaged in, the number of competitions participated in the past year, and the duration of training cessation due to sports injuries or illnesses in the past year.

The second section utilized five items from the scale developed by Feldman et al. (46) to collect relevant data on respondents’ mindfulness levels. Sample items included “I try to pay attention to my thoughts without judging them.” The third section employed four items from the scale developed by Lundqvist and Hassmén (47) to gather data on respondents’ cognitive state anxiety. Sample items included “I worry about my performance when I am engaged in sports.” The fourth section used four items from the scale developed by Vallerand et al. (37) to collect data on respondents’ obsessive passion. Sample items included “There is a strong impulse to engage in this sport that I cannot resist.” The fifth section utilized five items from the scale developed by Steffen and Brehm (48) to compile data on respondents’ obligatory exercise. Sample items included “I feel guilty when I do not exercise.” All four scales were measured using a Likert five-point scale, with response options ranging from 1 (Strongly Disagree) to 5 (Strongly Agree).

The researchers adapted some items in the scales to better suit the habits and context of Chinese individuals, as well as to ensure the scale items were appropriate for our research context. Please refer to Table 2 for the content of the scales. A pilot test was conducted in advance to ensure its reliability (49). A total of 96 valid questionnaires were anticipated, and the results indicated that Cronbach’s alpha coefficients were all above 0.8, confirming the researchers’ modifications to the scales based on the context were justified.

In this study, AMOS v.23 was employed to construct a structural equation model (SEM) to examine how athletes, through mindfulness meditation, reduce cognitive state anxiety and obsessive passion, thereby improving their tendency toward obligatory exercise. The maximum likelihood (ML) estimation method was utilized to estimate the model parameters. A two-step modeling approach was employed to assess both the measurement and structural models. First, a comprehensive evaluation of the model’s reliability and validity was conducted. Subsequently, fit indices and path coefficients of the hypothesis model were measured, and the presence of mediating effects was examined.

To mitigate potential common method variance (CMV) arising from self-reported behaviors, researchers followed the recommended approach outlined by Mossholder et al. (50). As part of this approach, a comparative analysis was conducted between model one and model two, focusing on variations in degrees of freedom and chi-square values. The results indicated that the chi-square value for model one was 3078.434, encompassing 135 degrees of freedom and yielding a p-value below 0.001. Similarly, for model two, the chi-square value was 252.106, accompanied by 129 degrees of freedom and a p-value below 0.001. These findings affirm the congruence of fit between model one and model two, suggesting that CMV is not a concern within the context of this study.

The reliability and validity assessment of latent variables incorporated confirmatory factor analysis (CFA) using AMOS v.23. Table 2 displays the factor loadings, Cronbach’s α coefficients, composite reliability (CR), and average variance extracted (AVE) for each item on the scale. As shown in Table 2, all variables demonstrated Cronbach’s α values exceeding 0.8, affirming robust internal consistency within the model structure as guided by Fornell and Larcker (51). Additionally, the AVE for each variable exceeded 0.6 (as noted in Table 2), surpassing the minimal acceptable threshold of 0.5. Furthermore, the CR of each latent variable surpassed 0.8, underscoring the model’s robust convergent validity. The resilience of convergent validity across proposed models was well-established. Factor loadings from principal component factor analysis ranged from 0.751 to 0.908, reinforcing the measurement model’s robust construct validity. Table 3 shows that the square root of the AVE is on the diagonals, while the off diagonals are Pearson’s correlations of constructs. The square root of the AVE on the diagonals exceeds the inter-construct correlations, confirming discriminant validity.

Following a comprehensive evaluation of the measurement model’s reliability and validity, this study advanced to scrutinize the structural model using AMOS v.23, aimed at substantiating the formulated hypotheses. Using 5,000 bootstrap samples, the outcomes of CFA consistently aligned with established criteria (χ²/df = 2.755, GFI = 0.907, NFI = 0.940, TLI = 0.954, CFI = 0.961, RMSEA = 0.066), unequivocally indicating an optimal fit between the model and empirical data. Moreover, the interconnectedness of variables was corroborated by the Pearson correlation results outlined in Table 3. The standardized coefficients for variables within the structural equation model are visually depicted in Figure 2.

As illustrated in Figure 2, mindfulness exhibited a direct and negative correlation with cognitive state anxiety (β = −0.378, p < 0.001) and a negative correlation with obsessive passion (β = −0.150, p < 0.01), effectively substantiating H1 and H2, respectively. Additionally, cognitive state anxiety showed a direct and positive correlation with obsessive passion (β = 0.571, p < 0.001), robustly supporting H3. Similarly, cognitive state anxiety had a direct and positive association with obligatory exercise (β = 0.504, p < 0.001), affirming H4. Notably, obsessive passion was identified as having a direct and positive correlation with obligatory exercise (β = 0.290, p < 0.001), substantiating H5.

The evaluation of mediating effects, as depicted in Table 4, was conducted using bootstrap estimation with 5,000 resamples and 95% bias-corrected confidence intervals. The findings distinctively highlighted the indirect effect of mindfulness on obligatory exercise, mediated through cognitive state anxiety and obsessive passion. This mediation yielded a robust estimate of −0.297 (SE = 0.041, CI = [−0.377, −0.213], p < 0.001), thus providing comprehensive support for H6.

This study makes several theoretical contributions to the analysis of athletes’ obligatory exercise. Firstly, unlike existing research on athletes’ obligatory exercise that focuses on athletes’ body dissatisfaction (11) and socio-cultural factors (10), this study is pioneering in combining cognitive and behavioral aspects to explore the relationship between mindfulness and obligatory exercise among young athletes. The results reveal a significant negative correlation between mindfulness and cognitive state anxiety (see Figure 2), supporting the findings of Bühlmayer et al. (19). In our model, the impact of mindfulness on obligatory passion was the most pronounced, followed by cognitive state anxiety. These factors, cognitive state anxiety and obligatory passion, mediate the relationship between mindfulness and obligatory exercise, explaining 52% of the variance in obligatory exercise as depicted in Figure 2. This indicates that the prevalence of obligatory exercise issues among young athletes is intimately linked to how they manage anxiety and passion. Mindfulness training can potentially diminish the impacts of obligatory passion and cognitive state anxiety on their cognitive and judgment skills, thereby effectively reducing the occurrence of obligatory exercise issues. This insight provides a solid pathway and theoretical foundation for further research into the relationship between mindfulness training and obligatory exercise issues among young athletes.

Considering the positive impact of mindfulness on reducing athletes’ obsessive passion and cognitive state anxiety, as well as its indirect effect on alleviating obligatory exercise problems, the following recommendations are proposed: Firstly, at the national level, the National Sports Administration and its subordinate government departments should prioritize mental health in young athletes. Given their less mature mindset, athletes are more susceptible to cognitive distortions (4), making the incorporation of mindfulness training into daily plans crucial. Establishing a reasonable evaluation system and regularly inspecting the psychological conditions of athletes nationwide will provide a strong support system for their mental health. It is essential to collect and analyze data on the cognitive impact of mindfulness training methods on athletes, allowing for timely adjustments to ensure their effectiveness and applicability.

Secondly, at the level of athletes’ surroundings, including their team psychologists, coaches and training partners, team sports psychologists play a crucial role. They must promptly understand the latest research on the impact of mindfulness on athletes, encourage active participation in mindfulness training, and meticulously observe and record the athletes’ psychological states to maintain mental health. Coaches should identify athletes suffering from obligatory exercise issues and adjust training plans accordingly, in consultation with sports psychologists, while eliminating the unhealthy training atmosphere of “the more, the better.”

Finally, at the individual level, athletes should actively participate in mindfulness training and maintain good communication with coaches and teammates. When early symptoms such as anxiety and stress arise, athletes should provide honest feedback on their psychological issues (2) to prevent obligatory exercise problems. Early detection and treatment are crucial for realizing athletic potential.

For a long time, the focus has been on athletes’ performances, overlooking the negative impacts of long-term high-intensity training. Despite studies confirming that mindfulness training alleviates negative emotions like anxiety, tension, and fatigue (29) and positively impacts performance (52), its importance is often unrecognized by management organizations, coaches, and athletes themselves. Furthermore, influenced by personal self-esteem, many young athletes are reluctant to seek help for psychological issues such as training pressure and panic (53), leading to the loss of control over their training behavior. This study highlights the positive influence of cognitive improvements and encourages greater participation in mindfulness training. We recommend that young athletes in China try the Mindfulness-Acceptance-Insight-Commitment (MAIC) program, suited to the Chinese social context (54), to gain a comprehensive understanding of their training activities and reduce the occurrence of negative situations such as obligatory exercise.

Firstly, the participants of our study were mainly drawn from universities and sports teams within the central and southern regions of China. Recognizing the need for a more diverse and representative sample, future studies should aim to extend the geographical scope and include a wider range of sports disciplines. This expansion would not only enhance the representativeness of the sample but also improve the generalizability of the research findings across different contexts. Secondly, given the cross-sectional nature of our study, which lacks a temporal dimension, we acknowledge the limitation this presents in understanding the dynamic nature of mindfulness training effects. Future research could benefit from adopting a longitudinal approach, allowing for a detailed examination of how the relationships between variables evolve over time and the sustained impacts of mindfulness training on athletes’ behavior and psychological well-being. Thirdly, our research concentrated on young competitive athletes, leaving a gap in our understanding of mindfulness training’s relevance for non-competitive or recreational athletes. Future inquiries should explore the applicability of our findings to these groups, taking care to address potential biases in self-reporting by participants. Fourthly, our study was limited by its exclusive reliance on quantitative research methods. Incorporating qualitative insights on top of quantitative data could significantly enrich our understanding of how athletes personally perceive the impact of mindfulness on their training and psychological states. Therefore, future research should aim to integrate qualitative methodologies, such as interviews or focus groups, to capture the nuanced perspectives of athletes regarding mindfulness practices. This mixed-methods approach would offer a more holistic view of the mindfulness training effects, providing deeper insights into the athletes’ subjective experiences and the complex interplay between mindfulness, training habits, and mental well-being.