This study employed a cross-sectional design, collecting data from sports-disadvantaged college students through a questionnaire survey. Data collection was conducted in April 2023 across six universities in Southwest China.

A stratified random sampling method was used to ensure representation across different universities and academic disciplines. Each university was treated as a stratum, within which students were randomly selected to participate. This approach aimed to enhance the representativeness of the sample and reduce sampling bias.

A total of 600 questionnaires were distributed, and 588 valid responses were obtained, resulting in an effective response rate of 98%. The sample size was determined based on the principle of ensuring statistical power for structural equation modeling analysis, referencing the commonly recommended sample-to-parameter ratio of at least 10:1. Given the number of parameters to be estimated in this study’s model, a minimum of 300 respondents was deemed sufficient; hence, the final sample size of 588 met and exceeded this requirement, ensuring robust and reliable statistical analysis.

The research instruments included the Trait Mindfulness Scale, Exercise Motivation Scale, Exercise Adherence Scale, and Self-Efficacy Scale. All questionnaires were completed online, with detailed instructions provided beforehand to ensure clarity and standardization of responses.

To ensure the quality of the data and the appropriateness of respondents, this study established scientifically grounded inclusion and exclusion criteria based on national standards and psychometric considerations.

Participants were included if they met the definition of being “sports-disadvantaged,” operationalized as those scoring below 60 points on the National Physical Fitness Standards for Chinese Students (College Level). This threshold is commonly used in education and health research to identify individuals with low physical fitness levels (Erwin, 2008). The criteria include multiple indicators such as cardiovascular endurance, muscular strength, flexibility, and body composition, which provide a multidimensional and standardized framework for evaluating physical disadvantage. Therefore, using this cutoff score offers a valid and objective method for identifying the target population in a manner consistent with national health benchmarks.

(b–e) Exclusion criteria:

To improve data reliability, the following responses were excluded:

Regarding instrument validity, all scales employed in this study (Trait Mindfulness Scale, Exercise Motivation Scale, Exercise Adherence Scale, and General Self-Efficacy Scale) have been widely validated in Chinese university student populations (Luszczynska et al., 2005; Zhang et al., 2024; Tao et al., 2024). Each scale’s internal consistency was reassessed in the present sample, and all Cronbach’s α coefficients exceeded 0.80, indicating strong reliability. Moreover, confirmatory factor analysis (CFA) was conducted to verify structural validity, and the model fit indices met accepted thresholds, supporting the continued use of these instruments in the current research context.

In terms of analytical depth, while some scales were used as composite unidimensional variables for initial analyses, the multidimensional structures of key scales—particularly exercise motivation and trait mindfulness were preserved in mediation and moderation models. We acknowledge the potential theoretical richness offered by these subdimensions; therefore, in subsequent structural equation modeling (SEM) analyses, we explored them as separate mediators and independent variables. This approach revealed more nuanced psychological pathways and yielded greater explanatory power.

To ensure ethical compliance, the study strictly followed the principles of the 1964 Helsinki Declaration and its later amendments. All participants received informed consent documents that clearly outlined the study’s aims, procedures, and any potential risks. Participation was entirely voluntary, and respondents retained the right to withdraw at any time without penalty or loss of benefit.

Data entry for the collected questionnaires was performed using Excel 2013, followed by relevant data analysis. First, descriptive statistics were conducted on the collected data. After confirming a normal distribution, Pearson correlation analysis was used to examine the relationships between variables. When all four variables showed significant pairwise correlations, a mediation effect analysis was conducted. The mediation effect analysis was conducted using the PROCESS macro (version 4.1) for SPSS, developed by Hayes (2013). To explore the relationship between independent variable (trait mindfulness) and dependent variable (self-efficacy), which may be affected by mediator variable (exercise motivation) and mediator variable (exercise persistence). In PROCESS, Model 6 was used for mediation analysis, with 5,000 bootstrap resamples, utilizing bias-corrected percentile bootstrap confidence intervals (CIs) to assess effect sizes. A mediation effect was considered significant if the 95% CI did not include zero. In the mediation effect analysis, participants’ age, gender, smoking status, and drinking habits were included as covariates to control for potential confounding.

Additionally, to examine the potential presence of common method bias, Harman’s single-factor test was conducted using exploratory factor analysis on all measured variables. The results indicated that nine factors had eigenvalues greater than 1, and the first unrotated factor accounted for only 30.91% of the total variance—below the critical threshold of 40%—suggesting that common method bias was not a serious concern in this study.

In this study, Harman’s single-factor test was applied to all variables using exploratory factor analysis. The results revealed nine factors with eigenvalues exceeding 1. Notably, the most significant factor accounted for only 30.91% of the variance, below the 40% threshold, indicating the absence of common method bias in this research.

This study included 588 participants for analysis. Male participants comprised 50.3% of the sample. The participants ranged in age from 18 to 23. Additionally, a relatively high proportion of participants reported smoking, accounting for 30–45% of the total sample, with a balanced distribution between those who consume alcohol and those who abstain. Detailed results are presented in Table 1.

The correlation analysis results (Table 2) indicate that trait mindfulness is significantly positively correlated with exercise motivation (r = 0.585, p < 0.01), exercise adherence (r = 0.545, p < 0.01), and self-efficacy (r = 0.581, p < 0.01). Additionally, exercise motivation is positively correlated with exercise adherence (r = 0.592, p < 0.01) and self-efficacy (r = 0.679, p < 0.01). Finally, exercise adherence shows a significant positive correlation with self-efficacy (r = 0.639, p < 0.01). All four outcome variables are positively correlated, which serves as an essential prerequisite for the subsequent analysis of the chain mediation effect.

The regression analysis results (Table 3) reveal that trait mindfulness positively predicts self-efficacy among sports-disadvantaged college students (β = 0.591, p < 0.001), confirming hypothesis H1. When mediating variables are introduced, the direct effect of trait mindfulness on participants’ self-efficacy diminishes (β = 0.195, p < 0.001), while trait mindfulness significantly predicts both exercise motivation (β = 0.583, p < 0.001) and exercise adherence (β = 0.330, p < 0.001). Additionally, exercise motivation significantly predicts exercise adherence (β = 0.406, p < 0.001) and self-efficacy (β = 0.385, p < 0.001), and exercise adherence significantly predicts self-efficacy (β = 0.303, p < 0.001). These findings suggest that exercise motivation and adherence act as mediators, forming a chain of mediation in the influence of trait mindfulness on self-efficacy in this group.

The mediation analysis results (Table 4; Figure 2) show a significant direct effect of trait mindfulness on self-efficacy within the mediation model (β = 0.195, 95% CI [0.125, 0.266]). Additionally, significant indirect effects were observed, with all three mediation pathways reaching significance:

Path 1: Trait mindfulness → Exercise motivation → Self-efficacy (β = 0.224, 95% CI [0.176, 0.277]);

Path 2: Trait mindfulness → Exercise adherence → Self-efficacy (β = 0.100, 95% CI [0.067, 0.136]);

Path 3: Trait mindfulness → Exercise motivation → Exercise adherence → Self-efficacy (β = 0.072, 95% CI [0.050, 0.099]).

The respective contributions of the three indirect effects are 37.90, 16.69, and 12.18%, thereby validating hypotheses H2–H4.

The analysis first confirms that trait mindfulness positively predicts self-efficacy, consistent with earlier research (Li et al., 2022). Within the framework of Social Cognitive Theory, individuals’ beliefs in their capabilities are shaped by prior experiences and cognitive processing. Students with higher trait mindfulness are more likely to regulate emotions effectively and display greater cognitive flexibility (Fino et al., 2021; Goilean et al., 2023). These qualities help reduce anxiety and self-doubt during physical activity, thereby fostering mastery experiences that reinforce self-efficacy.

Beyond this, mindfulness has been described as a metacognitive process involving awareness and regulation of one’s own thinking (Thomson and Van Hedger, 2025). Empirical studies have shown that mindful agency predicts metacognitive ability, which in turn enhances self-leadership among student populations (Chen and Zhang, 2022). This suggests that metacognition may act as a pathway between mindfulness and self-efficacy: through improved awareness and cognitive control, individuals reinterpret challenges as opportunities for growth, further strengthening competence and confidence. Importantly, mindfulness also supports attentional control and reduces rumination, which is particularly relevant for sports-disadvantaged students who often begin with lower motivational baselines.

Building on this direct relationship, the findings further indicate that exercise motivation serves as a partial mediator, adding nuance to the psychological processes involved. This result aligns with Self-Determination Theory, which emphasizes the importance of intrinsic motivation in sustaining health behavior. Mindfulness enhances individuals’ awareness during exercise and helps them align with their intrinsic goals (Jankauskiene and Baceviciene, 2024). For sports-disadvantaged students, who often lack external incentives or social support, this internalization of meaningful exercise goals is particularly critical (Cayleff, 1995).

Mindfulness promotes such internalization by fostering autonomy, a core element of SDT, and encouraging motivation that is self-endorsed rather than externally imposed (Ryan et al., 2021). In addition, it enhances perceived autonomy and competence—two essential needs within SDT—thereby facilitating the transition from controlled to autonomous forms of motivation (Aldbyani et al., 2024). Research in academic contexts has shown that mindfulness positively relates to self-efficacy, which in turn partly mediates its influence on performance outcomes (Aldbyani et al., 2024). Applied to exercise, this mechanism suggests that mindfulness enhances intrinsic motivation through strengthened self-efficacy, leading to deeper engagement and more sustained activity.

Extending this pathway, exercise adherence emerged as another significant mediator. The Health Action Process Approach provides a useful framework for understanding this finding, as it highlights the transition from intention to long-term maintenance through planning and self-regulation. In this process, mindfulness strengthens volitional control, enabling individuals to manage fluctuations in motivation, cope with barriers, and maintain consistent routines. According to self-efficacy theory (Smith et al., 2023), repeated successful experiences reinforce personal agency (Xiang et al., 2024), and mindfulness contributes to this process by fostering non-judgmental acceptance and resilience in the face of setbacks (Eşkisu et al., 2020).

Mindfulness further enhances distress tolerance and pain endurance, both of which are essential for sustaining regular physical activity. Evidence shows that injured athletes receiving mindfulness training report higher pain tolerance, which supports rehabilitation and continued adherence (Mohammed et al., 2018). Likewise, mindfulness practices during exercise reduce perceived exertion, making sustained participation more manageable (Solk et al., 2023). Together, these findings demonstrate how mindfulness not only supports psychological resilience but also reduces subjective barriers, thereby reinforcing the adherence component of the pathway to self-efficacy.

Taken together, the mediating roles of exercise motivation and adherence point to a chain mechanism through which trait mindfulness influences self-efficacy. Rather than acting in isolation, motivational quality (SDT) and behavioral persistence (HAPA) interact with cognitive belief formation (SCT) to form a comprehensive explanatory model of sustained health behavior. This integrated view demonstrates how mindfulness extends beyond momentary awareness, contributing to both the initiation and maintenance of physical activity, and ultimately strengthening self-efficacy.

Such an integrative framework represents a theoretical advancement by moving beyond single-pathway explanations. It also reflects recent calls for multi-theoretical health behavior models that offer greater explanatory power than approaches based on a single framework (Remskar et al., 2022). For example, interventions combining mindfulness and exercise have been shown to yield superior outcomes compared to either component alone (Tsang et al., 2008). By empirically validating a chain mediation model, this study enriches existing literature and provides practical implications for intervention design, especially for vulnerable student groups who face barriers to physical activity.