Introduction

Therefore, this study suggests that regulatory emotional self-efficacy may serve as an indirect factor between mindfulness and learning burnout. This study examines how mindfulness interacts with regulatory emotional self-efficacy to mitigate learning burnout, addressing a critical gap in understanding the process underlying mindfulness’s benefits. By integrating these constructs, we aim to inform more precise interventions in academic settings.

Literature review

The Relationship between Mindfulness and Learning Burnout

The Relationship between Mindfulness and Regulatory Emotional Self-efficacy

This study aims to address these gaps by specifically examining how mindfulness relates to RESE. We hypothesize that mindfulness is positively associated with RESE (Hypothesis 2), as mindful awareness may strengthen individuals' confidence in regulating emotional responses. By evaluating potential contradictions in prior work, this research seeks to clarify the mechanisms linking mindfulness to regulatory emotional self-efficacy.

The Relationship between Regulatory Emotional Self-efficacy and Learning Burnout

The Mediating Role of Regulatory Emotional Self-efficacy

H2H3MindfulnessLearning BurnoutH1Regulatory Emotional Self-Efficacy ScaleH4Existing research has established associations between mindfulness and learning burnout, yet causal mechanisms remain unclear. While mindfulness-based interventions have been widely studied, few focus specifically on university students. This study examines how regulatory emotional self-efficacy mediates the mindfulness-burnout relationship in this population, as showed in Figure 1, addressing key gaps in understanding these interactions.

Figure 1. The Hypothetical Model

Methodology

Research Design

The present study sought to investigate the association among mindfulness, regulatory emotional self-efficacy, and learning burnout. A quantitative methodology was employed to gather and analyze empirical data. Specifically, a survey research design was utilized as it is particularly well-suited for collecting and quantifying relevant data.

Participants and Procedures

This research used cluster sampling. Since the questionnaire requires a certain amount of time to complete all items, this sampling method is not only easy to implement, but also can better ensure that students complete it in full within the specified time. At the same time, in order to make the study more reasonable and reduce sampling error, we conducted cluster sampling of subjects from different types of universities and different grades. Inclusion and Exclusion Criteria of participants was considered as follows: (a) Students are between 18-23 years old from different universities of Fujian; (b) No major physiological or psychological illnesses.

A total of 461 participants were selected during the fall semester of the 2024/2025 academic year to take part in this study.The participants include 240 males and 221females; the respondents were undergraduate students from four grades (118 from freshman, 135 from sophomore, 123 from junior year and 85 from senior year). Table 1 provides an overview of the demographic characteristics of the participants.

This research received approval from the local ethics committee, and all procedures were carried out in accordance with the approved guidelines. Informed consent was obtained from each participant. They completed a series of questionnaires during class breaks: the Five Facet Mindfulness Inventory (FFMI), the University Student Learning Burnout Scale (USLBS), and the Regulatory Emotional Self-Efficacy Scale (RESS). The questionnaires were administered via Sojump and collected within ten minutes.

Table 1. Descriptive Analysis of Respondents' Background

Measures

Five- Factor Mindfulness Question

University Student Learning Burnout Scale

Regulatory Emotional Self-Efficacy Scale

Data Analysis

Results

Levels of Mindfulness, Regulatory Emotional Self-efficacy and Learning Burnout

Table 2. The Mean and Standard Deviation of Mindfulness, Regulatory Emotional

Self-efficacy and Learning Burnout

Note: M=Mean; SD= Std. Dev.

Multicollinearity

Multicollinearity or collinearity occurs when two independent variables predict each other rather than predicting the dependent variable. In order to confirm that the variables are free from multicollinearity, the Variation Inflation Factor (VIF) must be less than 5. In this study, the highest VIF is 1.797 for safety reward as shown in Table 3. All the VIFs are less than 3.3, supporting the assumption that the data are free from the multicollinearity issues and there is no common method bias.

Table 3.Multicollinearity

Note: VIF=Variation Inflation Factor

Measurement Model Evaluation

Table 4. Reliability Statistics

Table 4. Continued

Notes: x1 to x39 = each item of Mindfulness; y1to y20 = each item of Learning Burnout; l1 to l17= each item of Regulatory Emotion Self-efficacy; MIN1 to MIN5 = each construct of Mindfulness; LB1 to LB5 = each construct of Learning Burnout; RES1 and RES2 = each construct of Regulatory Emotion Self-efficacy

Table 5.AVE Value of Constructs

Note: AVE=Average Variance Extracted

Table 6. Discriminant Validity: HMTM

Notes: Square root of the AVE on the diagonal (bold) * Com=Community; RES= Regulatory Emotion Self-efficacy; MIN = Mindfulness; LB = Learning Burnout

Structural Model

Figure 2. The PLS-SEM Model

Referring to Figure 2, mindfulness, and regulatory emotional self-efficacy are able to explain 53.7% of the variance in learning burnout. Meanwhile, mindfulness explains 40.7% of the variance in regulatory emotion self-efficacy. In the structural model, each path linking two latent variables represents a hypothesis. As shown in Table 7, SEM analysis result shows that learning burnout is influenced directly by mindfulness (β=-0.679, t=28.657). As a result, mindfulness has an extremely significant negative correlation with learning burnout (p<.001). Hence, hypothesis H1 is accepted. SEM analysis result shows that mindfulness influenced directly regulatory emotional self-efficacy (β=0.638, t=24.306). Consequently, mindfulness exhibits an extremely significant positive correlation with regulatory emotional self-efficacy (p < .001). Therefore, hypothesis H2 is accepted. SEM analysis result analysis shows that learning burnout is influenced directly by regulatory emotional self-efficacy (β=-0.357, t=8.592). As a result, learning burnout exhibits an extremely significant negative correlation with regulatory emotional self-efficacy (p < .001). Therefore, Hypothesis H3 is accepted.

Table 7.Results of Hypotheses Testing (Bootstrap Analysis with 5,000 Samples)

Notes: ***p < 0.001 (two-tailed); all confidence intervals are derived from 5,000 bootstrap samples using bias-corrected percentile method; Standardized beta coefficients (β) reported; Standard errors and t-values from bootstrap analysis; RES= Regulatory Emotion Self-efficacy; MIN = Mindfulness; LB = Learning Burnout

Besides, regulatory emotional self-efficacy is found to negatively influence learning burnout (β = -0.357, t = 8.592) and is positively influenced by mindfulness (β = 0.638, t = 24.306). As shown in Figure 3, by incorporating the mediating variable, the coefficient between mindfulness and learning burnout is reduced from -0.679 to -0.451. According to Baron and Kenny’s standards, this study determines that regulatory emotional self-efficacy partially mediates the relationship between mindfulness and learning burnout. Additionally, the introduction of regulatory emotional self-efficacy as a mediator increases the R² value from 0.461 to 0.537. Therefore, Hypothesis H4 is supported.

MindfulnessLearning Burnout-0.679R2=0.4610.638-0.357-0.451R2=0.407MindfulnessLearning BurnoutEmotionalSelf-efficacyR2=0.537

Figure 3. The Results of Post-hoc Analysis of Regulatory Emotional Self-efficacy

Table 8.Results of f²and Q²

Notes: Bootstrapping: 5,000 samples, bias-corrected CI ; Effect Size Thresholds: f² ≥ 0.02 (Small), ≥ 0.15 (Medium), ≥ 0.35 (Large); Q² Interpretation: Values > 0 indicate predictive relevance (Stone-Geisser test); RES= Regulatory Emotion Self-efficacy; MIN = Mindfulness; LB = Learning Burnout.

The Importance-Performance Matrix Analysis (IPMA) Results

To build upon the results of the structural model, an Importance-Performance Matrix Analysis (IPMA) was conducted for the path model focusing on learning burnout. Each indicator's performance was evaluated on a scale from 0 to 100. According to Hair et al. (2017), indicators with high importance and low performance identify key areas requiring managerial attention. In this study, both the importance and performance of indicators linked to the endogenous variable (learning burnout) were analyzed.

Table 9. The Results of IPMA

Notes: MIN1 = Observing; MIN2 = Describing; MIN3 = Acting with Awareness; MIN4 = Non-judging; MIN5 = Non-reacting; RES1 = Expressing Positive Emotions; RES2 = Managing Negative Emotion.

As shown in Table 9, managing negative emotions has the strongest negative impact on burnout (Importance = -0.210), suggesting interventions targeting this resilience factor (e.g., stress management training) could most effectively reduce burnout. Acting with awareness shows high importance (-0.176) and the highest performance (61.078), indicating it’s already a relative strength—leveraging this (e.g., through mindfulness programs) may amplify benefits. Non-reacting (Performance = 45.139) has low current effectiveness despite moderate importance (-0.167). This signals a priority area for improvement (e.g., revising specific mindfulness techniques tied to non-reacting).Expressing positive emotions combines high importance (-0.183) and performance (61.159), suggesting maintaining current strategies (e.g., peer support) is advisable.

Discussion

Conclusion

This study constructed and verified the relationship model between mindfulness, regulatory emotional self-efficacy and learning burnout, and expanded the application perspective of mindfulness theory in the field of educational psychology. The study found that mindfulness can not only directly reduce learning burnout, but also play an indirect role by enhancing regulatory emotional self-efficacy, revealing the core position of emotion regulation in the mindfulness intervention mechanism.

Besides, among the specific dimensions of mindfulness, observation, description, non-judgment, non-reaction and conscious action all have a significant impact on learning burnout, especially the "non-reactivity" dimension is highly correlated with learning burnout, but students currently perform relatively weakly in this dimension. Based on the results of IPMA analysis, this study recommends that "non-reactivity" be taken as a priority development goal for future mindfulness intervention to improve the pertinence and effectiveness of the intervention effect. In addition, targeted mindfulness intervention measures to manage negative learning emotions are of great significance for alleviating learning burnout.

Therefore, this study theoretically fills the research gap in the relationship between mindfulness and learning burnout as a mediating variable, and provides a basis for schools to formulate personalized psychological intervention strategies in practice. Policymakers and educators should attach importance to the role of mindfulness training in student emotion management and learning support, especially by incorporating relevant training into curriculum design and mental health services in a targeted manner to effectively alleviate learning burnout and enhance students' learning motivation and psychological resilience.

Recommendations

Based on existing research results, it is recommended that when selecting mindfulness training techniques (e.g. Academic Body Scan,Affect Labeling,Loving-Kindness and so on) for coping with learning burnout, priority should be given to interventions focused on managing negative emotions. Besides, the expression of positive emotions has the greatest direct impact on burnout in the study (β=-0.210), which can be used as a core module of mindfulness intervention. It is recommended that the ability to express positive emotions and manage negative emotions, as well as mindfulness elements such as observation and conscious action, be incorporated into university mental health courses.

In addition, we also recommend the use of mixed research methods, through comparative studies of mindfulness intervention groups and non-intervention groups or multi-point tracking studies, to effectively elucidate the potential mechanisms by which mindfulness affects learning burnout. In addition, further exploration of the impact of mindfulness on research groups with greater scientific research pressure should also be considered, such as the intervention of mindfulness on learning burnout among graduate students or scientific researchers.

Limitations

This study has certain inherent limitations. First, the majority of participants in our sample were undergraduate university students, which restricts the generalizability of our findings, as this group may not be representative of a broader population. Furthermore, the relatively recent introduction of mindfulness within the Chinese cultural context limits the applicability of our findings to other educational settings and demographics. Additionally, the use of specific scales to collect data may constrain generalizability, as different instruments could yield varying results. These limitations underscore the need for caution when extrapolating our findings to populations outside the scope of this study and highlight potential avenues for future research to investigate these issues in more diverse and representative.

Ethics Statements

The studies involving human participants were reviewed and approved by Fujian University of Technology. The participants provided their written informed consent to participate in this study

Acknowledgements

The authors are very grateful to all those involved who contributed/supported this research.

Conflict of Interest

The authors declare no conflicts of interest.

Funds

No resources were available to support this research

Generative AI Statement

The authors acknowledge that no AI tool was used for the purpose of this paper. We, as the authors, take full responsibility for the content of our published work.

Authorship Contribution Statement

Cai: Conceptualization, design, analysis, writing. Mydin: reviewing, supervision. Amran: critical revision of manuscript, supervision.