This study included a total of 3,137 middle and high school students from the Chinese regions of Inner Mongolia, Henan, and Chongqing, using a simple random sampling method between October and December 2023. In the Chinese education system, middle school students typically encompass those in Grades 7 through 9 (ages approximately 12–15), while high school students include those in Grades 10 through 12 (ages approximately 16–18). These grade levels correspond to the junior and senior stages of secondary education. The sample thus spans the full spectrum of secondary education, providing a representative understanding of the phenomena across different developmental stages and diverse geographical areas within China.

A total of 3,137 questionnaires were distributed and collected. After excluding invalid, duplicate, and logically incorrect responses, 2,720 valid questionnaires were obtained, with an effective recovery rate of 86.7%. Logical errors primarily included extreme response patterns, such as selecting the same option for all questions or providing conflicting answers to opposing items; responses outside the target range, such as unreasonable values for age, weight, or height; and inconsistencies between reverse-scored and positively scored items, which made it impossible to verify the credibility of the responses. These errors were likely due to inattentive answering, misinterpretation of the questions, or intentional errors, and their exclusion significantly improved the accuracy and reliability of the data analysis.

Among the respondents, 1,336 were boys (49.1%) and 1,384 were girls (50.9%), with ages ranging from 11 to 18 years and an average age of 14.48 ± 1.47 years. This study was approved by the Ethics Committee of Capital University of Physical Education and Sports, and permission was obtained from the surveyed schools. The questionnaire survey was conducted after obtaining informed consent from all participating students and their parents.

Including gender, only-child status, grade, family residence, boarding status, height, and weight.

The Chinese version of the Academic Self-Efficacy Scale, originally developed by Pintrich and De Groot (1990) and revised by Liang (2002), was used to assess middle school students’ academic self-efficacy. The scale consists of two dimensions: self-efficacy in learning abilities and self-efficacy in learning behaviors, with 22 items in total (11 items per dimension). A 5-point Likert scale was used, ranging from “1” (not at all true) to “5” (completely true). In this study, the Cronbach’s alpha coefficient for the scale was 0.897, and the coefficients for the two dimensions were 0.927 and 0.712, respectively.

The Academic Pressure Scale from the Chinese Mental Health Scale for Middle School Students was used to assess academic stress (Wang and Xinhua, 2002). The scale includes six items corresponding to items 31, 33, 36, 38, 40, and 55 of the original questionnaires. A 5-point Likert scale was used, ranging from “1” (never) to “5” (very often). Higher scores indicate greater academic stress. In this study, the Cronbach’s alpha coefficient for the scale was 0.9.

The Simplified Coping Style Questionnaire, developed by Xie (1998), was used to assess coping styles. The scale includes two dimensions, with a total of 20 items: 12 items for positive coping and eight items for negative coping. A 4-point scale was used, ranging from “0” (never used) to “3” (frequently used). The average scores for each dimension were calculated, with higher scores indicating a tendency to use that coping style. In this study, the Cronbach’s alpha coefficient for the scale was 0.827.

The survey was conducted using an online platform, with a QR code and access link generated for the questionnaire. The research team conducted trial fills and verification to adjust and improve the questionnaire content. The survey was distributed with the help of school administrative staff, who were informed of the study’s purpose. After data collection, the research team conducted a second screening of the returned questionnaires, removing incomplete or invalid responses, and used double entry to record the valid data into statistical analysis software.

SPSS 26.0 and Amos 26.0 statistical software were used for data management and analysis. Prior to the formal data analysis, common method bias was tested using SPSS 26.0 and Amos 26.0. Formal data analysis involved descriptive statistics and correlation analyses conducted in SPSS 26.0, with count data presented as frequencies and percentages. Amos 26.0 was then employed to test the hypothesized mediation model through latent variable structural equation modeling (SEM), with the bootstrap method set to 5,000 iterations to explore the relationships among academic self-efficacy, simplified coping styles, and academic stress. The significance level was set at α = 0.05.

Since the data for this study were collected through self-reports from participants, there is a potential risk of common method bias. To minimize this bias during the administration of the survey to middle school students, the researchers emphasized the anonymity and confidentiality of the questionnaire, explicitly stated that the data would be used exclusively for scientific research purposes, and incorporated reverse-coded items into the questionnaire as a control measure.

Harman’s single-factor test was employed to assess common method bias. The analysis revealed that, in the unrotated factor solution, eight factors had eigenvalues greater than 1, with the first factor accounting for only 27.31% of the total variance, which is below the critical threshold of 40%. Additionally, a single-factor confirmatory factor analysis (CFA) was conducted using all self-reported items to further examine common method bias (Podsakoff et al., 2003). The results indicated poor model fit: χ²/df = 99.94, CFI = 0.67, GFI = 0.74, AGFI = 0.62, NFI = 0.67, and RMSEA = 0.19. Therefore, this study did not exhibit serious common method bias issues (Zhou and Long, 2004).

Table 1 reports the scores on various scales among students from different educational stages and family residences. From the perspective of educational stages, there were significant differences between junior high school students and senior high school students in academic self-efficacy, academic stress, positive coping style, and negative coping style (p < 0.01). Specifically, senior high school students scored higher on academic stress and negative coping style than junior high school students, while they scored lower on academic self-efficacy and positive coping style. From the perspective of family residence, there were no significant differences in academic self-efficacy, academic stress, and negative coping style between urban and rural students, but urban students scored significantly higher on positive coping style than rural students (p < 0.01).

Academic stress among middle school students was significantly negatively correlated with academic self-efficacy and positive coping style (r values were −0.37 and −0.3, respectively, both p < 0.001), and positively correlated with negative coping style (r = 0.32, p < 0.001). Positive coping style was significantly positively correlated with academic self-efficacy (r = 0.49, p < 0.001), while negative coping style was significantly negatively correlated with academic self-efficacy (r = −0.05, p < 0.01), provide a basis for subsequent hypothesis testing (Table 2).

Using latent variable structural equation modeling, the relationships among variables were further examined. Demographic variables, primarily gender and age, were first controlled, and the total effect of academic self-efficacy on academic stress was found to be significant. Subsequently, mediating variables (positive coping and negative coping) were incorporated into the model. The model’s fit indices indicated an overall good fit, with all metrics showing highly satisfactory results except for χ²/df, which was slightly larger due to its sensitivity to sample size (Abd-El-Fattah, 2010; Kline, 2016; Kenny and McCoach, 2003). The specific indices were as follows: χ²/df = 21.72 (p < 0.001), CFI = 0.92, GFI = 0.92, AGFI = 0.88, NFI = 0.91, RMSEA = 0.08, SRMR = 0.05. Standardized path analysis results demonstrated that all paths in the model were statistically significant (see Figure 1).

The bootstrap procedure was employed to test the mediation effects, with a resampling size of 5,000 to estimate the 95% confidence intervals for the mediation effects. Results showed that academic self-efficacy significantly predicted positive coping, negative coping, and academic stress (β = 0.56, p < 0.001; β = −0.13, p < 0.001; β = −0.26, p < 0.001). Both positive and negative coping significantly predicted academic stress (β = −0.22, p < 0.001; β = 0.36, p < 0.001).

The mediation effects were generated through two pathways: academic self-efficacy → positive coping → academic stress and academic self-efficacy → negative coping → academic stress, accounting for 47.38 and 18% of the total effect, respectively. The 95% confidence intervals for both paths did not include zero, indicating statistical significance. These findings support the hypotheses of this study.

Furthermore, the standardized indirect effects were β = −0.12 for the positive coping pathway and β = 0.05 for the negative coping pathway. According to conventional benchmarks (Cohen, 2013; Preacher and Kelley, 2011), the former represents a medium effect and the latter a small effect. This suggests that while both coping styles are significant mediators, positive coping plays a relatively stronger role in linking academic self-efficacy to academic stress.

Figure 1 illustrates the structural mediation model tested in this study. As shown in the figure, academic self-efficacy not only exerts a direct effect on academic stress but also influences it through indirect pathways.

The study found significant differences in academic stress among students of different ages and regions. Descriptive statistics revealed that high school students experienced significantly higher academic stress compared to junior high school students (Liu, 2016). This aligns with previous findings and can be attributed to the increased complexity of academic tasks in high school (Liu and Lu, 2011), the pressure of university entrance exams, and the heightened competition for limited spots at prestigious universities. High school students face stricter evaluation standards, more demanding coursework, and an intense extracurricular training environment, all of which contribute to elevated academic stress.

In terms of regional differences, middle school students in urban areas were found to adopt more positive coping strategies than those in rural areas. This disparity may stem from the advantages of educational resources in urban regions, including high-quality teaching staff, diverse extracurricular activities, and higher parental education levels (Zhao and Yuchun, 2019). These factors provide urban students with greater psychological support and learning resources, enabling them to approach academic stress with a more positive attitude. Therefore, when addressing academic stress among middle school students, it is essential to consider urban–rural disparities, particularly the academic stress levels, coping strategies, and potential impacts on the mental health of students in rural areas.

The results of this study demonstrate that academic self-efficacy significantly predicts lower academic stress. Students with high academic self-efficacy are more likely to believe in their ability to effectively handle academic tasks, displaying stronger emotional regulation skills (Honicke and Broadbent, 2016). This confidence enables them to view challenges as opportunities for growth rather than threats, thereby reducing their perceived levels of academic stress (Bandura and Wessels, 1997; Luo et al., 2023). Furthermore, students with high academic self-efficacy tend to make better use of external resources, such as teacher guidance, peer support, and family care (An et al., 2023). These sources of social support not only alleviate stress but also enhance academic achievement and intrinsic motivation (Wang and Eccles, 2013). In contrast, students with low academic self-efficacy often lack the belief in their ability to succeed, which can lead to feelings of helplessness and anxiety when faced with challenges. Such students are more likely to attribute failures to their lack of ability, reinforcing a cycle of stress and negative emotions. For example, they may give up easily when encountering difficulties, showing a lack of initiative and persistence in problem-solving (Achmad et al., 2023). This passivity leads to reduced effort and resource utilization in academics, creating a downward spiral (Frese and Fay, 2001). Negative attributions further amplify the adverse psychological impact of academic stress, affecting mental health (Usher and Pajares, 2008). Therefore, enhancing students’ academic self-efficacy serves as an effective psychological intervention, helping mitigate the detrimental effects of academic stress on mental wellbeing.

From a psychological perspective, academic self-efficacy, defined as an individual’s belief in their abilities within an academic context directly influences how students cognitively appraise and respond to academic challenges. According to the theory of stress appraisal (Lazarus and Folkman, 1984), an individual’s evaluation of a stressful situation determines their choice of coping strategies, with academic self-efficacy playing a pivotal role in this process (Bandura and Wessels, 1997; Pajares, 2002). Students with high academic self-efficacy tend to perceive academic challenges as manageable tasks rather than threats. This positive cognitive orientation encourages them to adopt more effective coping strategies, such as problem-solving and seeking support, thereby mitigating the negative effects of academic stress. Additionally, these students are more likely to use stress as motivation, enhancing their self-regulation abilities over time.

Conversely, students with low academic self-efficacy often lack confidence in their ability to succeed and perceive academic tasks as insurmountable challenges. They are more likely to adopt negative coping strategies, such as avoiding problems or venting emotions, which not only fail to address academic challenges but can also lead to the accumulation of stressors, creating a vicious cycle (Gao, 2023). For instance, when students avoid exams or procrastinate on completing tasks, the resulting backlog of academic work can intensify feelings of guilt and anxiety, further exacerbating their academic stress (Jagiello et al., 2024; Krispenz et al., 2019). This maladaptive coping approach acts as an amplifier of academic pressure. In contrast, enhancing academic self-efficacy can effectively break this cycle by fostering a sense of control and promoting more constructive responses to stress.

Research further demonstrates that coping strategies partially mediate the relationship between academic self-efficacy and academic stress, providing empirical support for the effectiveness of coping strategies as an intervention tool. Specifically, students with high academic self-efficacy are more likely to adopt positive coping strategies, which buffer the effects of stress by reducing its subjective perception and enhancing emotional adaptability. This finding aligns with existing research supporting the stress-buffering hypothesis of coping strategies (Compas et al., 2001; Folkman and Moskowitz, 2000; Skinner and Zimmer-Gembeck, 2007). However, it is worth noting that the mediating effect of coping strategies accounts for only part of the total effect. This suggests that, beyond coping strategies, academic self-efficacy may indirectly influence academic stress through other mechanisms. One such mechanism could be social support (Rueger et al., 2008). Students with high academic self-efficacy often exhibit stronger social skills, enabling them to actively seek assistance from teachers, parents, and peers. This social support system significantly alleviates academic stress through emotional encouragement and academic guidance (Malecki and Demaray, 2006; Wentzel, 1998).

The findings have important practical implications for alleviating academic stress among secondary school students. First, schools should focus on enhancing students’ academic self-efficacy by setting individualized academic goals, providing constructive feedback, and organizing interdisciplinary learning projects. These efforts can strengthen students’ sense of control and achievement in their academic tasks. Particularly in high school, teachers should assist students in creating effective time management plans and guide them to break down academic goals into achievable steps, thereby reducing the psychological burden of large-scale objectives. Second, the mediating role of coping strategies underscores the importance of mental health education in schools. Programs should incorporate scenario-based training and psychological counseling to cultivate positive coping strategies such as problem-solving, emotion regulation, and seeking social support while minimizing maladaptive behaviors. Finally, families and communities should serve as critical support systems for relieving academic stress. Parents need to provide both academic guidance and emotional support, encouraging children to share their difficulties and offering reassurance. At the societal level, optimizing resource allocation can help provide students in rural areas with greater access to quality educational resources and mental health services, thereby strengthening their ability to manage academic stress. Through the collaborative efforts of schools, families, and communities, the adverse effects of academic stress on students’ mental health can be effectively mitigated, fostering an environment conducive to the holistic development of secondary school students.

This study explores the relationships among academic self-efficacy, coping strategies, and academic stress, emphasizing how academic self-efficacy shapes students’ approaches to managing stress. The findings demonstrate that students with high academic self-efficacy are more likely to view stress as a challenge and adopt positive coping strategies, such as problem-solving, which alleviate stress and enhance resilience. Conversely, low self-efficacy leads to negative coping, exacerbating stress.

Coping strategies were also found to mediate the relationship between academic self-efficacy and academic stress, indirectly reducing stress and improving emotional regulation and academic outcomes. Moreover, the role of social support as a potential pathway warrants further exploration, as students with higher self-efficacy are more inclined to seek help from teachers, parents, and peers.

In conclusion, fostering academic self-efficacy and promoting effective coping strategies are critical for reducing stress and improving academic performance, offering valuable insights for educational interventions.

Although this study provides valuable insights into the relationships among academic self-efficacy, coping strategies, and academic stress, several limitations should be addressed. First, the cross-sectional design reveals correlations but does not establish causal relationships. Future research should adopt longitudinal or experimental designs to explore how enhancing academic self-efficacy causally impacts academic stress.

Second, the findings on coping strategies’ mediating role are based on statistical analyses of survey data, which may be influenced by confounding variables and individual differences. Experimental studies, such as randomized controlled trials, are needed to confirm these effects and assess the impact of interventions targeting positive coping strategies.

Third, this study did not include other potentially important variables, such as family environment, parental expectations, or social support, which may also influence students’ academic stress. In the Chinese cultural context, Confucian academic pressure and filial piety may further shape how students perceive and respond to academic demands, while urban–rural disparities in educational resources and social support may lead to differences in self-efficacy, coping strategies, and stress experiences. Future research should therefore consider incorporating these contextual and cultural factors into the models to provide a more comprehensive understanding of the mechanisms involved.

Finally, the sample, primarily drawn from specific regions, limits the generalizability of the results. Differences in educational resources, cultural contexts, and social support across regions may affect academic stress. Future research should broaden the sample scope, employ multi-level analyses, and conduct longitudinal studies to uncover more generalizable patterns and effective interventions.