This study adopted a simple random sampling method and selected a total of 3,600 students from middle schools in the Anshun, Liupanshui, Bijie, and Qiannan areas of Guizhou Province from May to June 2019. This study was cross-sectional design. According to the Chinese education system, secondary school includes the seventh grade, the eighth grade, the ninth grade, the tenth grade, the eleventh grade, and the twelfth grade. The data of students who did not answer all items or chose the same answer for all items were excluded. Of these, 3,566 responded with valid answers to the questionnaires—a 99% return rate. There were 1,539 male students (43.2%) and 2,027 female students (56.8%); 436 students (12.2%) in the seventh grade, 647 students (18.1%) in the eighth grade, 582 students (16.3%) in the ninth grade, 562 students (15.8%) in the tenth grade, 1,108 students (31.1%) in the eleventh grade, and 231 (6.5%) in the twelfth grade (see Table 1). Informed consent was obtained from all respondents. The studies involving human participants were reviewed and approved by School of Educational Sciences, Anshun University, China.

This study was conducted using a group questionnaire with standardized administration procedures. SPSS 24.0 and the macro program Process developed by Hayes were used for data analysis and processing. The moderating effect of gender and grade level on the stress and school bullying behavior was explored using the Bootstrap method proposed by Hayes (2013), with a sample size of 5,000 and Bootstrap CI using bias correction at 95% confidence level. Model testing criteria were used for confidence intervals not containing 0 and p < 0.05 and the model established met the fit criteria (Hayes, 2013). The direct and interactive effects of stress, gender, and grade on school bullying were then analyzed, as well as the specific moderating effects of gender and grade on said relationship.

Pearson correlation analysis found statistically significant correlations between gender, grade, stress, and school bullying (r = |0.19–0.28|, p < 0.01), and no statistically significant correlations between grade, gender, and stress (r_gender = −0.03, r_grade = −0.01, p > 0.05). According to the MacArthur method determination criteria, M was determined to be a moderating variable if it was not significantly correlated with X and Y or if the correlation was small (Kraemer et al., 2008). Based on the prior hypothesis and the related results, one concludes that gender and grade can be used as moderating variables. The descriptive statistics are shown in Table 2.

The variables were transformed prior to effect analysis, and the continuous variables “school bullying” and “stress” were transformed in a standardized manner. The categorical variables “gender” and “grade” were coded as male = 1, female = 0, middle school = 1, and high school = 0. The analysis showed that the confidence intervals of the coefficients of the established model did not contain zero, all of them were statistically significant, as were the interaction effects, and the model met the fit criteria.

This study explored the relationship between stress and school bullying behaviors among secondary school students as well as its mechanisms of action. Results found that stress had a significant positive predictive effect on school bullying behavior, which is consistent with previous national and international studies (Estrada-Martiner et al., 2012; Herts et al., 2012; Hsieh et al., 2014; Wang and Zhang, 2015; Shao et al., 2018; Hu et al., 2021), and the higher the stress level of an individual, the higher the risk of developing school bullying behavior. The “frustration–aggression” theory suggests that when their needs are not met, individuals may feel frustration and use aggression to relieve and mask it. When students experience certain stressful events in their daily lives, they may feel frustrated in achieving their goals or even feel that their interests or safety are threatened, which can provoke a differential response, thus gradually developing reactive aggression to defend and protect themselves. Secondary school students in the adolescent stage have not yet completed/integrated the appropriate cognitive-emotional mechanisms to properly handle various stressful situations. They are characterized by immature mental development, mood swings, and lack of self-control, therefore, reactive aggression caused by frustration is more likely to occur. Consequently, to mitigate the effects of stress on school bullying behaviors of secondary school students, families and schools should actively focus on those facing stressful events at school and help them channel their emotions in a timely manner as well as to properly navigate through the crisis.

This study found that the effect of stress on school bullying behavior of secondary school students was moderated by gender, with male students being more likely to develop school bullying behavior after being affected by stress. Buss proposed the theory of “gender differences in aggression” (Buss et al., 1998), which suggests that men and women differ in the targets, behaviors, and causes of aggression. Anthropological, physiological, and evolutionary studies have revealed that genetic factors explain 50% of the variance in aggressive behavior (Miles and Carey, 1997), with gender playing a moderating role between genes and aggression, and that genetic polymorphisms are significantly associated with the occurrence of aggressive behavior mainly in male populations (Lachman et al., 1998; Estrada-Martiner et al., 2012). Moreover, there are large gender-based differences in stress responses (Atkinson and Waddell, 1997), and a study by Bernhard et al. (2021) revealed that alterations in sex hormones (increased androgens in males and decreased estrogens in females) are strongly associated with conduct disorder (Bernhard et al., 2021), with the male group being more affected, in multiple ways, than the female group. In addition, the socialization theory of gender roles mentions that men and women show gender differences in many aspects due to the different roles and responsibilities they hold in the socialization process (Beall and Sternberg, 1995). The gender role stereotype that males should be assertive, independent, competitive, and aggressive, while females should be affectionate, caring, sensitive to the demands of others, and cooperative, lead to different gender characteristics and behaviors for men and women, making the former more likely to be involved in school bullying than the latter. Furthermore, both genders differ in the resolution strategies they use when faced with a problem that needs to be solved (Chen, 2011), with girls being stronger in “seeking social support” and problem-oriented coping strategies (Frydenberg and Lewis, 2000; Eschenbeck et al., 2007), while boys tend to use the strategy of “distracting themselves from unpleasant emotions” by distancing themselves more from the problem and becoming less emotionally invested in it. This increases their susceptibility to externalizing behaviors. As girls mature earlier and engage in more socialization than boys, they use more problem-solving strategies when faced with stress, generally employing non-aggressive approaches to conform their behavior to socially accepted role requirements; thus, females are less likely to use aggression as a form of problem solving than males. Consequently, schools should establish precise school bullying prevention and control mechanisms to give adequate attention to male high school students in stressful situations.

This study found that stress has an impact on school bullying behaviors among middle school students while being moderated by grade level, with middle school students more likely to be affected by stress and engage in school bullying than high school students. Previous research has suggested that aggression and victimization are highly stable from late childhood to adolescence (Scholte et al., 2007); however, research findings of Ryoo et al. (2015) showed that bullying and victimization are dynamic phenomena in groups that are frequently involved in bullying, suggesting that interventions should address the heterogeneity of the experiences of those involved in school bullying during adolescence. The present study further confirms this idea. The development of adolescents’ coping styles depends on the cognitive, biological, emotional, and social development state (Backhaus et al., 2010), and research has confirmed that coping strategies diverge and expand with age, showing some flexibility (Zimmer-Gembeck and Skinner, 2011), especially in the cognitive domain (Beck et al., 2016). According to the Social Information Processing Model of Children’s Social Adaptation (SIP model) proposed by Crick and Dodge (1994) children have a specific social cognitive process that influences or determines the final behavioral response before they respond to a particular social situation in the face of stress. High school students with higher cognitive levels are more inclined to make more comprehensive and objective interpretations at all stages of the information processing process, thus enabling appropriate behavioral responses. Zhang et al. (2011) further confirmed that dialectical thinking can reduce the tendency for aggressive behavior. Moreover, as secondary school students advance in their cognitive level, they also significantly develop their social emotions, increasing their ability to reflect on their emotions and developing an increasingly sophisticated emotion regulation strategy that manifests as “positive self-affirmation” (Zimmer-Gembeck and Skinner, 2011). As high school students get older and socialize more, they are better able to discipline themselves with social norms and control their own behavior, they begin to seek the approval and support of others, they naturally behave in a way that is consistent with social morality, and social moral norms are gradually internalized. Chen’s study concluded that children become more socialized and use more problem-solving strategies as they get older, more so after the age of 12 (Eschenbeck et al., 2007). Therefore, high school students at an older age will be less likely to engage in school bullying than middle school students, and more attention should be paid to the prevention and control of school bullying on middle school campuses and the establishment of a comprehensive prevention and control system.

This study has some limitations. First, the sample size was small and limited to secondary school students in Guizhou Province. Therefore, a larger and broader sample is needed to confirm our findings in future studies. Second, this study relied on secondary school students’ self-reports to conduct the assessment. Although the scale used in this study had high reliability, it could be further confirmed in the future using more rigorous and objective methods. Third, we did not adequately consider the effects of other potentially relevant variables (e.g., confounding factors such as coping strategies) when examining the relationship between stress and school bullying, which should be considered in future research.