Cluster sampling was adopted to recruit 500 high school students with experience on the Internet to participate in this study. All the participants were recruited from three high schools in Wuhan China. This school age group was considered at elevated risk in social, emotional, academic difficulties, and cyberbullying (Chu et al., 2019; Fan et al., 2019; Shutzman and Gershy, 2023). Before the investigation, all participants were told that the research was being conducted anonymously and that their information would remain confidential. They were informed of the requirements of this survey by using standard instructions, emphasizing the authenticity, independence, and integrity of all answers. All the questionnaires were conducted in the form of paper-and-pencil in different classrooms taking a class as a unit in 45 min. All participants completed an informed consent form prior to completing the questionnaire, and this study was reviewed and approved by the Ethical Committee for Scientific Research of correspondence author. After getting rid of the invalid questionnaires (questionnaires with lots of blanks or repeated answers), data from 477 participants were retained, and the valid response rate was 95.40%. The effective sample included 240 (50.31%) boys and 237 (49.69%) girls. The mean age of participants was 15.84 years old (SD = 0.67), with an age range of 14–17 years old. Moreover, a total of 248 (51.99%) first-grade students and 229 (48.01%) second-grade students participated in this survey, the third-grade students did not participate in this survey because of the pressure of college entrance exams.

SPSS 25.0 and the PROCESS macro developed by Hayes (2013) were used for statistical analysis. Firstly, we calculated descriptive statistics and correlational analyses for the main variables. Secondly, the mediation effects (model 4) and moderated mediation effects (model 14) were investigated. The bootstrapping method was used to obtain 95% confidence intervals (95% CI) with 5,000 re-samples for the model. Finally, all the potential significant interaction effects were decomposed by simple slope tests.

Gender and age were included as control variables in our statistical analysis, as previous literature found that they were closely related to the observed variables (Pearson et al., 2011; Yang XJ et al., 2021; Gao et al., 2023; Xiong et al., 2023).

The data of the present study were all from self-report questionnaires. Therefore, Harman’s single-factor test was used to examine common method bias (Podsakoff et al., 2003), the results indicated that there were 14 factors with eigenvalues higher than 1. The first factor accounted for 25.70% of the variance so common-method variance was not an issue in this study.

Table 1 presented the means, standard deviations, and correlations for all core observed variables. As hypothesized, PCSA was positively correlated with rumination and cyberbullying perpetration, and negatively correlated with mindfulness. Rumination was positively correlated with cyberbullying perpetration, and negatively correlated with mindfulness. Cyberbullying perpetration was negatively correlated with mindfulness. Gender was positively correlated with PCSA and negatively correlated with mindfulness. Age was negatively correlated with PCSA. Whereas, gender and age showed no significant correlation with other core observed variables.

Using the model 4 in Hayes (2013) the macro program PROCESS, mediation effects were tested after controlling for gender and age. All data were standardized. As Table 2 illustrates, PCSA had a significant total effect on cyberbullying perpetration (β = 0.27, t = 6.10, p < 0.001). Moreover, PCSA was significantly and positively associated with rumination (β = 0.20, t = 4.48, p < 0.001), rumination was significantly and positively associated with cyberbullying perpetration (β = 0.40, t = 9.74, p < 0.001). The indirect effect of PCSA and cyberbullying perpetration mediated by rumination was significant (β = 0.08, SE = 0.02, 95% CI = [0.04, 0.13]). The mediation effect accounted for 29.63% of the total effect. Thus, rumination partially mediated the relationship between PCSA and cyberbullying perpetration. H1 and H2 were supported.

To test the moderating role of mindfulness, we ran the PROCESS macro (model 14). Moderation effects were tested after controlling for gender and age. All data were standardized. As Table 3 illustrates, the interaction of rumination and mindfulness had a significant predictive effect on cyberbullying perpetration (β = −0.07, t = −2.02, p < 0.05). Explain that mindfulness plays a moderating role in the prediction of rumination on cyberbullying perpetration (see Figure 2). H3 was supported.

To better describe the moderating effect of mindfulness, simple slope tests were conducted in this study (see Figure 3). For adolescents with lower mindfulness (M-1SD), rumination had a significant positive predictive effect on cyberbullying perpetration (b_simple = 0.41, SE = 0.05, p < 0.001, 95% CI [0.31, 0.51]). For adolescents with higher mindfulness (M + 1SD), the positive predictive effect of rumination on cyberbullying perpetration is still significant, but the predictive power is weaker (b_simple = 0.28, SE = 0.05, p < 0.001, 95% CI [0.17, 0.38]). It showed that as mindfulness increases, the predictive effect of rumination on cyberbullying perpetration is weakened. This suggested that mindfulness can alleviate the negative impact of rumination on cyberbullying perpetration (see Table 4).

The total effect model demonstrates that PCSA significantly and positively associated with cyberbullying perpetration among adolescents. Adolescents with higher PCSA are more inclined to bully others in cyberspace, thus validating H1. Support the general strain theory and previous research findings (Huang et al., 2017; Ding et al., 2018).

Adolescents who have higher PCSA have accumulated a large amount of aggression-related clues (such as social exclusion and frustration, etc.) through long-term interaction with negative life events, which increase their susceptibility to social adversities (Zhang J et al., 2017; Davis et al., 2018; Gabbiadini and Riva, 2018). They will automatically implement an immediate assessment of the social adversity when they encounter certain clues. This assessment process is spontaneous and unconscious and requires almost no cognitive effort. Adolescents can directly and automatically engage in bullying perpetration and aggression based on the original cognitive scripts when they think they face threatening events in their daily lives (Anderson and Bushman, 2002; Allen et al., 2018; Wang et al., 2019). Additionally, adolescents with heightened PCSA are prone to feel frustrated in reality and form latent negative self-schemas, resulting in a negative affect state for a long time, which also may enhance their motivation to implement aggressive behaviors in cyberspace to rebuild confidence (Ding et al., 2018). Particularly in cyberspace with its unique characteristics including anonymity and accessibility, the activation level of the adolescents’ behavioral inhibition system is lower than that in reality (Yen et al., 2012). Therefore, adolescents with higher PCSA are more likely to learn that bullying others in cyberspace is an acceptable way to alleviate their chronic negative effects and escape from negative life events in real social interaction (Zhang et al., 2019; Yuan et al., 2020). Previous studies found that PCSA can lead to a series of offline maladaptive problems including social avoidance, self-injury, and suicidal tendencies (Ding et al., 2018; Liu et al., 2022). Our findings extend existing research, indicating that adolescents with high levels of PCSA may not only become victims in reality but also be more likely to bully others in cyberspace.

This study’s results indicate that PCSA is associated with cyberbullying perpetration through the partial mediation of rumination, thereby validating H2. This finding supports the general aggressive model (Anderson and Bushman, 2002). PCSA may be a necessary condition in the formation of cyberbullying perpetration as a distal contributory, while rumination may be a sufficient condition for the formation of cyberbullying perpetration as a proximal contributory. Therefore, PCSA is not only directly associated with cyberbullying perpetration, but also indirectly associated with cyberbullying perpetration through rumination. This finding also supports the views of the social information processing theory that individuals’ perception of chronic social adversity may be a triggering factor for cyberbullying perpetration, and individuals will engage in emotional processing of the triggering factor, promoting the expression of cyberbullying perpetration (Crick and Dodge, 1994).

On the one hand, this study’s results are consistent with previous findings that a positive association exists between subjective perceptions of negative life events and rumination (Michl et al., 2013; Liu and Wang, 2017; Jin et al., 2020). Adolescents with high PCSA are often most sensitive to negative events, so they have to constantly and repeatedly perceive social adversity information in their daily lives (Ding et al., 2018). However, their cognitive and emotional management abilities are not yet mature, and they have not formed an appropriate cognitive style, which makes it difficult for them to form reasonable assessments of such events (Dickson et al., 2017). They have to repeatedly try to integrate the social adversity information into new life narratives to reduce the psychological pain caused by it (Platte et al., 2022). Therefore, high PCSA may encourage adolescents to develop non-adaptive passive coping styles like rumination into negative psychological habits related to emotional reactions, thus causing incorrect cognitive style tendencies, inducing excessive negative emotional experiences, and repeatedly thinking about possible strategies and consequences (Zhang J et al., 2017; Jin et al., 2020).

On the other hand, the results are consistent with previous studies that have shown that individuals with higher rumination are more likely to exhibit bullying perpetration and aggression (Pedersen et al., 2011; Ruddle et al., 2017; Jin et al., 2020). In online contexts, Yang J et al. (2021) also found a significant positive correlation between adolescents’ anger rumination and cyberbullying perpetration. Adolescents with higher rumination are more likely to immerse in the perception of social adversity, this will reduce their critical value of the potential threat evaluation system (Mathews and MacLeod, 2002), resulting in their abnormal sensitivity to negative events and difficulty to shift their attention from the negative emotions and experiences, ultimately leading to the deviation of cognition, the reduction of empathy, the formation of hostile prejudice, and the phenomenon of “violent desensitization” (Wang et al., 2018). That is to say, rumination reduces the self-regulation ability of adolescents when facing social adversity, and they are more likely to bully others in cyberspace (Gao et al., 2021). As such, adolescents’ subjective perceptions of negative life events affect their cyberbullying perpetration by exacerbating their rumination.

Further, this study discovered that the mediating effect of rumination was moderated by mindfulness. Specifically, the positive association between rumination and cyberbullying perpetration among adolescents was progressively weakened as mindfulness levels increased, confirming H3. These findings align with the buffering hypothesis within the “delivering carbon in the snow model,” which suggests that a protective factor may diminish the predictive effect of a risk factor on the outcome variable (Zhang et al., 2022). In this study, mindfulness served as a protective factor and rumination served as a risk factor for cyberbullying perpetration among adolescents. These two adverse factors interacted, which had an impact on the outcomes, with mindfulness effectively buffering the association between rumination and cyberbullying perpetration and aggression.

This may be because, in the limited resource of self-control theory (Baumeister et al., 2007), individuals with higher mindfulness will accumulate self-control resources and emphasize the subjective attitude of acceptance and non-judgment, as well as the process of “re-perception” towards external stimuli, including negative life events (Zou et al., 2021). They will facilitate their capacity to receptively observe negative life events as they arise with acceptance and equanimity through mindfulness, and make flexible rather than automated responses to those events objectively, which in turn buffer initial threat appraisals and increase secondary appraisals of coping resources (Creswell and Lindsay, 2014). Specifically, adolescents with higher mindfulness usually have higher self-efficacy, psychological resilience, and other positive psychological resources (Yue et al., 2020; Xiong et al., 2023), which can effectively alleviate the insufficient self-control resources caused by rumination and enhance individuals’ abilities to execute self-control, reduce repetitive passive thinking about adverse events, and promote individuals to seek more proactive strategies such as self-withdrawal (Petrova et al., 2021). Therefore, mindfulness enables adolescents facing social adversity to consciously choose and recognize their thoughts, emotions, and feelings but does not generate habitual reactions, then gradually eliminating the automated evaluation process of distressed emotions (Chambers et al., 2009), thus reducing the likelihood of engaging in deviant behaviors in cyberspace. Just as the superego in psychoanalytic theory strengthens its control over the id, it avoids the implementation of socially unacceptable behaviors that result from attempts to vent negative emotions (Chen et al., 2022).

Previous research emphasized that mindfulness includes two components: control of attention and guidance of individual experience (Duan, 2014). Individuals with low mindfulness might find it difficult to pay attention to the awareness of current cognition, thinking, and behavior, so they may lack confidence in their capacity to interpret challenging events or stimuli as new experiences and then have to repeatedly judge them, thereby amplifying their hostile reactions to the outside world and themselves (Borders et al., 2010). Conversely, individuals with higher mindfulness possess a more objective evaluation of capabilities (Creswell and Lindsay, 2014). They gain a stronger sense of control over their attention resources (Kraemer et al., 2016). Hence, they are more capable of shifting the resources from self-reflection to exploration of new experiences, which can alleviate their negative self-evaluation and negative emotions caused by rumination, promote their stable emotional states, and enhance their monitoring and restraint of their behaviors, then alleviate the risk of cyberbullying perpetration (Kowalski et al., 2014). Due to the constraints of adolescents’ social experience, those with low mindfulness will face great challenges in their tolerance and acceptance of negative life events (Xiong et al., 2023). Mindfulness helps adolescents to assess their inner thoughts and external environment as comprehensively as possible, without being limited by their own experiences when facing social adversity (Zhang et al., 2022). Therefore, a mindfulness-based intervention would be helpful to enhance the self-control function of adolescents with high PCSA, they can be more inclined to adjust cognitive strategies to buffer negative emotional experiences caused by rumination and then reduce the risk of cyberbullying perpetration.

This study revealed the association between PCSA and cyberbullying perpetration as well as the mechanism in between, which added to our understanding of the explored constructs and brought practical implications. From a theoretical perspective, the results provide further validation for the general strain theory in the context of cyberbullying perpetration by revealing a significant positive correlation between PCSA and cyberbullying perpetration among adolescents. Additionally, few studies have investigated cyberbullying perpetration from the perspective of the interaction of multiple factors (Fan et al., 2018). This study provided an attempt to explore the mechanism of cyberbullying perpetration more systematically and comprehensively by focusing on the perspectives of cognitive strategies, emotional reactions, and social relationships, and further revealed the mediating role of rumination between PCSA and cyberbullying perpetration. The results verified the applicability of the general aggressive model in cyberbullying situations. Furthermore, previous studies have mostly used environmental variables such as negative life events as independent variables to explore the effects on individuals. Our findings explored the issue of individual consistency of the effect caused by subjective perception of negative life events from a new perspective. The moderated mediation model examined in this study revealed the mechanisms of perceived chronic social adversity, rumination, and mindfulness on aggressive behaviors in cyberbullying situations, which not only validated the limited resource of self-control theory in the context of cyberbullying among adolescents, but also explored whether perceptions of social adversity have the same impact on individuals with different levels of mindfulness, which is beneficial for further revealing the interaction between mindfulness and subjective perception of negative life events on individuals, thus supplied new insights for preventing and intervening cyberbullying perpetration among adolescents.

From a practical standpoint, the results offer guidance on promoting mindfulness-based intervention through educational measures in response to cyberbullying incidents and present new perspectives for educators to reduce the prevalence and negative impact of cyberbullying from the perpetrators. To prevent adolescents from experiencing long-term psychological pain and accumulating the risk of cyberbullying, educators should help adolescents improve their perception abilities to integrate various types of information in life, correct negative cognitive biases, and seek a positive meaning from adversity based on their advantages. For instance, conducting training programs in social skills and team games to enhance adolescents’ adaptive social information processing, emotional response, and social skills, then guide them to establish reasonable expectations for social adversity. Assisting adolescents to explore their unique feelings of exclusion and rejection in multiple identities and social environments, helping them to conduct other supportive relationships and areas of achievement, and executing principles of dialectical-behavioral therapy like establishing core skills in the regulation of emotional responses, distress tolerance, and interpersonal efficacy. Moreover, this study revealed that educators should pay prompt attention to the perceived chronic social adversity and rumination of adolescents, and assist them in cognitive adjustment through mindfulness-based intervention. For instance, Mindfulness-Based-Cognitive Therapy (MBCT) is a systemic strategy that combines elements of cognitive behavioral therapy and mindfulness skills, which can help adolescents accept negative events with a more open mindset. MBCT can also make individuals aware of the possibility of increasing their flexibility in response when they are about to fall into habitual rumination, then reducing rumination and promoting positive mental health (Proeve et al., 2018).

This study had several limitations. Firstly, all variables were measured by self-reported methods, so the objective details of social adversities received little attention, which may be subject to social desirability bias and lack of ecological validity. Future studies could employ a unified standardized stimulus to evaluate the response and emotional arousal level of adolescents to adversities, which could improve the validity of data. For instance, standardized negative life events (presented by scenarios or vignettes that describe common experiences and feelings of adolescents who implied such events) could be used as stimuli to gather the adolescents’ reports of their emotions and feelings about social adversities in strictly controlled situational experiments, so that we would be more confident that we were assessing change in responses related to social adversities rather than change in stressors. This method could improve validity and accuracy over retrospective recall of actual events and reactions and has been valuable as a commonly used approach to assessing responses to stress among large samples in other research areas (Catterson and Hunter, 2010; Zimmer-Gembeck et al., 2011). Secondly, the cross-sectional data could not permit the causality portrayed by the existing theoretical models. Longitudinal experimental design could be applied to explore the direction and the long-term impact between variables. For instance, the use of the ecological momentary assessment method may be more effective in capturing the fluctuation patterns of participants’ cognitive emotional reactions and rumination under different clues such as negative or ambiguous stimuli in social situations, even evaluating the negative emotion and attention allocation of participants when facing different levels of potential threats, thus revealing the dynamic impact of PCSA on cyberbullying perpetration. Thirdly, this study investigated the degree of subjective perception of negative life events among adolescents but did not specifically explore the role of early cyber victimization experience in the formation of cyberbullying perpetration. Compared to other negative life events, as a typical adverse event in cyberbullying scenarios, cyber victimization is more likely to lead to cyberbullying, thus creating a vicious cycle (Sun et al., 2020). Prior literature has found a strong correlation between cyber victimization and cyberbullying perpetration, but the underlying mechanisms remain unclear (Luo et al., 2023). Future studies could further explore the association between early cyber victimization experience and cyberbullying perpetration to examine the mechanisms underlying the role transition from victim to bully in cyberbullying incident. Lastly, the hypothesis mainly focused on rumination that produces negative effects, but rumination includes multiple dimensions, such as maladaptive dimensions of negative passivity and adaptive dimensions of positive problem-solving (Watkins, 2008; Pearson et al., 2011). Future studies could further investigate the interaction between PCSA and different dimensions of rumination to examine their distinct roles in various emotional regulation strategies and aggressive behaviors.