The Effects of Covert Narcissism on Chinese College Students Cyberbullying: The Mediation of Hostile Attribution Bias and the Moderation of Self-Control

Introduction

People’s online lives are getting increasingly affluent as the Internet becomes more popular, which gives rise to a new phenomenon known as cyberbullying. Cyberbullying is the term for a group or individuals who routinely and persistently use electronic tools like mobile phones or the Internet to engage in aggressive and hostile behavior,¹ which is about spreading untrue information about others, sending obscene text messages, and using derogatory language. Compared with conventional bullying, cyberbullying is a more recent research topic, whose incidence surged drastically between 2015 and 2019.² More importantly, researchers cannot ignore the consequences of cyberbullying on people’s mental health.³ Cyberbullying has a wide range of adverse consequences on individuals, such as anxiety, depression,^4,5 suicidal tendencies, and suicidal behaviors.^6,7 A literature review summarized that cyberbullying is related to various adverse outcomes, comprising psychological health issues, drug abuse, melancholy, stress, autism spectrum disorders, developmental disorders, obesity, and asthma.⁸ Therefore, intervention and prevention of cyberbullying have become crucial due to the severe consequences of cyberbullying in modern society. The current study attempted to provide more references for preventing cyberbullying.

According to the 51st Statistical Report on China’s Internet Development released by the China Internet Network Information Center (CNNIC),⁹ as of December 2022, Chinese Internet users have increased by 35.49 million compared with December 2021. As the scale of netizens continues to expand, the phenomenon of cyberbullying among Chinese groups may become more and more serious, especially among college students. First, college students depend heavily on the Internet in their studies and daily lives. Long-term exposure to the Internet may impact their cognition. Second, anonymous online communication allows college students to talk freely without fear of punishment, providing more chances to bully others online. Furthermore, research indicated that college students are more prone than younger people to bully others online regularly,^2,10 because they frequently exhibit more impulsivity, lack of sympathy, and may even be more aggressive and antisocial.¹¹ Although cyberbullying has become one of the research hotspots worldwide, there is relatively little empirical evidence concerning Chinese people. Most studies in China have focused on cyberbullying of teenagers^12,13 and paid less attention to cyberbullying among college students.¹⁴ Therefore, this study mainly investigated cyberbullying among college students to enrich the literature.

Previous research has explored many factors that influence cyberbullying, including the anonymity of cyberspace,¹⁵ family factors,¹⁶ school factors,¹⁷ childhood experiences,¹⁸ personalities,¹⁹ cognition,²⁰ and others. However, very few studies (especially research on Chinese people) have examined the impact of narcissism on cyber behavior, particularly covert narcissism. Consequently, the purpose of this study was to investigate how covert narcissism and cyberbullying are related. For the first time, we developed a model that took into account both the mediating role of hostile attribution bias and the moderating role of self-control. In addition, previous research has demonstrated that gender has a significant role in cyberbullying.²¹ Generally, males are more inclined to value aggression than females due to gender role socialization,²² so males are more likely to become cyberbullies than females.^23,24 Therefore, this study will consider gender differences in cyberbullying to more clearly explore the relationship between cyberbullying and other variables.

Covert Narcissism and Cyberbullying

Researchers have had a long-standing interest in narcissism, and many have investigated the links between distinct narcissistic traits and diverse social behaviors. However, few have linked covert narcissism to cyberbullying. Covert narcissism refers to depression, shame, sensitivity, shyness, low tolerance for others’ attention, high sensitivity to criticism or failure, and more passiveness in society.²⁵ Compared with overt narcissism, covert narcissism is usually associated with psychological abnormalities.²⁶ Consequently, examining the impact of covert narcissism on cyberbullying is more helpful in understanding the occurrence mechanism of cyberbullying.

Covert narcissism and aggression are related. The psychodynamic mask model of narcissism proposes that the exquisite appearance of narcissists may be a representation to cover up their potential sense of insecurity and inferiority.^27,28 Then, under external stimuli, the repressed negative emotions of narcissists are very likely to transform into aggressive behavior. According to the threatened egotism theory,²⁹ narcissists may confront threats with anger rather than sadness or anxiety when encountering situations inconsistent with their mindset. The exaggeration and self-righteousness of narcissists quickly make them attribute mistakes to others and rarely reflect on themselves, resulting in anger. Therefore, narcissists may become aggressive and manipulate others to reach their goals if they feel they have not received the expected recognition or appreciation.^30,31 In addition to theoretical explanations, relevant research has proved that covert narcissism could predict aggression.^32,33

As one of the manifestations of aggressive behavior, cyberbullying may also be closely related to covert narcissism. First, covert narcissists show potentially low self-esteem and empathy,²⁵ which may affect cyberbullying. When they receive evaluations that fall short of their expectations, they do not directly show anger and hostility like overt narcissists. They are more likely to vent negative emotions subtly or indirectly, such as attacking others online.³⁴ Second, the anonymity of cyberspace also provides convenience for covert narcissists.³⁵ Covert narcissists skilled at masking themselves^27,28 are inclined to release their suppressed negative feelings on social network sites. Although few studies exist on the association between covert narcissism and cyberbullying, one study taking Chinese teenagers as the research object proved that covert narcissism could positively predict cyberbullying.³⁶ Hence, we may conclude from the analysis above that covert narcissists frequently employ indirect assaults and that cyberbullying is likewise an indirect attack.

Hostile Attribution Bias as a Mediator

Cognitive factors are crucial in the study of aggression. As one of the factors influencing aggressive behavior, hostile attribution bias is the propensity to interpret vague social information as a threat or hostility.³⁷ Prior research has linked hostile attribution bias and aggressive behavior,^38,39 but few studies have discussed its mediating effect in the association between covert narcissism and cyberbullying. This study will conduct a preliminary theoretical analysis and literature review of the mediating role of hostile attribution bias.

First, hostile attribution bias and cyberbullying may be positively correlated. There is theoretical support for the relationship between hostile attribution bias and aggressive behavior, which may also apply to cyberbullying. The General Aggression Model (GAM)⁴⁰ suggests that individuals’ hostile attributions to situational cues may induce aggressive behavior; the Social Information Processing Model (SIPM)⁴¹ suggests that individuals’ aggressive behavior results from hostile attributions of others’ behavior during the cue-encoding phase. Combining these two theories, we can conclude that individuals may attack others when they make hostile attributions to a situation or cue. These theories may apply to research on cyberbullying.^42,43 Due to the virtuality and decentralization of the network environment, individuals may be unable to obtain more clues from other people’s body movements and speech, resulting in a one-sided understanding of the network environment. Thus, in online environments with ambiguous cues, individuals are inclined to make false or hostile attributions to the cues they receive.⁴¹ And the anonymity of cyberspace also creates conditions for cyberbullies.¹⁵ Existing research has also found that hostile attribution bias boosted cyberbullying.^44,45 However, some scholars suggests that bullying is an unprovoked and active attack, which means that hostile attribution bias is not associated with cyberbullying.⁴⁶ Therefore, more research needs to explain their relationship.

Second, hostile attribution bias may also relate to narcissism and mediate the relationship between covert narcissism and cyberbullying. The threatened egotism theory²⁹ proposes that narcissists will interpret social circumstances with more hostility when there is ambiguity,²⁷ resulting in hostile behavior and anger.^47,48 Furthermore, narcissists frequently exhibit improper emotional and behavioral responses during social interactions and frequently display hostility without a threat.²⁶ Previous research indicated that covert narcissists tend to have hostile attribution bias more than overt narcissists.^49,50 Specifically, covert narcissists exhibit heightened sensitivity, as do individuals with high degrees of hostile attribution bias.⁵¹ Baumeister also proposed that narcissism may affect aggressive behavior through hostile attribution bias.⁵² In particular, the oversensitivity of covert narcissists makes them easy to misunderstand others, and after forming a hostile attribution bias, they may indirectly attack others in hidden places. Hence, this study infers that covert narcissism may affect cyberbullying through hostile attribution bias. Nevertheless, empirical studies also show that narcissism and hostile attribution bias are unrelated.^47,53 It is required to conduct further study on the connections between aggression, hostile attribution bias, and narcissism.

Self-Control as a Moderator

The impact of covert narcissism on cyberbullying may vary due to individual self-control. The ability to withstand, harness, and regulate one’s emotions and impulses is known as self-control.⁵⁴ This study examined whether self-control could moderate the mediation and direct pathways of covert narcissism and cyberbullying.

First, self-control may affect individual behavior. The self-control theory⁵⁵ suggests that individuals who lack self-control frequently engage in aggressive behaviors, while those with great self-control typically exhibit little antisocial behavior.^56,57 Individuals who lack self-control are typically selfish, indifferent, and primarily concerned with their immediate needs.⁵⁸ When they encounter conflicts, they may impulsively make aggressive behaviors immediately. Prior studies have also concluded that self-control and aggression are associated: individuals with high levels of self-control show less aggressive behavior;^59,60 low self-control can predict cyberbullying.^61–64 Moreover, a recent study has found that self-control moderated the association of narcissism with antisocial tendencies.⁶⁵ That is, high narcissists who lack self-control are predisposed to aggression,⁶⁶ whereas narcissists with moderate to high self-control should have low antisocial tendencies.⁶⁷ Thus, we speculate that self-control may mediate the direct pathways of covert narcissism and cyberbullying.

Second, self-control may affect individual cognition. According to the risk-buffering model,⁶⁸ protective variables have the potential to reduce or even neutralize the negative impacts of risk factors. The protective effect of self-control can reduce the degree of personal hostility.⁶⁹ Existing research suggested that high levels of effortful control may help individuals suppress hostile attribution bias,⁷⁰ and poor self-control and high degrees of hostile attribution bias are related.⁷¹ Individuals may use self-control to reduce hostility or suppress the impulse to do certain behaviors⁶⁹ after the formation of hostile attribution bias. Thus, self-control may mediate the relationship between hostile attribution bias and cyberbullying.

The Current Study

Based on existing theories and research, this study aimed to explore the connection between covert narcissism and cyberbullying, the mediating effect of hostile attribution bias, and the moderating effect of self-control. Hence, the current study constructed a moderated mediation model (Figure 1) and presented the following hypotheses:

Hypothesis 1. Covert narcissism would positively predict cyberbullying.

Figure 1 Hypothesized moderated mediation model.

Materials and Methods

Participants and Procedure

Through the convenient sampling, 734 college students from a university in Anhui province, China, filled out the questionnaire. Ultimately, the final sample comprised 672 valid data (91.55% out of the original 734). Some students fill in the questionnaire through the questionnaire collection website (https://www.wjx.cn/), and other parts of students fill in the questionnaire in the classroom. Before conducting the questionnaire survey, we guaranteed the informed consent and the right to withdraw from the survey freely for all participants. Table 1 is the descriptive statistics of demographic variables.

Table 1 Descriptive Statistics of the Demographic Variables

Measurement

Covert Narcissism

We measured covert narcissism using a 15-item scale suitable for the Chinese cultural background compiled by Zheng and Huang⁷² based on previous scales.⁷³ The scale has three dimensions: privilege, self-admiration, and susceptibility. College students rated the items (eg, “I often feel useless”) on the 5- point Likert type from 1 (very nonconforming) to 5 (very conforming), with higher values indicating higher degrees of covert narcissism. The Cronbach’s alpha in this research was 0.878.

Cyberbullying

A 12-item cyberbullying scale for college students compiled by Çetin and his colleagues⁷⁴ and revised by Xu⁷⁵ was adopted. The scale includes two dimensions: direct and indirect bullying. For each item (eg, “I have used abusive language online”), college students rated on a 4-point Likert scale ranging from 1 (never) to 4 (always), with higher values indicating higher frequencies of cyberbullying. The Cronbach’s alpha in this research was 0.901.

Hostile Attribution Bias

The hostile attribution bias scale was partially back-translated from the Social Information Processing-Attribution Bias Questionnaire⁷⁶ recomposed by Li.⁷⁷ The scale includes four dimensions: direct hostile attribution, indirect hostile attribution, neutral attribution, and instrumental attribution. College students graded the questions on a 4-point Likert scale, with higher values indicating higher degrees of hostile attribution bias. The Cronbach’s alpha in this research was 0.863.

Self-Control

The brief self-control scale compiled by Tangney and his colleagues⁷⁸ was employed. The scale has 13 items in total, including 9 reverse-scoring questions. College students rated the items (eg, “I can resist temptation very well”) on the 4- point Likert scale ranging from 1 (very nonconforming) to 4 (very conforming), with higher values indicating higher degrees of self-control. The Cronbach’s alpha in this research was 0.846.

Data Analysis

We used SPSS 23.0 and PROCESS 4.1 to analyze the data statistically. First, we used the independent-samples t-test to assess demographic differences in continuous variables. Second, we tested correlation analysis on research variables. Finally, using Hayes’ PROCESS macro program Model 4 and Model 15,⁷⁹ we examined the mediating effect of hostile attribution bias and the moderating effect of self-control. All analyses adopted the bootstrap method (N=5000) and used standardized variables. Given that the existing research revealed that gender and registered residence could affect individuals’ cyberbullying,^23,80 we regarded gender and registered residence as covariates in this study.

Results

Common Method Bias Test

According to Harman’s single-factor test findings,⁸¹ this study did not show a significant common method bias, with the total variance explanation rate of the first common factor being 21.48% (<40%).

Demographic Difference Analyses

We used the independent samples t-test to analyze the difference in demographic variables. As shown in Table 2, there are only gender differences in covert narcissism, self-control, and cyberbullying (t = −3.03, p < 0.01; t = 2.41, p < 0.05; t = 4.16, p < 0.001): females displayed higher levels of covert narcissism than males; males perpetrated cyberbullying more frequently than females; males had higher levels of self-control than female.

Table 2 Demographic Difference Analysis

Correlational Analyses of the Main Variables

We performed the correlation analysis for all main variables. Table 3 shows that cyberbullying was positively correlated with covert narcissism (r = 0.43, p < 0.01) and hostile attribution bias (r = 0.33, p < 0.01). Covert narcissism was positively correlated with hostile attribution bias (r = 0.52, p < 0.01) but negatively correlated with self-control (r = −0.18, p < 0.01); hostile attribution bias was negatively correlated with self-control (r = −0.14, p < 0.01).

Table 3 Means, Standard Deviations, and Correlations of the Main Variables (N=672)

Mediation Test of Hostile Attribution Bias

Based on the analysis of differences in demographic variables, we took gender as a covariate. The results of the mediation effect analysis are shown in Table 4 and Figure 2, which indicated that covert narcissism positively predicted cyberbullying, β = 0.46, p < 0.001 (path c); covert narcissism significantly and positively predicted hostile attribution bias, β = 0.51, p < 0.001 (path a); hostile attribution bias significantly positively predicted cyberbullying, β = 0.16, p < 0.001 (path b). Hostile attribution bias partially mediated the relationship between covert narcissism and cyberbullying: ab = 0.08, SE = 0.02, 95% CI = [0.04, 0.12]. The mediating effect accounted for 17.66% of the total effect. After adding the mediating variable, the direct effect of covert narcissism on cyberbullying was significant, β = 0.38, p < 0.001 (path c’). Hence, Hypothesis 1 and Hypothesis 2 were verified.

Table 4 Testing the Mediating Effect of Hostile Attribution Bias on Cyberbullying

Figure 2 Mediation model. c is the total effect of covert narcissism on cyberbullying, a is the effect of covert narcissism on hostile attribution bias, b is the effect of hostile attribution bias on cyberbullying and c’ is the direct effect of covert narcissism on cyberbullying. ***p<0.001.

Moderation Test of Self-Control

As shown in Table 5, self-control moderated the relationship between covert narcissism and cyberbullying (β = −0.07, p < 0.05) but did not moderate the relationship between hostile attribution bias and cyberbullying (β = 0.02, p > 0.05). The final model of this study is shown in Figure 3.

Table 5 Testing the Moderated Mediation Effect of Self-Control of Covert Narcissism on Cyberbullying

Figure 3 The final model of this study. *p<0.05; ***p<0.001.

We conducted the simple slope test to explain further the moderating effect of self-control in the relationship between covert narcissism and cyberbullying. As shown in Figure 4, under low levels of self-control, covert narcissism predicted cyberbullying positively and significantly (β_simple = 0.26, t = 10.38, p < 0.001); under high levels of self-control, the predictive effect of covert narcissism on cyberbullying was diminished (β_simple = 0.21, t = 10.21, p < 0.001). Accordingly, the positive predictive effect of covert narcissism towards cyberbullying steadily diminished as self-control increased.

Figure 4 Moderating effect of self-control on the relationship between covert narcissism and cyberbullying. Abbreviations: CN, covert narcissism; SC, self-control.

Discussion

While many studies have revealed the connection between narcissism and aggressiveness, very few have examined the connection between covert narcissism and cyberbullying. It is essential to study the occurrence mechanism of cyberbullying from the personality perspective. This study is the first to propose a model with hostile attribution bias as a mediator and self-control as a moderator. The results revealed that covert narcissism predicted cyberbullying positively and significantly, despite in part through hostile attribution bias; self-control was a moderator of the connection between covert narcissism and cyberbullying.

Covert Narcissism and Cyberbullying

This study showed that covert narcissism positively predicted cyberbullying, consistent with Hypothesis 1 and previous studies.³⁶ Individuals with covert narcissistic personalities are likely to engage in cyberbullying. First, covert narcissists tend to make indirect attacks to cover up their potential inferiority complex and anger.²⁷ The anonymity of cyberspace also provides convenience for them.¹⁵ Second, some covert narcissists are more active online than in real life and expect compliments from others online to satisfy their narcissistic needs. If these expectations are unmet, they could become irritated and hurt others through cyberbullying. Furthermore, covert narcissists may bully others online to control others and establish dominance by proving their importance and authority.⁸² The findings of this study also corroborate the threatened egotism theory. When individuals with higher levels of covert narcissism interpret social situations, they perceive more hostility and anger, which increases aggression.²⁹ Thus, covert narcissists with low self-esteem and low empathy who always feel threatened are likelier to engage in cyberbullying behaviors in anonymous cyberspaces than in real life. Furthermore, covert narcissists among college students are the focus of future research. They spend more time using electronic media online and have easier access to the Internet, leading to a higher frequency of cyberbullying.⁸³

In addition, there was no significant difference in cyberbullying by registered residence, consistent with previous research.^84,85 The difference in Internet access between rural and urban of China is gradually decreasing,⁹ and college students have long been influenced by campus culture. Thus, the effect of registered residence on college students cyberbullying is relatively small and did not reach a significant level in this study. However, this study found that women have more covert narcissism than men but exhibit fewer cyberbullying behaviors. The possible explanation is that women behave as if they may have more latent narcissism, rage, and hostility than men.⁸⁶ However, they may subtly satisfy their narcissistic demands and behave less aggressively due to societal and individual expectations around gender roles.⁸⁷ It is worth mentioning that Field⁸² believes that the characteristics of cyber victimization include shyness, anxiety, and inferiority, which are similar to those of covert narcissism. Therefore, there may also be a relationship between covert narcissism and cyber victimization. Future research can explore the influencing factors and occurrence mechanism of cyberbullying from the victims’ perspective.

Mediation of Hostile Attribution Bias

Evidence from this study indicating hostile attribution bias mediated the relationship between covert narcissism and cyberbullying supported Hypothesis 2. That is, cognitive factors can mediate between personality and cyberbullying. Nonetheless, hostile attribution bias only accounts for a portion of the association between covert narcissism and cyberbullying. Future studies can consider additional mediating factors.

Divide the mediation pathway in this study into two sections to explain. First, covert narcissism positively predicted hostile attribution bias, which supports previous studies.^49,50 Explained by the threatened egotism theory,²⁹ because of their inherent sensitivity, repression, and shame, covert narcissists tend to see situations or clues as threatening, which triggers anger. That is, covert narcissists are likely to make mistakes during the online coding phase,⁴¹ making hostile attributions of others’ intentions. Therefore, while covert narcissists are highly concerned about themselves, they are also susceptible to the evaluation of others. Their underlying low self-esteem may make them more inclined to make aggressive assumptions about the intentions of others, especially in unclear circumstances. Moreover, the information given by the network environment is one-sided and lacks verbal, physical, or practical cues, which increases the possibility of hostile attribution bias for covert narcissists. Contrary to the views of this study, Law and Falkenbach⁵³ suggest that narcissists are too selfish and do not care about the intentions of others, so they will not show hostile attribution bias. Notably, this study investigated the characteristics of covert narcissism, which is different from other forms of narcissism (eg, state narcissism and overt narcissism). Unlike overt narcissists, who are exaggerated and arrogant, covert narcissists are introverted, have low self-esteem, and are overly sensitive and vigilant to others’ judgments.²⁶ Immersed in their world, they can quickly become hostile towards others when something does not meet their expectations, which is also why the mental health level of covert narcissists is relatively low.²⁶ However, more empirical evidence is needed to draw clear conclusions about the relationship between covert narcissism and hostile attribution bias.

Second, this study discovered that hostile attribution bias predicted cyberbullying, consistent with previous research.^44,45 When covert narcissists develop a hostile attribution bias, they will likely engage in aggressive behavior. Due to the anonymity of the Internet,¹⁵ covert narcissists are more likely to translate their emotions and hostile attribution tendencies into actual behaviors, leading to cyberbullying. However, previous research suggested that the victim rather than the bully is related to hostile attribution bias.⁴⁶ That is, victims may develop negative beliefs about the motives of others after experiencing multiple injuries. Nevertheless, it is worth discussing whether the victim will transform cognition into action after forming a hostile attribution bias and then attack others. If victims tend to make hostile attributions and develop the attribution style, they may attack others and become bullies. Then whether it is a victim or a bully is related to hostile attribution bias. Hence, future studies can examine the mechanism of cyberbullying from both the perspective of the victim and the bully to provide more evidence and literature support.

In conclusion, this study interpreted the mediation model of covert narcissism and cyberbullying as a path of “personality → cognition → behavior”, revealing the formation process of cyberbullying by narcissists in the information age.

Moderation of Self-Control

The association between covert narcissism and cyberbullying was moderated by self-control, in line with the research hypothesis and previous findings.^66,67 In particular, the positive predictive effect of covert narcissism on cyberbullying dwindled as the degree of self-control rose. Hence, via self-control, covert narcissists can control their cyberbullying behavior. First, the self-control theory⁵⁵ suggests that individuals with poor self-control exhibit more antisocial actions, implying that self-control is negatively associated with cyberbullying, consistent with the result of this study. Second, individuals with low self-control exhibit deficits such as self-centeredness, short tempers, impulsiveness, and poor control. Narcissists lack self-control, are prone to impulsiveness,⁶⁷ and are more prone to be cyber bullies. However, to satisfy narcissistic needs, some individuals may enhance self-control and reduce their bad behavior to obtain external praise. Covert narcissists, in particular, fear negative evaluation⁸⁸ and may use self-control to overcome their aggressive tendencies to avoid social exclusion. Therefore, as self-control increases, the cyberbullying behavior of covert narcissists can be controlled to a certain extent.

Inconsistent with research hypothesis, the current study found that self-control did not moderate the relationship between hostile attribution bias and cyberbullying. Although the results revealed that self-control was negatively correlated with hostile attribution bias and cyberbullying, which is consistent with previous studies,^64,71 this correlation cannot explain the role of self-control in the mediation pathway of covert narcissism and cyberbullying. Covert narcissists have hypersensitivity to hostility,⁸⁹ and once they feel hostility from the outside world, it is not easy to regulate their emotions or behavior.⁹⁰ For covert narcissists, it is feasible to improve their self-control ability to suppress cyberbullying behavior. However, it is difficult to overcome this bias through self-control to change their behavior tendency after forming a hostile attribution bias. Existing studies suggest that self-control can suppress personal hostility,^69–71 but covert narcissists’ low tolerance²⁵ makes it difficult to suppress potential cognition by improving self-control. Regardless, the connection between self-control, covert narcissism, hostile attribution bias, and cyberbullying needs further study.

Implications and Limitations

Although existing research has examined many influencing factors of cyberbullying, there is relatively little evidence of covert narcissism and hostile attribution bias. This study explained the mechanism of covert narcissism in cyberbullying from the perspective of bullies and proved the role of hostile attribution bias and self-control among the pathways, which provide new evidence for the research field and contribute to cyberbullying prevention and intervention. Based on this study, educators should pay attention to their personality development and cognitive behavior when educating college students, which requires the joint efforts of families, schools, and society to cultivate their wholesome personalities and form correct attribution; educators should help students strengthen their self-control ability and prevent them from indulging in cyberbullying.

In addition, this study also has some things that could be improved. First, making causal inferences about the complex links between research variables is difficult because the study is cross-sectional. Second, this study adopted the self-report method, and social desirability may influence the participants. Future research can adopt various research methods to examine the mechanism of cyberbullying.

Conclusion

The current study makes an effort to look at how covert narcissism links cyberbullying and other influencing factors from the viewpoint of bullies. The results of this study discovered that covert narcissism could affect cyberbullying through hostile attribution bias. Self-control moderated the relationship between covert narcissism and cyberbullying. More importantly, this study has significant implications for the intervention and prevention of cyberbullying and additional evidence for the relationship between covert narcissism and cyberbullying.