Cognitive and affective empathy

Empathy, commonly understood as a stable personality trait, consists of two components, cognitive and affective empathy (^{Jolliffe & Farrington, 2004}). Cognitive empathy allows individuals to recognise and understand others' emotions, while affective empathy involves the ability to share and experience those emotions, leading to compassion and concern. Some studies have demonstrated the negative correlation between empathy and general aggression (e. g., ^{Vachon et al., 2014}). However, there are inconsistencies in the association between empathy and cyberbullying perpetration in the online context. For instance, while a few studies (e. g., ^{Chen et al., 2020}; ^{Zych et al., 2019}) found a negative association between cognitive and affective empathy with cyberbullying perpetration, other studies reported that only affective (e. g., ^{Ang et al., 2017}; ^{Schultze-Krumbholz & Scheithauer, 2013}; ^{Zych & Llorent, 2018}) or only cognitive (e. g., ^{Šincek et al., 2020}) empathy is associated with cyberbullying perpetration. These findings suggest that the role of empathy in relation to cyberbullying may not be uniform across different online contexts and populations. Moreover, the inconsistencies observed across studies could stem from differences in the age of the participants, as adolescents may interact with digital platforms differently compared to young adults. This highlights the necessity of further investigation to clarify the complex association between empathy and cyberbullying perpetration.

Callous unemotional traits

Callous unemotional (CU) traits are closely linked to the affective dimension of psychopathy (^{White & Frick, 2010}). Individuals with elevated levels of these traits often depreciate moral emotions of guilt and remorse, appearing emotionally cold, and unable to empathise with others' feelings, show compassion or affection (^{Frick & White, 2008}). Moreover, they typically exhibit emotional unresponsiveness, apathy, and a self-centred and deceitful interpersonal style marked by narcissistic tendencies, manipulation, impulsive and irresponsible behaviour, and a tendency to become easily bored. Despite that there is some research on the association between CU traits and aggression (e.g., ^{Fang et al., 2020}; ^{Li et al., 2023}), the specific connection of these traits to cyberbullying perpetration on SNSs remains underexplored. Facebook might offer a platform where individuals with high CU traits can find a community or environment that reinforces their behaviours.

Moral disengagement

Moral disengagement (MD) refers to the cognitive process through which individuals rationalise harmful behaviour and enables them to temporarily suspend self-sanctions by justifying their actions, diffusing responsibility, minimising the harm caused, or dehumanising the victims, among other strategies (^{Bandura, 1999}). Individuals with increased levels of MD tend to be more inclined to distance themselves from their moral principles without experiencing moral emotions such as feelings of guilt and remorse (e. g., ^{Kowalski et al., 2014}). While some prior research has established a link between MD and cyber perpetration in online contexts (e. g., ^{Gini et al., 2014}; ^{Lee & Jang, 2023}), there remains limited research exploring the association between MD and cyberbullying perpetration on SNSs. This scarcity highlights the need for further investigation in this area.

Number of Facebook friends

A larger Facebook network can facilitate prosocial behaviour by providing a platform for users to connect with a diverse community, thereby enhancing access to support and resources. Prosociality on Facebook might manifest in several ways, such as offering advice, sharing resources, expressing compassion, or organising group events (e. g., ^{Marshall et al., 2023}). While these positive interactions offer many benefits, they may also have the potential to increase risky behaviour. As ^{Chan et al. (2021}) highlight in their review, essential features of SNSs such as digital profiles, interpersonal connections, search functionalities, privacy settings, and social visibility create opportunities for cyberbullying. The increased visibility and connectivity associated with larger networks could possibly influence online behaviour. In more details, the number of Facebook friends could act as a moderating variable, potentially influencing how personality traits associate to CPF. While there is limited research directly linking Facebook network size to CPF, ^{Lee (2017}) found that cyberbullying perpetration was marginally negatively associated with the number of Facebook friends. This finding suggests that individuals with fewer Facebook friends were more likely to engage in cyberbullying perpetration. Moreover, cyber perpetrators with smaller networks may also exhibit higher levels of CU traits and MD, along with lower levels of cognitive and affective empathy (^{Chan et al., 2019}; ^{Decety & Jackson, 2004}; ^{Fang et al., 2020}). This lack of empathy and moral flexibility could allow them to inflict harm with limited emotional inhibition, while without a larger network to reinforce social norms, these individuals may feel less constrained by positive behavioural expectations or social accountability on SNSs (e. g., ^{Neben et al., 2015}). In other words, individuals with fewer connections might lack this regulatory social structure, leading to lower accountability and a potentially higher likelihood of CPF. Hence, the number of Facebook friends can serve as a moderating variable that potentially amplifies or mitigates the association between personality traits and CPF. These findings emphasise the need to clarify the role of the number of Facebook friends and its association with cyberbullying perpetration, particularly its moderating effect on the association between personality traits and CPF.

Gender

The literature presents conflicting evidence on gender differences in cyberbullying perpetration on SNSs. While some studies report no significant gender differences (^{Chan et al., 2019}; ^{Hood & Duffy, 2018}; ^{Kwan & Skoric, 2013}), others suggest that males are more involved (^{Chan et al., 2021}; ^{Kokkinos et al., 2016}; ^{Lowry et al., 2016}). These inconsistencies may arise from variations in demographics, cultural contexts, and methodologies, as well as the evolving nature of digital interaction (e. g., ^{Bayraktar, 2015}). While there might be inconsistencies in the expression of cyberbullying perpetration across genders, early studies suggest that both cognitive and affective empathy tend to be higher in females due to socialisation processes that encourage emotional attunement, caregiving, and relationship-building from an early age (^{Baron-Cohen, 2004}). Individuals who are more empathetic and emotional, demonstrating greater emotional contagion, sensitivity to distress, and prosocial behaviours tend to be less likely to engage in actions that hurt others (e. g., Chen et al., 2014; Christov-Moore et al., 2014). Gender differences are also evident in CU traits, with males being more likely to exhibit low emotional arousal (Awada et al., 2022). Males with high CU traits tend to report positive emotions when imagining unethical behaviours. Similarly, men demonstrate a higher level of MD compared to women. This difference can be attributed to several factors, including societal expectations (e. g., ^{Polanco-Levican & Salvo-Garrido 2023}). Traditionally, men are encouraged to be more dominant and competitive, leading to a greater acceptance of aggression in their attitude. Specifically, in the context of cyberbullying, men tend to exhibit higher levels of MD comparing to women (e. g., ^{Ding et al., 2023}). It has been reported that men tend to downplay the severity of their actions and tend to blame or dehumanise the victim for provoking the cyberbullying (e. g., ^{Jeong et al., 2024}). Given these gender differences, this study also aims to investigate how personality traits differently associate to CPF for males and females, shedding light on distinct gender patterns.

The present study

The study has two objectives. The first one is to examine whether the association between personality traits (cognitive and affective empathy, CU traits, and MD) and CPF is moderated by the number of Facebook friends a user has. The second objective is to explore potential gender-specific patterns in how these individual factors associate to CPF. In light of previous findings, cognitive and affective empathy are expected to have a negative relationship with CPF, while CU traits and MD are anticipated to show positive associations (H1). Furthermore, it is hypothesised that the relationship between personality traits and CPF will vary depending on the number of Facebook friends (H2). More, it is hypothesised that males will exhibit higher involvement in CPF compared to females (H3). Lastly, it is hypothesised that distinct gender-specific associations between personality traits and CPF will emerge (H4).

Method

Results

Discussion

Two key objectives guided the current study: 1) to clarify whether the association between personality traits and CPF is moderated by the number of Facebook friends, and 2) to explore whether gender-specific patterns exist in the relationship between personality traits and CPF.

The study revealed significant associations between CPF and CU traits, MD, and empathy, ratifying thus the hypothesis 1 of the study. CPF was positively correlated with CU traits and MD, while negatively with both cognitive and affective empathy, supporting earlier research which highlights their role in facilitating cyber perpetration (e. g., ^{Lee & Jang, 2023}; ^{Li et al., 2023}; ^{Šincek et al., 2020}; ^{Zych et al., 2019}). According to Bandura's theory of MD, individuals with elevated levels of CU traits and MD are more likely to rationalise harmful behaviour through cognitive processes such as dehumanising the victim or distorting the consequences of one's actions (e. g., ^{Fang et al., 2020}; ^{Li et al., 2023}). According to ^{Li et al. (2023)}, SNSs create a sense of detachment, making it easier for individuals to morally disengage and avoid confronting the consequences of their actions. Moreover, Tan (2024) suggests that the lack of social cues in online communication reduces the emotional visibility of victims, making it harder for perpetrators to understand the impact of their behaviour.

The interaction analyses between personality traits and the number of Facebook friends were not significant to the association with CPF, thus the initial hypothesis 2 was rejected. As for the conditional effects, significant associations between the personality traits and CPF were observed across various levels of the moderator. These findings reflect that social network size is probably less critical than personality factors in activating perpetration and the impact of personality traits on CPF remains stable, regardless of the size of one's Facebook network (e. g., ^{Chan et al., 2019}).

The study found no significant gender differences in the overall frequency of CPF, even when analysis was performed only for the group of the perpetrators, thus rejecting the hypothesis 3 of the study. Some research focusing on young adults has also found no gender differences (e. g., ^{Chan et al., 2019}). The lack of gender differences in CPF could be attributed to specific characteristics of the sample, such as cultural beliefs, gender norms, or the particular cyberbullying perpetration scale used. Gender differences were observed in personality traits. Males reported lower levels of both cognitive and affective empathy and higher levels of CU traits and MD compared to females. These findings align with existing literature on gender differences in empathy, CU traits, and MD (e. g., Awada et al., 2022; ^{Decety & Yoder, 2016}; Eilts & Baker, 2023; ^{Polanco-Levican & Salvo-Garrido 2023}). Lower empathy in males may be rooted in both biological factors and evolutionary pressures, such as differences in parental investment and socialisation process during early years (e. g., ^{Decety & Yoder, 2016}). Additionally, males' higher CU traits and MD may be linked to a greater tendency towards physical aggression and the suppression of moral emotions (Camara et al., 2025; ^{Ding et al., 2023}).

The gender-specific analyses revealed that CPF was significantly correlated with MD, CU traits, and cognitive empathy in males, ratifying thus the hypothesis 4 of the study. This finding mirrors previous research showing that males typically score higher on MD while lower on empathy (Eilts & Baker, 2023; ^{Jeong et al., 2024}; Navarro-Rodriguez et al., 2023; ^{Lee & Jang, 2023}). Further regression analysis showed that MD and cognitive empathy made significant contributions to the explanation of the statistical variance of CPF with number of Facebook friends contributing marginally significantly to the model and all effect sizes, albeit significant, being small. These findings suggest that the visibility of actions to a wider audience could provide social validation and encourages perpetration while diffusing personal responsibility, as individuals may feel less accountable when others are involved or observing (^{Bandura, 1999}). Males with higher cognitive empathy may use this ability to manipulatively engage in CPF while employing MD strategies to justify their actions (e. g., ^{Frick & Kemp, 2021}; ^{Logoz et al., 2023}; ^{Ritchie et al., 2022}).

In female's group, CPF was significantly negatively correlated with cognitive and affective empathy, as well as positively with CU traits and MD. Further regression analysis showed that cognitive and affective empathy made significant contributions to the explanation of the statistical variance of CPF among females. It may seem counterintuitive that empathy, typically linked to prosocial behaviour (e. g., ^{Ritchie et al., 2022}), associates with CPF in females. One explanation is that females with high empathy can exploit victims' vulnerabilities, make their antisocial actions more targeted and harmful, navigate social dynamics, protect their status, or address perceived threats, thus they may make Facebook an effective tool for relational aggression (e. g., ^{Ding et al., 2023}; Eilts & Baker, 2023).

Conclusions

The size of an individual's Facebook network did not significantly moderate the association between personality traits and cyberbullying perpetration on Facebook. Also, no significant gender differences were found in the frequency of cyberbullying perpetration on Facebook, suggesting both genders engage in this behaviour at comparable rates. However, distinct gender patterns emerged in the association between personality traits and cyberbullying perpetration on Facebook. Male perpetrators displayed lower levels of cognitive empathy and higher levels of moral disengagement. Female perpetrators were found to have lower cognitive and affective empathy. These findings contribute to the literature of cyberbullying perpetration in emerging adulthood and have possible practical implications for the development of gender-specific policies.