Bullying is conceptualized as an interpersonal act of systematic, repetitive, and intentional aggression toward someone who lacks power to self-defend (Olweus, 1994). For victims, it is a highly stressful and adverse experience. A factor that has been suggested both as precursor and consequence of victimization is dysfunctional interpersonal relationships. For example, compared to non-involved individuals, victims have more trouble forming and sustaining friendships and romantic relationships (e.g., Ellis and Zarbatany, 2007). Additionally, victims describe their interpersonal relationships as being of lower quality and lacking trust and affection (e.g., Goldbaum et al., 2003; Jantzer et al., 2006). Having such interpersonal struggles has also been suggested to play a role in explaining mental health problems such as depression, in bullied (Hansen et al., 2012; Arseneault, 2018), and non-bullied populations (Hammen, 2006; Beevers et al., 2007).

To better understand victims’ interpersonal struggles and thus also their mental health problems, researchers have recommended studying concepts such as emotional intelligence and competence (Mayer and Cobb, 2000; Lomas et al., 2012). Emotional intelligence represents the abilities to recognize and interpret one’s own and others’ emotions, along with using emotional information to guide and manage one’s own thoughts and emotional responses (Salovey and Mayer, 1990; Mayer et al., 1997, 2008). This also includes affective empathy (i.e., the ability to experience how another person feels) and cognitive empathy (i.e., the ability to understand how another person feels; see van Noorden et al., 2015). Emotional competence entails applying emotional intelligence skills to guide socially acceptable behavior (Zych et al., 2017). The appropriate utilization of emotional competencies is considered essential for successful social interactions and interpersonal relationships (Keltner and Haidt, 1999; Romasz et al., 2004). In the present study, our focus was on emotion recognition as well as understanding of other’s emotions.

Currently, there is mixed evidence regarding differences in emotion recognition and understanding emotional expressions between victims of bullying and non-involved individuals. Victims may have more difficulties with correctly recognizing and interpreting emotions. Lower emotion recognition has been prospectively related to a higher chance of experiencing peer victimization 6months later (Miller et al., 2005). Other cross-sectional studies have also reported victims being generally less accurate in recognizing facial emotional expressions, and specifically in recognizing anger, fear, and disgust compared to non-involved individuals (e.g., Ciucci et al., 2014). Victims have also been shown to overinterpret others’ intentions as hostile (Ziv et al., 2013) and to be prone to misclassify emotions such as fear as anger (Ciucci et al., 2014) but also to misclassify anger as fear (i.e., a fearful bias; cf. DiLalla and John, 2020). Of note however, most of these findings were qualified by type of victimization (Woods et al., 2009), the victim’s gender (Ciucci et al., 2014), or the intensity of the expressed emotion (Pozzoli et al., 2017). Additionally, one study found no evidence for emotion recognition difficulties in victims compared to bullies, bully-victims (i.e., persons who both bully others and are victimized), or non-involved individuals (Guy et al., 2017).

Results on the association between victimization and empathy are similarly inconsistent. In their systematic review of 40 studies, van Noorden et al. (2015) found victimization to be associated with reduced cognitive but not affective empathy, suggesting they are able to experience others’ feelings but not understand them well. Nonetheless, in a later empirical study, the same group found that perceived severity of victimization was positively associated with both types of empathy (van Noorden et al., 2016). Moreover, severe victims reported higher levels of both cognitive and affective empathy than non-involved individuals. While this could suggest that being victimized more severely might positively influence how victims understand and experience how others feel, the systematic review of van Noorden et al. (2015) suggests that being victimized might either not be related to empathic competencies, or that a negative association exists.

Emotional intelligence skills such as correctly identifying and interpreting emotions are fundamental to correctly process social information and to instigate appropriate interpersonal behaviors (cf. Social Information Processing Model; Crick and Dodge, 1996). Potential alterations in victims regarding emotion recognition and understanding, as described above, could therefore lead to altered interpersonal behaviors. In line with this, compared to non-involved individuals, victims have been reported to behave more submissively and lack assertiveness, while also reacting more aggressively (e.g., Perren and Alsaker, 2006; Lansford et al., 2010; Manring et al., 2018). These behaviors are generally perceived as dissatisfying and unpleasant by others (Moskowitz, 2009, 2010) and can perpetuate negative social interactions and bring about re-victimization (Dodge et al., 2003; Brendgen and Poulin, 2018). This in turn may explain why victims are more often rejected and have lower-quality social relationships than non-involved individuals (Ellis and Zarbatany, 2007; Veenstra et al., 2007). It seems therefore warranted to study multiple components of emotional intelligence and competence in victims simultaneously (Lansford et al., 2010; Lomas et al., 2012). This might help to further increase our understanding of victims’ interpersonal functioning.

Previous studies in victims mainly used non-immersive computer tasks showing static photographs of faces to assess emotion recognition (e.g., Woods et al., 2009; Ciucci et al., 2014), and self-report questionnaires to assess empathy (see van Noorden et al., 2016). Victims’ behaviors were often assessed through other-report (e.g., Perren and Alsaker, 2006; Lansford et al., 2010). In the present study, to better capture the dynamics and complexities of emotions and interpersonal situations (Fiorentini and Viviani, 2011; Hopwood et al., 2019), we examined aspects of emotional intelligence and competence using novel methodologies. Firstly, we employed a Virtual Reality (VR) emotion recognition task (ERT-VR). Compared to computer tasks showing morphed stimuli of gradually increasing emotional facial expressions on a screen (e.g., Pozzoli et al., 2017), VR offers increased ecological validity by creating a three-dimensional, immersive experience while keeping controlled laboratory conditions (Parsons, 2011; Gutiérrez-Maldonado et al., 2014). Secondly, in a separate empathic accuracy task (EAT), we asked participants to watch videos of people recounting real-life autobiographical events and rate how these targets felt. These ratings were then compared with the targets’ own ratings to create a measure of empathic accuracy (EA; considered a form of cognitive empathy) based on ecologically valid stimuli (cf. aan het Rot and Hogenelst, 2014). Finally, a third computer task was used to examine likely behavioral responses to facial emotions (aan het Rot et al., 2014). While this task was not designed with a focus on ecological validity and providing participants with an immersive experience, it does offer a controlled setting with set facial emotional expressions which enabled us to assess and compare potential interpersonal behaviors in social situations. In sum, by studying multiple components of emotional intelligence and competence using various and novel methods, we aimed to enhance knowledge about social-emotional skills of victims of bullying and to better understand victims’ interpersonal struggles.

The study protocol was positively reviewed by the Ethics Committee of the Department of Psychology at the University of Groningen.

Detailed information on descriptive statistics for participant characteristics, mental health variables, and the ERT-VR and EAT can be found in Table 1; for FERT see Figure 1. Pearson’s correlations between study variables can be found in Table 2.

In the present study, we examined aspects of emotional intelligence and emotional competence in teenagers with and without bullying victimization experiences. We found some subtle differences regarding emotion recognition accuracy when taking depression symptoms into account and no significant differences in empathic accuracy or behavioral responses to facial emotions. Thus, in contrast to what we expected, victims mostly showed similar social-emotional skills compared to non-involved individuals.

Victims have been reported to have interpersonal struggles (e.g., Ellis and Zarbatany, 2007). As a potential reason, some research suggests alterations in victims’ emotional intelligence and competence. For example, victims have been reported to be less accurate in recognizing emotions (e.g., Pozzoli et al., 2017). In line with this, we found that victims had more trouble recognizing emotions and specifically neutral faces compared to their non-involved peers. Angry-misattribution (i.e., labeling a neutral face as angry) appeared more frequent for victims, although group differences only approached significance. Nevertheless, overall, victims were statistically less accurate in recognizing neutral faces. This suggests that victimized individuals had a biased interpretation of social information. Specifically, they appeared to attribute hostile intentions to others. Such a hostile attribution bias could stem from adverse experiences with bullies that generalize to other situations. Victims’ increased social anxiety scores (compared to non-involved peers) also point toward them fearing or expecting to be negatively judged by others. Having such negative expectations and perceptions of others can influence the quality of social relationships (Guy et al., 2017) and has also been found to be associated with depression symptoms (Gadassi and Rafaeli, 2015; Belmans et al., 2019).

Of note, we did find these differences in emotion recognition between victims and non-involved individuals when considering underlying depression symptoms but not when we tested the association without taking depression symptoms into account. Victims in our sample, similar to previous research (e.g., Moore et al., 2017), had significantly more depression symptoms than non-involved individuals. Therefore, and due to research reporting depression symptoms to influence interpersonal skills such as emotion recognition (e.g., Dalili et al., 2015), we chose to include it in our final analyses. Our significant results suggest a negative relationship between bullying victimization and emotion recognition accuracy independent of underlying depression symptoms. The fact that, although pointing in the same direction, results did not reach significance when disregarding depression symptomology, could be explained by our small group sizes and therefore low statistical power. While there are no standard power analyses available for multilevel models, deducing from the Cohen’s ds of our significant effect (d=0.81) compared to the non-significant effects (ds<0.67), we assume that we had enough power to detect large but likely lacked power to detect medium and small effects. However, we would like to highlight that sample size might not be the only relevant factor to increase the detection of such associations. From a more theoretical point of view, a dose effect of victimization should also be considered. More specifically, apart from large effects between general victimization vs. no bullying experiences, there might be some smaller associations between (only) the more intense or more frequently victimized and emotional intelligence characteristics. This calls for not only a general increase in the sample size but also a more diverse representation of victimization occurrences.

Previous studies which have examined emotion recognition accuracy of victims had considerably more participants (e.g., Woods et al., 2009) and some also more frequently bullied victims (e.g., Ciucci et al., 2014) and reported significant differences between victims and non-involved individuals. However, there is also one large study (Guy et al., 2017; frequency of victimization not apparent) that did not find victims to recognize emotions differently than non-involved peers. This discrepancy in findings in previous and in the present study warrants more research to establish whether there are indeed differences in emotion recognition between victims and non-involved individuals and how depression symptoms influence this association.

Contemporary integrative interpersonal theory (Pincus, 2005; Pincus and Ansell, 2013) proposes that interpersonal situations are a dynamic interplay of perceptions, behaviors, and affect. Therefore, similar to what we hypothesized, one could expect that a biased perception of others (as indicated by a potential hostile attribution bias) would also lead to an adjustment of behavioral responses, such as behaving more submissively when thinking others want to do harm. However, victims in our sample did not show a differential response to negative emotions compared to non-involved peers, or at least we were not able to detect such associations (see discussion of statistical power above). Previous research suggests that perceiving negative situations as uncontrollable and unchangeable is associated with processing negative stimuli more internally by exhibiting so-called characterological self-blame (cf. Graham and Juvonen, 1998). Engaging in characterological self-blame is associated with interpersonal and internalizing problems. Specifically in victims, high tendencies for characterological self-blame have been suggested to not only partly explain re- victimization (Schacter et al., 2015) but also victims’ depression symptoms (Perren et al., 2013). Whether victims in our sample also exhibited characterological self-blame remains to be determined.

Our results suggest that victims have similar cognitive empathy skills as non-bullied individuals. This potentially implies that having been bullied is not associated with an altered understanding of how others feel, at least not when it is about happy or sad content as in our task. It is worth noting however, that the entire sample performed poorly on the EAT compared to previous studies, specifically when watching the negative/sad videos. Research describes adolescence as a period characterized by a maturation of social and interpersonal competencies (Crone and Dahl, 2012). Notably, concepts such as cognitive empathy and theory of mind are said to develop until adulthood (e.g., Blakemore, 2012). Therefore, empathic accuracy skills of teenagers in our sample were possibly not as far developed. The task itself could have also influenced our participants’ performances. Participants might have had problems relating to the targets, possibly due to their age (mainly mid-twenties, i.e., 23–26years and one target with the age of 62) or content of the autobiographical events (see Materials and Methods). Therefore, we are uncertain to what degree our (non-significant) findings could be explained by participant’s cognitive maturity or the stimuli of the task itself.

We cannot say to what degree maturation of social and interpersonal competencies of our participants might have also potentially impacted the performance on the other two tasks. Though, regarding the tasks themselves, we did choose novel methodologies to assess interpersonal skills. VR is considered an appealing tool among adolescents and enabled us to assess emotion recognition in a more ecological valid and immersive manner as compared to some previous studies (Gutiérrez-Maldonado et al., 2014). In the present study, 62% of participants rated the VR-task as fairly or quite realistic and 28% as a little realistic (the remaining 10% as not realistic). These numbers are comparable to another study which applied the same VR task and in which the majority of participants also indicated that VR characters and their facial expressions looked realistic (cf. Nijman et al., 2020), which supports ecological validity of the task. Regarding the FERT, we assessed potential behavioral responses to others’ emotions while keeping a controlled and comparable environment. However, we are aware that indicating likely behavior based on a static photograph does not necessarily very well represent actual behavioral responses in real-life interpersonal encounters. Nevertheless, the FERT offers the assessment of likely interpersonal behaviors in response to facial emotions as compared to assessing behaviors more generally, without taking situational context such as the other person’s emotional state into account.

Interpersonal situations are complex. Understanding them requires skills to interpret verbal and non-verbal cues which can differ per emotion (cf. Hall et al., 2000). Emotions included in our tasks were intentionally mainly negative; however, the type and amount of emotions differed between tasks. For example, the ERT-VR and FERT did not include sadness, while the EAT mainly consisted of negative videos with sad content, and fearful stimuli were only included in the ERT-VR. This is because we used pre-set tasks and did not adjust them. Therefore, the comparability between tasks and emotions regarding social-emotional competencies is somewhat reduced.

Research has reported gender differences regarding emotional intelligence and competence in victims (Ciucci et al., 2014; van Noorden et al., 2017). Similarly, type and severity of victimization can also influence social-emotional skills (e.g., Woods et al., 2009; van Noorden et al., 2016). Due to small group sizes, we did not run additional tests including gender or victimization specific characteristics to prevent an increased false positive rate through a large number of tests. However, we did statistically control for underlying variation due to participants’ depression symptoms, which have also been shown to effect interpersonal skills (e.g., Dalili et al., 2015). While future studies should increase the sample size to test for additional influential effects, the time span of the assessed victimization should also be considered. Possibly, the association between victimization and social-emotional competencies is different for more recent (i.e., in the past month) compared to victimization that potentially happened some time ago (i.e., during high school, as in our study).

While we did choose tasks which were at parts more ecologically valid and immersive than previous studies to examine social and emotional competencies in victims, these computerized tasks cannot fully represent actual real-life interpersonal situations. As interpersonal situations are complex and situation-specific (Hopwood et al., 2019), researchers have suggested studying interpersonal processes with ecologically valid approaches such as ecological momentary assessment methodology (Reis, 2014). Specifically, event-contingent recording of social interactions can be used to examine the link between interpersonal perceptions and behavior (Moskowitz and Sadikaj, 2014). Therefore, future research examining interpersonal processes of victims would potentially benefit from applying such methodologies.

We used novel methods to examine multiple aspects of emotional intelligence and emotional competence in victims of bullying and non-involved individuals and illustrated how these potentially contribute to interpersonal struggles of victims. Overall, the findings indicated that how victims perceive facial expressions, and how they potentially respond to facial expressions as well as their ability to understand how others feel were largely similar to non-involved individuals. Of note, as the present study only had sufficient power to detect large effects, generalization of our findings is limited and it cannot be ruled out that more subtle group differences remained undetected.

The data that support the findings of this study are available from the corresponding author, MF, upon reasonable request.

The studies involving human participants were reviewed and approved by Ethics Committee of the Faculty of Behavioural and Social Sciences at the University of Groningen (EC-BSS). Written informed consent from the participants’ legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.

MF conceived the study, participated in its design, coordinated and gathered data, performed statistical analyses, and drafted the manuscript. PJ participated in the design of the study and helped to draft the manuscript. WV provided parts of the measurement equipment, and helped to draft the manuscript. MR conceived the study, participated in its design, helped with statistical analyses and interpretation of the data, and drafted the manuscript. All authors contributed to the article and approved the submitted version.

The present study was funded by a general PhD-research-grant of the Faculty of Behavioral and Social Sciences, University of Groningen (no grant number available).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.