Cyberbullying and Psychological Well-being in Young Adolescence: The Potential Protective Mediation Effects of Social Support from Family, Friends, and Teachers

International Journal of Environmental Research and Public Health Article Cyberbullying and Psychological Well-being in Young Adolescence: The Potential Protective Mediation Effects of Social Support from Family, Friends, and Teachers Karin Hellfeldt 1,*, Laura López-Romero 2and Henrik Andershed 1 1School of Law, Psychology and Social Work, Örebro University, SE-701 82 Örebro, Sweden; [email protected] 2Department of Clinical Psychology and Psychobiology, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain; laura.lopez.romer[email protected] *Correspondence: [email protected]; Tel.: +46-(0)-19301329 Received: 30 November 2019; Accepted: 17 December 2019; Published: 19 December 2019   Abstract: In the current study, we tested the relations between cyberbullying roles and several psychological well-being outcomes, as well as the potential mediation effect of perceived social support from family, friends, and teachers in school. This was investigated in a cross-sectional sample of 1707 young adolescents (47.5% girls, aged 10–13 years, self-reporting via a web questionnaire) attending community and private schools in a mid-sized municipality in Sweden. We concluded from our results that the Cyberbully-victim group has the highest levels of depressive symptoms, and the lowest of subjective well-being and family support. We also observed higher levels of anxiety symptoms in both the Cyber-victims and the Cyberbully-victims. Moreover, we conclude that some types of social support seem protective in the way that it mediates the relationship between cyberbullying and psychological well-being. More specifically, perceived social support from family and from teachers reduce the probability of depressive and anxiety symptoms, and higher levels of social support from the family increase the probability of higher levels of subjective well-being among youths being a victim of cyberbullying (i.e., cyber-victim) and being both a perpetrator and a victim of cyber bullying (i.e., cyberbully-victim). Potential implications for prevention strategies are discussed. Keywords: cyberbullying; adolescents; cyber-victim; cyberbully-victim; mental health; psychological well-being; social support; depression; anxiety; subjective well-being 1. Introduction The negative consequences of school bullying have been relatively well established within research [ 1 – 3 ]. However, due to technology development, bullying is not only restricted to the physical and real-life school context. Cyberbullying refers to an intentional act of aggression, carried out to harm another individual using electronic forms of contacts or devices [ 4 ]. Previous studies have rather consistently linked cyberbullying with several negative psychosocial well-being outcomes in adolescence (for review, see Reference [ 5 ]). However, much more research is needed to more clearly establish how and to what extent the distinctive cyberbullying roles (i.e., being a cyberbully, cyber-victim, and cyberbully-victim) are related with various psychological well-being outcomes. More research investigating if and to what extent various types of social support can mediate these relations is also needed. Social support from family, friends, and teachers has proven to be able to mitigate the negative impacts of traditional bullying but have been scarcely studied in relation to cyberbullying [ 6 – 11 ]. Therefore, the aim of the present study is to examine the relations between Int. J. Environ. Res. Public Health 2020,17, 45; doi:10.3390/ijerph17010045 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2020,17, 45 2 of 16 cyberbullying and its different roles with several psychological well-being outcomes. In addition, and in order to identify potential variables that may influence these associations, we examine if and to what extent perceived social support from the family, friends, and teachers mediate the association between cyberbullying roles and psychological well-being outcomes. 1.1. The Association between Cyberbullying and Adolescent Psychological Well-Being Some scholars suggest that cyberbullying is more stressful than traditional forms of school bullying [ 12 , 13 ]. Cyberbullying has emerged as a distinct form of bullying, with features such as publicity, permeability of online messages and pictures, anonymity of offender, and limitless boundaries, which distinguish it from traditional school bullying [ 4 ]. Hence, it is possible that cyberbullying renders other and more serious consequences as compared to traditional bullying. Although less studied than traditional bullying, involvement in cyberbullying has been linked to a range of psychological problems (for review, see Reference [ 5 ]). However, very few studies have compared psychological outcomes for the different cyberbullying roles. Youths may participate in cyberbullying either in the role of cyber victims, cyberbullies, or cyberbully-victims. Previous research has mostly focused on victims of cyberbullying and its potential consequences. A substantial amount of research has found associations between cyber-victimization and depressive symptoms ([ 13 – 16 ]; for reviews and meta-analyses, see Reference [ 17 ]). Fewer studies have studied other outcomes and found relations between cyber-victimization and anxiety symptoms [ 18 ] and between cyber-victimization and lower levels of subjective well-being [19,20]. Although less studied, research also indicates an association between cyberbullying (i.e., being in the role of the perpetrator of cyberbullying) and adverse outcomes. A majority of studies show that higher levels of cyberbullying relates to higher levels of depressive symptoms [ 14 , 18 ], anxiety symptoms [ 18 , 21 ], and lower levels of subjective well-being [ 19 ]. However, some studies indicate that cyberbullies might be better offthan those who are victimized, also with examples of studies finding no relation between the role of cyberbullying and depressive symptoms [15]. Some studies suggest that children and adolescents who are both victims and perpetrators of cyberbullying, (i.e., cyberbully-victims), constitutea distinctgroup withthehighestriskforpsychosocial problems, such as depressive and anxiety symptoms, as well as for lower levels of well-being in general [ 22 – 25 ]. Nevertheless, a limited amount of research has focused on the relationship between cyberbully-victims and different well-being outcomes. In sum, existing research suggests that children involved in any way in cyberbullying can be at increased risk for psychological distress, including depressive and anxiety symptoms, as well as lower subjective well-being. Existing studies are, however, not entirely consistent in these findings and more research is therefore clearly needed. 1.2. The Potential Positive and Protective Role of Social Support Very little is known from research concerning how to prevent or ameliorate the potential negative consequences of cyberbullying. One such potential protective factor that could mediate the relationship between cyberbullying and negative well-being could be social support. Existing studies indicate that different types of social support could potentially buffer against negative consequences of traditional, non-cyber, bullying [ 6 – 11 ]. The term social support often refers to different kinds of supportive social relations or interactions that can increase or promote an individuals’ well-being by acting as a buffering factor against negative outcomes [ 26 – 28 ]. Social support can be defined as including both an emotional dimension, (i.e., the individuals’ perception of being valued and cared for by others in their social network), as well as an instrumental dimension, (i.e., the individuals’ perception of having access to practical help with different tasks or obstacles in life) [ 29 , 30 ]. Theoretically, the potential benefits of social support can be understood in terms of the stress-buffering model [ 26 ], that is, that social support works as a buffer in stressful situations by serving as an important coping mechanism on which youths can draw [ 31 ]. Ongoing involvement in cyberbullying can be seen as a chronic stressor. Int. J. Environ. Res. Public Health 2020,17, 45 3 of 16 Hence, emotional and instrumental social support could serve as important resources when youths experience bullying, by both offering support when it occurs, but also to help stop bullying at an early stage [ 9 ]. For example, greater family support has been shown to be able to protect adolescents from being cyberbullied and cyber-victimized [ 32 ]. In addition, telling a friend about the bullying situation has been identified by adolescents themselves as the most helpful coping strategy when being cyberbullied [33]. Social support may derive from a number of sources and can thus be of various types. Among youths, the two primary sources or types of support seem to be parents and friends [ 7 ]. However, which of these two types of social support that serves as the primary resource seem to differ with age during youth. Younger children receive their primary support from parents but as the child approaches and enters adolescence, the role of parents can become less prominent, and the support from friends can increase and become more important [ 34 ]. Turning to a friend for support has been shown to be more commonly used by cyberbullied children, as compared to other types of support [ 35 ]. Regarding youths’ social network in the context of the school, teachers could be expected to play an important role in offering support [ 36 ]. Although cyberbullying generally takes place outside the school, it has been shown that most of the victims know their perpetrator from school [ 37 ]. Hence, teachers can potentially serve as an important support system when children experience cyberbullying. Importantly, seeking support from parents, friends, or teachers have been shown to be quite common strategies used by adolescents to cope with cyberbullying experiences [ 35 ]. Although there are several studies on different programs aimed at preventing bullying, few studies have focused on factors that may help youths to handle cyberbullying [ 38 ]. Social support from parents, friends, or teachers has been indicated to be able to mitigate the consequence of being a victim of traditional bullying but studies concerning this topic have yielded mixed findings [ 6 , 9 , 10 ]. In one study, moderate levels of peer support were shown to buffer against anxiety/depression among bullies and victims, as well as bully-victims [ 7 ]. In contrast, others found that support from friends and family protected cyber-victims from poor academic achievement but not from mental health difficulties [ 9 ]. Also, little research has explicitly investigated various types of social support in relation to cyber-victimization in general, and different roles of cyberbullying more specifically. An exception is a study including 765 Swiss seventh-graders, which examined if certain coping strategies could moderate the relation between cyber-victimization and depressive symptoms [ 39 ]. In this study, seeking support from friends and family showed a significant buffering effect on depressive symptoms. In addition, in a study including 1416 adolescents living in Cyprus, family support protected both cyber-victims and cyberbullies from being cyber-victimized one year later [ 32 ]. These previous studies offer promising indications that social support can protect adolescents involved in cyberbullying from negative consequences. However, little empirical work has been done in this area and more research is needed to understand how the various cyberbullying roles among youths can benefit from different types of social support. 1.3. The Present Study Even though more and more research has linked cyberbullying to different aspects of psychological distress, there is still a lack of studies focused on how distinctive cyberbullying roles (i.e., being a cyberbully, cyber-victim, and cyberbully-victim) are associated with various adverse outcomes [ 22 ]. Furthermore, few studies have examined processes that could mitigate the potential negative consequences of cyberbullying. Previous studies suggest that social support from family, friends, and teachers may mitigate the effect of traditional bullying [ 6 ] but this has not been thoroughly studied in relation to cyberbullying in general, nor more specifically in relation to distinctive cyberbullying roles. Hence, in the current study, we examine the relations between cyberbullying and its different roles with symptoms of depression, anxiety, and levels of subjective well-being. In addition, and in order to identify potential variables that may mediate these associations, we test whether and to what extent youths’ perceived social support from the family, friends, and teachers mediate the association between Int. J. Environ. Res. Public Health 2020,17, 45 4 of 16 cyberbullying roles and the three studied psychological well-being outcomes. This was examined using data from a cross-sectional study, in which 1707 youths aged 10–13 years in a mid-sized Swedish municipality responded to a questionnaire. The specific research questions we aimed to answer are: • How are different cyberbullying roles (i.e., being a cyberbully, cyber-victim, and cyberbully-victim) associated with depressive and anxiety symptoms, as well as with subjective well-being? • To what extent can youths ´ perceived social support from family, friends, and teachers mediate the association between various cyberbullying roles and depressive and anxiety symptoms, as well as to subjective well-being? In terms of hypotheses, previous research indicates a negative influence of cyberbullying on children’s psychological well-being [ 5 , 19 ]. Hence, we hypothesized higher levels of depressive and anxiety symptoms, and lower levels of subjective well-being for the cyber-victim and cyberbully-victim roles. We also expected protective mediation effects of social support in the way that social support from family, friends, and teachers will reduce the risk for depressive and anxiety symptoms and lower the levels of subjective well-being among those exposed to cyberbullying. We also controlled for gender since prior research indicates that seeking support may be more beneficial for girls compared to boys [ 9 ], and also that different sources of support might be of different importance for bullied girls and boys [27]. 2. Materials and Methods This study uses cross-sectional self-report data from the fifth wave of data collections in an ongoing prospective longitudinal study, the SOFIA-study (Social and Physical Development, Interventions and Adaptation). The SOFIA-study aims to provide better understanding of children’s behavior, social adjustment, and psychological and physical health. The target population for the SOFIA-study was all children born in 2005, 2006, and 2007, attending preschools during the spring of 2010 (2542 children) in a midsized (approximately 85,000 citizens) Swedish municipality. In terms of demographics of the municipality (i.e., gender, age, educational level, and employment, and the mix of urban and rural areas), the municipality is proportional to the rest of Sweden. Wave 5 of the SOFIA-study used in the current study was conducted in the year 2018 when the children were in their young adolescent age of 10–13 years. In the first data collection, the parents of 2121 children (85.7% of target population; 47% girls) gave active consent to participate. The SOFIA-study uses parents’, teachers’, and (pre)schoolteachers’ reports for all five waves of data collections. In the fifth data collection, the children were themselves the respondents for the first time. In Wave 5, 1707 children (approximately 80% of the original sample) completed the self-report questionnaire. The current study used these cross-sectional youth self-reports from the fifth wave of data collection of the SOFIA-study. 2.1. Participants The sample consisted of 1707 youths, which represent 80.4% of the original target sample (47.5% girls). The sample age ranged between 10 to 13 years (Mean age =11.89, Standard Deviation age = 0.86) and included 33.7% (n =576) children born in 2005, 33.7% (n =576) children born in 2006, and 32.5% ( n=554 ) children born in 2007. In addition, 16.7% of participants had at least one parent born in another country than Sweden. Regarding non-participants, caregivers declined participation in the study for about 16% of the target population in 2010. The non-participants did not differ significantly from participants regarding relevant aspects, such as the children’s levels of conduct problems and internalizing problems, or the caregivers’ socio-economic status and origin [40]. 2.2. Measurements Cyberbullying/Cyber-victimization. Two items from the Revised Olweus’ Bully/Victim Questionnaire (OBVQ, [ 41 ]) were used to assess cyberbullying. Each participant was introduced Int. J. Environ. Res. Public Health 2020,17, 45 5 of 16 with a detailed definition of bullying. The definition included three common criteria of bullying, i.e., intentionality, repetitiveness, and power imbalance between perpetrator(s) and a victim [ 41 , 42 ]. This definition was followed up by two cyberbullying questions, one on cyber-victimization, “How often have you been cyber-victimized during the past six months?”, and one on cyberbullying, “How often have you cyberbullied other students at school during the past six months?”. Items were rated as 1 (Never), 2 (1 or 2 times), 3 (2 or 3 times a month), 4 (Once a week), and 5 (Several times a week). The cyberbullying questions were preceded by the following definition: “Here are some questions about cyberbullying. When we say ‘cyberbullying’, we mean bullying through e-mail, instant messaging, in a chat room, on a website, or through a text message sent to a cell phone”. Following recommendations from previous studies, a cutoffpoint for frequent involvement in bullying is “two or three times a month”, and “only once or twice” for occasional involvement [43]. Depressive symptoms. To assess Depressive symptoms, items from The Youth Self-Report (YSR) was used, an instrument based on the Achenbach System of Empirically Based Assessment [ 44 ]. The subscale used in this study, i.e., Affective problems (used to measure depressive symptoms) reflects DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, 4th) problem dimensions, which comprise items that experienced psychiatrists and psychologists from 16 cultures have rated as being very consistent with DSM-IV diagnostic categories [ 45 ]. We included all items from the Affective problems subscale but one (i.e., “Thinks about suicide”), resulting in 11 items (e.g., “I am unhappy, sad, or depressed”; α =0.79; mean inter item correlation (MIC) =0.44). Participants rated each of the 11 included items on a three-point scale ranging from 1 (Not true) to 3 (Very true or often true). Respondents were requested to base their ratings on the preceding 6 months. Anxiety Symptoms. Anxiety symptoms were assessed with items from the Spence Children ´ s Anxiety Scale (SCAS [ 46 ]). The SCAS corresponds to DSM-IV anxiety disorder categories, and the scale has shown good psychometric properties with empirical support for test–retest reliability and internal consistency [ 47 ]. In the current study, we used the subscale for general anxiety, in total 6 items (e.g., “I worry that something awful will happen to me”; α =0.81, MIC =0.58). Participants rated each item on a four-point scale ranging from 1 (Never) to 4 (Always). Respondents were requested to base their ratings on the preceding 6 months. Subjective well-being. Subjective well-being was assessed using one item (“I enjoy life very much”), using a response scale ranging from 1 (Does not apply at all) to 4 (Applies very well). Respondents were requested to base their ratings on the preceding 6 months. Perceived social support. Social support was assessed with The Multidimensional Scale of Perceived Social Support [ 48 ]. This scale consists of 12 items intended to assess perceived social support from Family (four items; e.g., “My family is willing to help me making decisions”; α =0.81, MIC =0.63 ), Friends (four items, e.g., “I can talk about my problems with my friends”; α =0.87, MIC =0.72 ), and Teachers (four items; e.g., “My teachers help me solve problems in a good way”; α =0.91, MIC =0.80 ). Participants rated each item in a response scale ranging from 1 (Does not apply at all) to 4 (Applies very well). Respondents were requested to base their ratings on the preceding 6 months. 2.3. Procedure Initially, the decision makers in the Child and Adolescent Department at the municipality, decided on the participation of all municipal preschools. Private preschool principals were contacted separately. In 2010, at the first wave of data collection, all concerned preschool teachers received written information about the study, who in turn, passed on this information to parents who gave active informed consent. In the fifth wave of data collection carried out in the year of 2018 (used in current study), the youths answered a web-based questionnaire during school hours. They did not receive any compensation for their participation. The web-questionnaire took approximately 20 minutes to complete. Each school was responsible for administrating the questionnaire, i.e., giving the children the possibility to answer the questionnaire via a secure web-based questionnaire. The youths received information about the purpose of the study, that their participation was voluntary, etc. Int. J. Environ. Res. Public Health 2020,17, 45 6 of 16 The SOFIA-study has been evaluated by a regional ethics committee (First three waves 2010–2012; Dnr #2009/429. Fourth wave 2015; Dnr #2015/024. Fifth wave 2018; Dnr #2017/486). The study has followed all stipulated ethical research principles by the Swedish Research Council and the Swedish Ethics Authority. 2.4. Statistical Analyses First, descriptive statistics were computed for all study variables, with additional tests for differences due to gender and age performed through Student’s t-test and zero-order correlations, respectively. Second, correlations among study variables were explored with zero-order and partial correlations, controlling for gender and age. Third, cyberbullying role variables (i.e., No cyberbully/cyber-victim, Cyberbully, Cyber-victim, Cyberbully-victim) were created by computing low/high groups from the cyberbullying and cyber-victimization items. These groups were then compared on psychological well-being outcomes (i.e., depressive symptoms, anxiety symptoms, and subjective well-being) and perceived social support (i.e., social support from family, friends, and teachers) through analysis of variance (ANOVA) and including the Bonferroni correction for multiple comparisons (p<0.008). The strength of differences was assessed with the partial effect size statistic ( η2 ), and interpreted as small (>0.05), medium (0.06 to 0.14), and large (<0.14). Both correlation analyses and comparisons across groups were replicated using non-parametric analytic approaches (i.e., Spearman rho and Kruskal–Wallis tests, respectively), with no relevant changes in main results (further information is available upon request). Descriptive statistics, correlation analyses, and ANOVAs were computed in IBM SPSS 20 (IBM, Armonk, NY, USA). Finally, a series of mediation models were examined via structural equation modeling (SEM) in Mplus 7.4 (Muth é n & Muth é n, Los Angeles, CA, USA). The estimated models tested the effects of cyberbullying roles (i.e., Cyberbully, Cyber-victim, and Cyberbully-victim), coded as dummy variables (0 =No, 1 =Yes), on depressive and anxiety symptoms, and subjective well-being through the potential mediation effect of perceived family, friends, and teachers support, controlling for age and gender. The meanand variance-adjusted maximum likelihood test statistic (MLMV) was used as the estimator, since it is robust to non-normal data, and yields the best combination of accurate standard errors and Type I error [ 49 ]. Goodness-of-fit was assessed using the root-mean-square error of approximation (RMSEA), the comparative fit index (CFI), and the standardized root mean square residual (SRMR). According to suggestions by Hu and Bentler [ 50 ], RMSEA and SRMR values lower or equal to 0.06 and 0.05 respectively, and CFI values of 0.95 or higher are considered indicators of good model fit, whereas a RMSEA and SRMR smaller than 0.08, and CFI larger than 0.90 indicate adequate model fit. 3. Results 3.1. Descriptive Statistics and Correlations between Main Study Variables Descriptive statistics and gender comparisons are presented in Table 1. As expected given the characteristics of the sample, participants showed low levels of cyberbullying and cyber-victimization, depression, and anxiety symptoms, and high levels of subjective well-being, as well as high levels of perceived social support. Gender comparisons revealed differences across gender groups in cyberbullying, anxiety, subjective well-being, and peer support, with effect sizes ranging from low to moderate. Results showed higher levels of cyberbullying and subjective well-being among boys, and higher levels of anxiety and social support from friends among girls. Significant differences were also observed in terms of age for all study variables (r s =0.05 to − 0.26; p<0.05). More specifically, results revealed higher levels of cyberbullying and cyber-victimization, depressive, and anxiety symptoms among the older adolescents, whereas higher levels of subjective well-being and perceived social support were observed among the younger adolescents. Int. J. Environ. Res. Public Health 2020,17, 45 7 of 16 Table 1. Descriptive statistics for the main study variables with tests for gender differences. Total Sample Boys Girls Min–Max Mean (SD) Mean (SD) Mean (SD) t Cyberbullying 1.00–5.00 1.04 (0.28) 1.05 (0.35) 1.02 (0.16) 2.45 * Cyber-victimization 1.00–5.00 1.18 (0.57) 1.19 (0.60) 1.18 (0.53) 0.31 Depressive symptoms 1.00–3.00 1.36 (0.33) 1.36 (0.61) 1.35 (0.34) 0.07 Anxiety symptoms 1.00–4.00 1.73 (0.54) 1.63 (0.50) 1.84 (0.55) −8.53 *** Subjective well-being 1.00–4.00 3.45 (0.79) 3.51 (0.76) 3.38 (0.82) 3.38 *** Family support 1.00–4.00 3.57 (0.55) 3.58 (0.52) 3.55 (0.58) 1.38 Friends support 1.00–4.00 3.45 (0.64) 3.35 (0.66) 3.56 (0.60) −7.07 *** Teachers support 1.00–4.00 3.27 (0.74) 3.29 (0.74) 3.25 (0.75) 0.87 Note. Min =Minimum score; Max =Maximum score; SD =Standard deviation. * p<0.05. ** p<0.01. *** p<0.001. Zero-order correlation results between the main study variables are displayed in Table 2. All variables were significantly correlated with each other, except cyberbullying and anxiety symptoms. Results from partial correlations controlling for age and gender yielded basically the same results, with similar values for all the analyzed variables. The only larger difference was observed for the correlation between cyberbullying and friends’ support, which was no longer significant when controlling for age and gender (results available upon request). Table 2. Zero-order correlations between the main study variables. 12345678 1. Cyberbullying - 2. Cyber-victimization 0.30 *** - 3. Depressive symptoms 0.18 *** 0.30 *** - 4. Anxiety symptoms 0.04 0.23 *** 0.57 *** - 5. Subjective well-being −0.07 ** −0.16 *** −0.49 *** −0.33 *** - 6. Family support −0.14 *** −0.18 *** −0.45 *** −0.28 *** 0.40 *** - 7. Friends support −0.05 * −0.10 *** −0.31 *** −0.17 *** 0.23 *** 0.40 *** - 8. Teachers support −0.11 *** −0.20 *** −0.41 *** −0.28 *** 0.36 *** 0.55 *** 0.42 *** - *p<0.05. ** p<0.01. *** p<0.001. 3.2. Cyberbullying Roles: Comparisons across Groups on Psychological Well-Being In order to examine the associations between distinctive cyberbullying roles and psychological well-being, we identified groups low and high from the cyberbullying and cyber-victimization continuous items. Based on prior recommendations, those participants who reported being cyberbullying/bullied “two or three times a month” or more often [ 42 ] were classified as cyberbullies (n =11; 0.6% of the sample) and cyber-victims (n =55; 3.2%), respectively. By combining these groups, four mutually exclusive groups representing distinctive cyberbullying roles were identified: No Cyberbully/cyber-victim (n =1640; 96.5%), Cyberbully (n =5; 0.3%), Cyber-victim (n =49; 2.8%), and Cyberbully-victim (n =6; 0.4%). Given the low prevalence rates in the Cyberbully, Cyber-victim, and Cyberbully-victim groups, and in order to ensure enough participants within each group for subsequent analyses, new cyberbullying groups were created using a less stringent cut-off. The less strict cutofffollows recommendations from previous studies, resulting in groups with youths more occasionally involved in cyberbullying [ 43 ]. Hence, those participants who reported being cyberbullying/bullied “one or two times” or more often were classified as cyberbullies (n =42; 2.5%) and cyber-victims (n =218; 12.8%). The combination of these groups yielded four different groups representing the following cyberbullying roles: No Cyberbully/cyber-victim (n =1469; 86.4%), Cyberbully (n =13; 0.8%), Cyber-victim (n =191; 11.2%), and Cyberbully-victim (n =27; 1.6%). Int. J. Environ. Res. Public Health 2020,17, 45 8 of 16 Comparisons across groups revealed no differences in terms of gender, χ2 (3) =7.39; p=0.060, nor age, F=1.73; p=0.158 in the distribution across these groups. Results of comparisons between these groups showed significant differences on all three of the psychological well-being outcomes as well as on all types of perceived social support, even after applying the Bonferroni correction (see Table 3). More specifically, the Cyberbully-victim group showed the highest levels of depressive symptoms, and the lowest levels of subjective well-being, and family support. Higher levels of anxiety symptoms were observed for both the cyber-victim and the cyberbully-victim groups. There were no significant differences between the three cyberbullying roles (i.e., cyberbully, cyber-victim, and cyberbully-victim) in social support from friends and teachers. Effect sizes were small for all the analyzed variables except depressive symptoms, which showed a medium effect size. Table 3. Comparisons between cyberbullying roles on psychological well-being and social support variables. No Cyberbully/ Victim (n =1469) Cyberbully (n =13) Cyber-Victim (n =191) CyberbullyVictim (n =27) Mean (SD) Mean (SD) Mean (SD) Mean (SD) F η2 Depressive symptoms 1.32 (0.30)a1.50 (0.34)ab 1.55 (0.38)b1.86 (0.48)c57.62 * 0.09 Anxiety symptoms 1.68 (0.49)a1.58 (0.52)a2.06 (0.64)b2.11 (0.83)b35.32 * 0.06 Well-being 3.50 (0.76)c3.54 (0.78)bc 3.20 (0.85)b2.56 (1.01)a20.61 * 0.04 Family support 3.61 (0.51)c3.31 (0.85)bc 3.38 (0.68)b3.00 (0.80)a21.42 * 0.04 Friends support 3.47 (0.63)b3.21 (0.71)ab 3.30 (0.70)a3.36 (0.55)ab 4.82 * 0.01 Teachers support 3.32 (0.72)b3.16 (0.78)ab 2.93 (0.80)a2.58 (0.93)a25.16 * 0.04 Note. η2 =partial effect size statistic. Means with different subscripts (a, b, c) were significantly different (p<0.05) in post hoc pairwise comparisons (subscript a represents the lowest score/s in the analyzed variable). * Significant value after applying the Bonferroni correction (p<0.008). 3.3. Cyberbullying Roles and Psychological Well-Being: The Potential Mediational Role of Perceived Social Support In order to include cyberbullying roles as independent variables in the mediation models, three dummy variables were created: Cyberbully (No (0) =1687; Yes (1) =13), Cyber-victim ( No (0) =1509 ; Yes (1) =191), and Cyberbully-victim (No (0) =1673; Yes (1) =27). Three independent models were initially estimated, examining the effects of cyberbullying roles and social support variables on depression (RMSEA =0.04, CFI =0.93, SRMR =0.04), anxiety (RMSEA =0.04, CFI =0.95, SRMR =0.03 ), and subjective well-being (RMSEA =0.04, CFI =0.96, SRMR =0.03), respectively. Because similar results were observed across models, and considering the covariance between depressive symptoms, anxiety symptoms, and subjective well-being, the final model tested the effects of cyberbullying roles and perceived social support on all three psychological well-being outcomes simultaneously. As can be seen in Figure 1, model fit indices ranged from acceptable (CFI =0.92) to good ( RMSEA =0.03 ; SRMR =0.04). Regarding direct effects, both Cyber-victim and Cyberbully-victim roles showed a positive and significant association with depressive and anxiety symptoms and negative associations with subjective well-being. Similarly, the Cyber-victim group was significantly and negatively related with all perceived types of social support (i.e., from family, friends, and teachers), whereas the Cyberbully-victim showed negative significant associations with family and teachers support. Finally, family, friends’, and teachers’ support showed negative associations with both depressive and anxiety symptoms. Only family and teachers’ support showed a positive and significant association with subjective well-being. Int. J. Environ. Res. Public Health 2020,17, 45 9 of 16 Int. J. Environ. Res. Public Health 2019, 16, x FOR PEER REVIEW 9 of 16 Figure 1. Mediation model of cyberbullying roles on psychological well-being through perceived social support. Root-Mean-Square Error of Approximation (RMSEA) = 0.03; Comparative Fit Index (CFI) = 0.92; Standardize Root Square Residual (SRMR) = 0.04. The model shows the standardized estimates for direct effects, covariance between mediators and dependent variables, and controlled effects for age and gender (in grey). Only statistically significant relationships are shown in the figure. * p < 0.05. ** p < 0.01. *** p < 0.001. Figure 1. Mediation model of cyberbullying roles on psychological well-being through perceived social support. Root-Mean-Square Error of Approximation (RMSEA) =0.03; Comparative Fit Index (CFI) =0.92; Standardize Root Square Residual (SRMR) =0.04. The model shows the standardized estimates for direct effects, covariance between mediators and dependent variables, and controlled effects for age and gender (in grey). Only statistically significant relationships are shown in the figure. *p<0.05. ** p<0.01. *** p<0.001. Int. J. Environ. Res. Public Health 2020,17, 45 16 of 16 52. Berne, S.; Fris é n, A.; Schultze-Krumbholz, A.; Scheithauer, H.; Naruskov, K.; Luik, P.; Katzer, C.; Erentaite, R.; Zukauskiene, R. Cyberbullying assessment instruments: A systematic review. Aggress. Violent Behav. 2013,18, 320–334. [CrossRef] 53. Felix, E.D.; Sharkey, J.D.; Green, J.G.; Furlong, M.J.; Tanigawa, D. Getting precise and pragmatic about the assessment of bullying: The development of the California Bullying Victimization Scale. Aggress. Behav. 2011,37, 234–247. [CrossRef] 54. Thomas, H.J.; Connor, J.P.; Scott, J.G. Integrating traditional bullying and cyberbullying: Challenges of definition and measurement in adolescents—A review. Educ. Psychol. Rev. 2015,27, 135–152. [CrossRef] 55. Cornell, D.G.; Brockenbrough, K. Identification of bullies and victims: A comparison of methods. J. Sch. Violence 2004,3, 63–87. [CrossRef] 56. Aricak, T.; Siyahhan, S.; Uzunhasanoglu, A.; Saribeyoglu, S.; Ciplak, S.; Yilmaz, N.; Memmedov, C. Cyberbullying among Turkish adolescents. Cyberpsychol. Behav. 2008,11, 253–261. [CrossRef] [PubMed] 57. Hoff, D.L.; Mitchell, S.N. Cyberbullying: Causes, effects, and remedies. J. Educ. Adm. 2009 ,47, 652–665. [CrossRef] 58. Slonje, R.; Smith, P.K.; Fris é n, A. Perceived reasons for the negative impact of cyberbullying and traditional bullying. Eur. J. Dev. Psychol. 2017,14, 295–310. [CrossRef] 59. Hodges, E.V.; Perry, D.G. Personal and interpersonal antecedents and consequences of victimization by peers. J. Personal. Soc. Psychol. 1999,76, 677–685. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).