الگوهای جنسیتی متناقض برای مزاحمت سایبری و زورگویی سنتی - تجزیه و تحلیل داده نوجوانان سوئدی

In the wake of the rapid development of modern IT technology, cyberspace bullying has emerged among adolescents. The aim of the present study was to examine gender differences among adolescents involved in traditional bullying and cyberbullying. Cross-sectional data from 2989 Swedish students aged 13–15 were analyzed using logistic regression analysis. The results show discrepant gender patterns of involvement in traditional bullying and cyberbullying. First, although there were only minimal gender differences among traditional victims, girls are more likely than boys to be cybervictims when occasional cyberbullying is used as a cut-off point. Second, whereas boys are more likely to be traditional bullies, girls are as likely as boys to be cyberbullies. In conclusion, compared to traditional bullying, girls are generally more involved in cyberbullying relative to boys. We discuss these results in the light of adolescents’ usage of computerized devices.

As a consequence of the global IT revolution, the use of and access to electronic devices for communication have become part of everyday life for most people. In Sweden, a country highly influenced by modern IT technology, almost all adolescents have access to the internet via mobile phones or computers (The Swedish Media Council., 2010). Following the rapid development of IT technology, the phenomenon of bullying in cyberspace has emerged among adolescents. The estimated prevalence rates for cyberbullying, as for traditional bullying, vary between countries (Beran and Li, 2007, Craig et al., 2009, Li, 2006, Molcho et al., 2009, Pornari and Wood, 2010, Raskauskas and Stoltz, 2007 and Wang et al., 2010). Although children and adolescents in Sweden report relatively low rates of cyberbullying compared to other countries (Beckman, Hagquist, & Hellström, 2012), it is still considered an important public health issue. 1.1. Gender differences Viewing Swedish adolescents’ internet habits and usage from a gender perspective, some obvious differences have been reported. For example, boys in Sweden (aged 9–16) play more games and watch more video clips on the internet, whereas girls are more active on social networking sites (apart from Facebook), chatting and blogging (The Swedish Media Council, 2010), and they use more sites where they can upload pictures for public display (Findahl, 2010). Instant messaging (MSN) is used by 85% of 11-year-old girls compared to 50% of the boys of the same age. Blogging is also more widely used by girls: about twice as many girls as boys blog and almost 70% read blogs, compared to 50% of boys (Findahl, 2010). International research has shown similar gender differences (Lucas & Sherry, 2004; Muscanell and Guadagno, 2012 and Pujazon-Zazik and Park, 2010). Gender patterns in traditional bullying have been evident over time. Boys are more likely than girls to engage in bullying, particularly direct physical bullying (Björkqvist et al., 1992, Crick et al., 2001 and Wang et al., 2009) as bullies or as bully/victims (i.e. individuals who are both a bully and a victim) ( Perren et al., 2010 and Solberg et al., 2007). In some studies, more boys than girls are victims of traditional bullying ( Olweus, 1993 and Solberg and Olweus, 2003). However, many do not report any gender differences (e.g., Perren et al., 2010, Ranta et al., 2009 and Solberg et al., 2007). Verbal bullying, on the other hand, seems to be just as common for boys as for girls, but girls more frequently engage in indirect, covert, relational bullying than boys do ( Björkqvist et al., 1992 and Crick et al., 2001). In contrast to research on traditional bullying, cyberbullying research shows inconsistent results regarding gender differences. For example, studies from the UK (Smith et al., 2008), the US (Wang et al., 2009) and Canada (Li, 2007) reported boys being overrepresented as cyberbullies. Another study from the UK showed no such gender differences (Smith et al., 2008). In the US, Kowalski and Limber (2007) and Wang et al. (2009) reported girls as more likely to be cybervictims. A study from Spain (Calvete, Orue, Estévez, Villardón, & Padilla, 2010) reported that girls are more often cybervictims and boys are more often cyberbullies. However, many studies do not report any particular gender differences either for cyberbullies or for cybervictims (e.g. Mishna et al., 2010, Patchin and Hinduja, 2006, Slonje and Smith, 2008 and Smith et al., 2008). To our knowledge, only two studies in Sweden have yet examined adolescents’ involvement in cyberbullying (Slonje and Smith, 2008 and Slonje et al., 2012), and their result showed no significant gender differences, either for victims or bullies. While the difference between direct and indirect bullying is rather distinct in traditional bullying, it is not clear-cut in cyberbullying research. Cyberbullying is itself sometimes generalized as a form of indirect bullying (Kowalski & Limber, 2007) or as a tool for social exclusion (Spears, Slee, Owens, & Johnson, 2009), but more sophisticated descriptions have also been outlined (Kowalski et al., 2008, Mason, 2008 and Vandebosch and Van Cleemput, 2009). Although cyberbullying is increasingly recognized, there are still unanswered questions concerning gender differences in cyberbullying and traditional bullying, especially regarding simultaneous analyses of mutually exclusive groups (victims, bullies and bully/victims) and types of bullying involvement (traditional bullying and cyberbullying). Including mutually exclusive groups in the analysis allows us to assess the real effect of different kinds of bullying involvement. This knowledge may help in the planning of preventive work and thus enhance students’ mental health in a wider perspective. Therefore, the current study will examine gender differences among adolescents involved in traditional bullying and cyberbullying using mutually exclusive bullying groups. 1.2. Definitions of traditional bullying and cyberbullying Bullying is conceptually part of the broader umbrella concept of peer victimization, which can include physical and verbal aggressive acts, exclusion from social activities, and threats of harm (Hawker & Boulton, 2000). Bullying is a narrower phenomenon, usually defined according to three criteria, namely, it is repeated, negative intentional actions arising from an imbalance in power and/or strength that can be social, physical or psychological ( Olweus, 1993). In contrast to traditional bullying, cyberbullying definitions are less consistent ( Kiriakidis and Kavoura, 2010 and Tokunaga, 2010), but most are based on Olweus, 1993 and Olweus, 1996a criteria (e.g. Smith et al., 2008). There is, however, an ongoing debate concerning the distinctive features of the two forms of bullying ( Kiriakidis and Kavoura, 2010 and Tokunaga, 2010) and it has been called into question whether the criteria for traditional bullying really apply to cyberbullying. For example, power imbalance is questioned because of the perpetrator’s opportunity to be at least virtually anonymous, for example when using someone else’s computer or e-mail account ( Dooley et al., 2009 and Hinduja and Patchin, 2008). While the repetition criterion is one of the aspects distinguishing bullying from aggressive behavior ( Olweus, 1993), the operationalization of pervasiveness in cyberbullying may be more in terms of the impact, irrespective of repetition ( Dooley et al., 2009). Bullies and victims can perceive the number of incidents differently and, although it might be an easy task to count the number of sent text messages, it is almost impossible to count visitor clicks on a web page containing a humiliating photo ( Slonje & Smith, 2008). Furthermore, a single cyber incident can spread rapidly via the internet, both in and outside the school ( Kiriakidis & Kavoura, 2010), and may be harder to escape from, since the victimization can take place even in the victim’s home ( Slonje & Smith, 2008). It has been argued that these aspects make cyberbullying a more severe phenomenon than traditional bullying ( Dooley et al., 2009 and Wang et al., 2010). 1.3. Socio-demographics It has been discussed in the literature whether socio-demographic factors such as family structure and ethnicity are associated with both being a bully and being victimized. For example, Nordhagen, Nielsen, Stigum, and Köhler (2005) reported an increased risk of being victimized, and Spriggs, Iannotti, Nansel, and Haynie (2007) found an increased risk of being a bully, among children and adolescents living in single-parent families. Graham and Juvonen (2001) discuss ethnicity in a school context, in terms of majority–minority status that could affect the balance of power and enhance perceptions of ‘us’ versus ‘them’. Carlerby, Viitasara, Knutsson, and Gillander Gådin (2012) also found Swedish adolescents with a foreign background to be more involved in bullying. Therefore, it is important to control for such factors. 1.4. What we know and where research is lacking Reviewing previous researchers’ methodology, there are many studies comparing traditional bullying and cyberbullying, but only a few studies report a clearly described procedure that distinguishes between bullies, victims and bully/victims and at the same time between traditional bullying and cyberbullying, which is critical to the assessment of the real effect of different kinds of bullying involvement. We found three studies making such distinctions that also examined gender differences between these groups (Gradinger et al., 2009, Gradinger et al., 2011 and O’Moore, 2012). Gradinger et al., 2009 and Gradinger et al., 2011 acknowledged the methodological issue with mixed groups and also studied combined groups, i.e. when bullies, victims and bully/victims engage in, or are exposed to, both traditional bullying and cyberbullying. In line with previous studies, O’Moore (2012) found boys to be more involved as bullies and bully/victims in traditional bullying, whereas there was no gender difference among traditional victims. Regarding cyberbullying, a larger proportion of girls reported being cybervictims and cyberbully/victims, but there was no gender difference in cyberbullying others. However, O’Moore only reported percentages and did not perform any statistical analysis. Similarly, Gradinger et al. (2011) reported boys to be more involved in physical bullying, both as victims and as bullies; girls were more involved as cybervictims (and victims of verbal bullying and social exclusion), but there were no gender differences in cyberbullying others. In contrast, Gradinger et al. (2009) reported boys outnumbering girls in cyberbullying others, but found no differences in cyber-victimization, and no difference in traditional bullying or victimization. In both studies by Gradinger et al., 2009 and Gradinger et al., 2011, gender analyses on bully/victims are lacking due to a paucity of observations. In conclusion, while some studies have used mutually exclusive groups and types in their analysis, there is still a lack of statistical analysis of gender differences for bully/victims that controls for possible confounders such as family structure and country of birth. Existing knowledge of gender differences in the use of social media sites does not clarify whether the comprehensive changes in use and availability of computerized technology have affected online bullying patterns as well. The scope of the internet is no different for boys than girls, and existing socialization processes may not play the same role in this context. According to Österman et al. (1998), it is important to distinguish between styles of aggression because, if they are neglected, gender-specific variations will go unnoticed. Therefore, the overall aim of the present study is to examine gender differences among adolescents involved in traditional bullying and cyberbullying using mutually exclusive bullying groups.

3. Results 3.1. Demographic characteristics Table 1 reports demographic characteristics of the participants. Table 1. Descriptive statistics of socio-demographic characteristics of participants (N = 2989) (percent). N (%) Sex Girl 1499 (50.1) Boy 1494 (49.9) Grade 7th 947 (31.4) 8th 1053 (35.0) 9th 1012 (33.6) Birth country – student Sweden 2885 (96.1) Other country 117 (3.9) Birth country – parents Both parents born in Sweden 2483 (83.2) 1 parent born in another country 272 (9.1) Both parents born in another country 231 (7.7) Family structure Live with other than parent/alone 48 (1.6) Shared living 624 (20.7) Single parent 442 (14.7) With both parents 1896 (63.0) Note: Percentage may not sum up to 100 since missing cases is included in the base. Table options Table 1 shows that 4% of the adolescents reported being born outside Sweden, 9% had one parent who was born outside Sweden and 8% reported both parents being born outside Sweden. Sixty-three percent of the adolescents were living with both of their parents, 15% lived in a single-parent home, 21% reported a shared living arrangement and nearly 2% were living alone or with someone other than their parents. Table 2a indicates the distribution of students owning their own mobile phone and having access to the internet at home, and Table 2b shows the distribution of number of computers in the home. Table 2a. Descriptive statistics of mobile phone ownership and access to internet at home, distributed by gender and grade (percent). Grade 7 8 9 Boys Girls Boys Girls Boys Girls n = 469 n = 465 n = 504 n = 536 n = 511 n = 485 % % % % % % Do you own your own mobile phone? Yes 97.2 98.9 96.8 99.1 97.7 99.4 No 2.8 1.1 3.2 0.9 2.3 0.6 n = 465 n = 464 n = 502 n = 536 n = 512 n = 488 % % % % % % Do you have access to the internet at home? Yes 97.8 97.8 97.2 99.2 97.6 99.0 No 2.2 2.2 2.8 0.7 2.3 1.0 Note: Percentage may not sum up to 100 due to missing cases. Table options Table 2b. Numbers of computers at home, distributed by gender and grade (percent). Grade 7 8 9 Boys Girls Boys Girls Boys Girls n = 468 n = 467 n = 507 n = 537 n = 514 n = 490 % % % % % % How many computers do your family own? None 1.0 0 0.8 0.4 0.5 1.0 1–2 49.0 51.0 42.0 47.1 40.5 50.0 3 or more 50.0 49.0 57.2 52.0 59.0 49.0 Note: Percentage may not sum up to 100 due to missing cases. Table options About 2% – slightly more boys than girls – answered that they do not have their own mobile phone or have access to the internet at home (Table 2a). Table 2b shows that nearly 1% of the adolescents had no computer at home. About half of the remaining students reported having ‘1–2’ computers at home and the other half reported ‘3 or more’. 3.2. Types and groups of bullying involvement The Venn diagram in Fig. 1 illustrates all unique types of bullying involvement using the cut-off point ‘once or twice’ or more often. Of the original 2989 students, 845 were involved in bullying behavior in some way. Full-size image (32 K) Fig. 1. 15 Unique categories (845 students), distribution and overlapping of students involved in bullying based on N = 2,989. The cut-off point is “once or twice” or more often. Bully-victim is abbreviated to b/v. The categories are: (A) traditional bully (29.2%); (B) traditional victim (19.5%); (C) cybervictim (8.8%); (D) cyberbully (4.5%). In ascending order; (E) traditional victim and cybervictim (7.5%); (F) traditional b/v and cyberbully (7.2%); (G) traditional b/v (6.5%); (H) traditional b/v and cyber b/v (5.1%); (I) cyber b/v (2.8%); (J) traditional b/v and cybervictim (2.1%); (K) traditional bully and cyber b/v (1.9%); (L) traditional bully and cybervictim (1.9%); (M) traditional victim and cyber b/v (1.4%); (N) traditional b/v and cyberbully (1.0%); (O) traditional victim and cyberbully (0.5%). Figure options Fig. 1 illustrates that the patterns of bullying involvement are complex and overlapping. The largest group consists of traditional bullies, which is almost six times as large as the group of cyberbullies. The group traditional victims is about twice the size of the cybervictims group. However, the group of victims of both traditional bullying and cyberbullying is of the same size as the group of bullies of both types. The traditional bully/victim (hereafter b/v) group is almost twice the size of the cyber b/v group. However, the group of both traditional b/v and cyber b/v is of the same proportion as traditional b/v, and only 1.4% reported being a combination of traditional victims and cyber b/v. The smallest group is both traditional victims and cyberbullies. Using a higher cut-off point (2–3 times a month or more) (not reported in any figure or table), the relationships between the groups change proportion and the prevalence rates decrease. The most notable change is that the relative proportion of traditional bullies decreases to about one third of the original size, and instead the group of traditional victims becomes the largest. Interestingly, the prevalence rates for the rather large cyber b/v group, as well as the group of both traditional bullies and cybervictims drops dramatically using the higher cut-off, because these groups almost entirely consist of occasionally victimized students. 3.3. Prevalence rates of mutually exclusive groups and types of bullying involvement with respect to gender Table 3 presents prevalence rates within bullying types and groups across gender. The total percentage does not add up to 100, since the table does not include all 15 unique categories. Table 3. The proportion of boys and girls in mutually exclusive types and groups of bullying involvement and noninvolved (percent). Groups and typesa Boys Girls n = 1489 n = 1488 % % Traditional victim 4.9 5.0 Traditional bully 10.0 5.8 Cybervictim 1.5 3.2 Cyberbully 1.1 1.3 Traditional victim and cybervictim 0.9 2.7 Traditional bully and cyberbully 2.8 1.3 Neither traditional victim nor cybervictimb 84.0 83.2 Neither traditional bully nor cyberbullyb 26.0 27.3 a The cut-off point for involvement in bullying is ‘once or twice or more often’. b Not mutually exclusive. Table options Table 3 shows that boys were overrepresented as traditional bullies, and as both traditional bullies and cyberbullies, compared to girls. Girls were overrepresented as cybervictims, and as both traditional victims and cybervictims. There is only a minimal gender difference regarding traditional victims. 3.4. Associations between gender and types of bullying involvement In Table 4a, lower and higher cut-off points are contrasted using multinomial logistic regression analyses for traditional victims, cybervictims, traditional bullies and cyberbullies respectively, with non-involved as a reference category. Grade, family structure, birth country of parents, and birth country of students were included as control variables. Table 4a. Multinomial logistic regression analysis for variables predicting involvement in traditional bullying and cyberbullying as victims, bullies based on higher and lower cut-off points c. Model 1a: N = 2255 traditional victim Model 2a: N = 2128 cybervictim Model 3a: N = 2473 traditional bully Model 4a: N = 2088 cyberbully B OR 95% CI B OR 95% CI B OR 95% CI B OR 95% CI 2–3 t/m or more often Girl −.420 .67 .36–1.20 .−035 .97 .34–2.68 −1.195 .3 .18–.52b −.724 .49 .09–2.71 Boy 0 1 0 1 0 1 0 1 Once/twice Girl −.022 .99 .67–1.42 .701 2.02 1.15–3.54b −.513 .6 .46–.78b −.110 1.12 .55–2.26 Boy 0 1 0 1 0 1 0 1 a Controls variables are grade, family structure, country of birth (coefficients not shown in the table). b Indicate significance at the 95% level (p < .05). c Noninvolved as reference category. Table options Table 4a shows no significant gender differences for traditional victims or cyberbullies. Girls were significantly less likely than boys to be traditional bullies, both at the lower cut-off point (OR 0.6) and at the higher cut-off point (OR 0.3). Furthermore, girls were significantly more likely than boys to be cybervictims at the lower cut-off point (OR 2.0), but not at the higher cut-off point (OR 1.0). Table 4b shows binary logistic regressions for traditional b/v and cyber b/v with non-involved as reference category. Due to low frequencies in the b/v groups, we did not conduct any multinomial regression analyses. Table 4b. Binary logistic regression analysis for variables predicting involvement as traditional bully/victims and cyberbully/victims based on the cut-off ‘one or twice or more often’c. Sex Model 5a: N = 2018 traditional bully/victim Model 6a: N = 1987 cyber bully/victim B OR 95% CI B OR 95% CI Girl −1.735 .18 .09–.36b 1.216 3.37 1.23–9.30b Boy 0 1 0 1 a Controls variables are grade, family structure, country of birth (coefficients not shown in the table). b Indicate significance at the 95% level (p < .05). c Noninvolved as reference category. Table options The results from the analyses show that girls were significantly less likely than boys to be traditional b/v (OR 0.18), and significantly more likely (OR 3.4) than boys to be cyber b/v.