School Social Workers' Perceptions of Cyberbullying

Although cyberbullying is a growing concern among students, parents, and school personnel, there has been little research exploring school social workers (SSWs) at the elementary, middle, and high school levels about their perceptions of the seriousness and pervasiveness of this issue as well as their responses to it. Data for this study came from a survey of SSWs (N = 399) who were members of the 11-state Midwest School Social Work Council at the elementary, middle, and high school levels. Results indicate that SSWs at all levels believed that cyberbullying can cause psychological harm, including suicide, and should be addressed by SSWs. However, nearly half of respondents believed they were not equipped to deal with cyberbullying. Multivariate analysis of variance with post-hoc comparisons suggested significant differences at the school level in reports of seriousness and pervasiveness of cyberbullying. Middle school SSWs' reports of seriousness were significantly higher than those of elementary SSWs. Pervasiveness of cyberbullying was reported to be significantly lower at the elementary level than at middle and high school levels. Among SSWs at all three levels, there were no significant differences in perceptions of responses to this issue. Implications for school social work are discussed.

KEY WORDS: cyberbullying; exploratory research; Internet; Midwest; school social workers

Traditional bullying is the most common form of violence in schools and has been shown to disrupt the emotional and social development of youths (Raskauskas & Stoltz, 2007) . Types of bullying - including physical, verbal, relational, and indirect - and age and sex trends have been well established (Olweus, l993;Smith,Madsen,& Moody, 1 999). There has been considerable research on bullying prevention programs and scholarship on best-practice guidelines for school social workers (SSWs) (Whitted & Dupper,2005).

An emerging concern among the general public and researchers is the recent use of the Internet or other digital communication devices to intentionally harm others, generally referred to as "cyberbullying" (Smith et al.,2008;Willard, 2007). Other terms for cyberbullying include "online bullying" (Nansel et al., 2001), "online harassment" (Wolak, Mitchell, & Finkelhor, 2007), and "electronic bullying" (Raskauskas & Stoltz, 2007). The common element in all definitions is the use of electronic medium to cause harm. This type of bullying can include online fighting, harassment, stalking, impersonation, outing, trickery, exclusion, and more via the Internet, personal Web sites, e-mail, blogs, instant messaging, cell phones, and chat rooms (Willard, 2007).

Although the most severe consequences of cyberbullying have recently garnered national attention due to the highly publicized suicides of two middle school students, Megan Meier (http://www.meganmeierfoundation. org) and Ryan Patrick Halligan (http: //www. ryanpatrickhaUigan.org), researchers have only recendy started to identify the nature and impact of cyberbullying (Smith et al., 2008). Most research has focused on the characteristics of students as victims and perpetrators and as users of electronic media (Smith et al., 2008). Only a few empirical studies have focused on policy, parents, or school personnel (Chibbaro, 2007; Dehue, Bolman, & Völlink, 2008;Willard, 2007) . A search of the literature found no studies of SSWs and cyberbullying.

LITERATURE REVIEW

Cyberbullying has been estimated to have affected as many as 29 percent of youth Internet users (Patchin & Hinduja, 2006), with more conservative surveys suggesting that 9 percent have been subjected to online harassment (Wolak et al., 2006). A recent survey of British secondary school children ages 11 to 16 years reporced a lifetime incidence of being the victim of cyberbullying of 22 percent (Smith ec al, 2008). Although these estimates suggest that most youths have not been victims of cyberbullying, as Internet use and access to electronic media increases, it is likely that the percentage of youths affected by cyberbullying will also increase. However, research has suggested that as recently as a few years ago, few school professionals were aware of cyberbullying among students (B eran & Li, 2005).

Cyberbullying appears to be both similar to and distinct from traditional bullying. Some researchers have suggested that cyberbullying is a form of indirect traditional bullying (Dehue, Bolman,&Vö.Uink, 2008); others have suggested that it is a distinct class of bullying (Slonje & Smith, 2008). As with traditional bullying, victims of cyberbullying have reported negative physical, social, and psychological problems, such as extreme stress, being upset, fear (Finkelhor, Mitchell, & Wolak, 2000) , and depressive symptomology (Ybarra, 2004). Perpetrators of cyberbullying also appear to have similar negacive outcomes (Mason, 2008). In contrasc to traditional bullying, cyberbullying is often anonymous, can occur at anytime and without regard to geographical barriers, occurs outside of school, can involve potentially millions of people (via the rapid distribution of harmful messages and pictures through social networking sites like MySpace.com or video sites HkeYouTube.com), and is subject to limited monitoring by adults (Kowalski & Limber, 2007, Patchin Sc Hinduja, 2006; Slonje & Smith, 2008 iWilliams & Guerra, 2007). These unique characteristics suggest thac the information and craining SSWs have received abouc traditional bullying might not be sufficient to develop and engage in effective cyberbullying prevention and intervention programs.

Sex and age effeccs appear co be different for cyberbullying and traditional bullying. Traditionally, boys are more likely than girls to be involved in traditional physical bullying as both victims and perpetrators, but the picture is not the same for cyberbullying (Smith et al., 2008). One cyberbullying study reported chat although boys are slightly more likely to be perpetrators of cyberbullying than girls, girls reported significantly higher experiences of cyberbullying victimization than boys did (Li, 2007). Another sCudy suggesced that girls and boys have almost the same rate of recent cyberbullying perpetration (within the past 30 days), bue girls'lifetime report of cyberbullying perpetration was significance/ higher. In addition, this study reported that victimization of cyberbullying within the past 30 days was almost identical for girls and boys, with girls reporting a significantly higher lifetime rate of being the victim of cyberbullying (Hinduja & Pacchin,2010).Alchough more research Co delineace sex differences in cyberbullying may be needed, these preliminary studies demonstrate that variations may exist between the perpecration and victimization of cyberbullying and traditional bullying.

With regard to age and cyberbullying, the limited research presents an unclear picture. Some research suggests that targets of cyberbullying are more likely to be older than traditional bully victims (Wolak, Mitchell, & Finkelhor, 2007). However, other research suggests that Internet bullying peaks in middle school and then demonstrates a decline in high school, whether appreciable or not (Smith et al., 2008; Williams & Guerra, 2007;Worthen, 2007). Research on craditional bullying suggests that these age differences correspond with differences in staff perceptions by grade level in that both middle school students and staff tended to report the greatest exposure to and concern about bullying (Bradshaw, Sawyer, & O'Brennan, 2007) . Comparable research on school scaff perceptions and cyberbullying has noe been published.Whac one can tencatively conclude from these findings is chac, in general, cyberbullying appears Co become elevated around middle school and tends to hover around the same level during high school. Certainly more research is warranted to ferret out the subtleties of cyberbullying between middle and high school levels.

CYBERBULLYING AND SCHOOL POLICY

School staff members, including social workers, have an obligation co maincain a safe and healthy learning environment for students, while at the same time upholding laws and regulations that, although designed to protect students, have the unintended consequences of restricting staff members' ability to protect against cyberbullying. Because cyberbullying occurs through electronic media such äs cell phones and the Internet, federal privacy laws limit school staff members' ability to access records and monitor use. Mason (n.d.) summarized the current standing of the school's ability to legally intervene on cyberbullying by disciplining a student when it occurs within school or outside of school:

The First Amendment places restrictions on school officials when responding with formal disciplinary actions in situations involving online speech by students. Case law is limited and provides unclear guidance. The basic legal standard is that school officials can place educationally based restrictions on student speech that appears to be sponsored by the school or that is necessary to maintain an appropriate school climate. This standard probably applies to student speech through the district Internet system or via cell phones used at school. For off-çampus online speech, the courts have ruled that there must be a substantial and material threat of disruption on campus. But how this standard might be applied to severe off campus, online speech by one student. against another student is unknown, (¶ 4)

Thus, schools administrators are often confused about their legal role when addressing the effects of cyberbullying that occurs off school grounds, where it is most likely to occur (Blair, 2003). In school, social workers are left to make decisions about cyberbullying on the basis of nonexistent, unclear, or nonspecific policies. However, there has been no research into SSWs' perceptions about the effectiveness of school policy on cyberbullying.

PURPOSE OF THE CURRENT STUDY

Addressing violence in schools has become a priority since it has been demonstrated that violence can manifest itself in numerous ways and even low levels of violence can be detrimental to a successful learning environment (Whitted & Dupper, 2005). Although the task of identifying youths at risk for violence is complicated for a variety of reasons (Klein, 2002) i one role of the SSW is to assist educators in properly identifying and assessing risk factors so appropriate responses can be initiated (Franklin & Harris, 2007). Noted differences in the characteristics of cyberbullying and the possibility of different age patterns suggest that direct translation of bullying research and corresponding policies and programs might not be sufficient to inform cyberbullying prevention and intervention programs. By providing accurate and current information on school violence, SSWs can help develop and implement effective programs and policies (Astor, Marachi, &Benbenishty,2007).

The purpose of the current study was twofold. First, we aimed to increase practitioners' awareness of cyberbullying as aft important trend in school social work practice by documenting SSWs' perceptions of cyberbullying. Second, because prior research has suggested that the incidence of cyberbullying increases during middle school and perhaps remains in high school, we wanted to find out if SSWs across school levels perceived differences in cyberbullying seriousness and pervasiveness as well as their responses to the issue. Because SSWs work at all grade levels and because they assume a variety of roles in the school system (Whitted & Dupper, 2005), their perceptions on the issue are important as a basis for confirming existing research noted earlier and offering additional insight. Because prior cyberbullying research has focused on students, teachers, school counselors, and parents, but not SSWs, the current study sought to explore SSWs' perceptions and attitudes about cyberbullying, rather than test hypotheses.

METHOD

Sample

Data were collected over a three-month period (January through March 2008) from 399 SSWs from nine of the 1 1 member states of the Midwest School Social Work Council (MSSWC): Indiana, Iowa, Illinois, Kansas, Kentucky, Michigan, Ohio, Minnesota, and Nebraska.These data represented SSWs in suburban (49.7 percent), urban (26.6 percent), and rural (23.6 percent) schools. Because SSWs can potentially work at more than one level of school (elementary, middle, and high school), respondents were asked to choose only one level on which to base their answers.The final sample reflects SSWs responding from the elementary (37.8 percent), middle (32.3 percent), and high school (29.8 percent) levels. The majority of respondents reported their race as white (87.2 percent), followed by African American (8 percent) and other (4.8 percent). Most respondents were female (88.5 percent); they reported an average age of 45.4 years (SD = 11.2) and 12.8 years of practice experience (SD = 8.5). The subsample characteristics of the 178 participants used in our multivariate analysis did not differ significantly from those of the full sample.

Instrument

The survey instrument was three pages in length and contained three substantive sections (cyberbullying, suicide, diversity) and a demographics section. Because the three substantive sections were designed to be independent of each other, we report on the responses from the cyberbullying and demographics data. The survey was designed to capture SSWs' perceptions of cyberbullying, with a focus on their views of the seriousness and pervasiveness of it and their responses to this issue. Respondents were instructed at the beginning of the survey to base their answers on their past two years of practice experience to retrieve current responses and to create a boundary to standardize responses across participants. Cyberbullying was defined for respondents as "sending or posting harmful or cruel text or images using the Internet or other digital communication devices" (Willard, 2007, p. 1).

There were 16 questions related to SSWs' perceptions of cyberbullying. The first 12 questions queried the seriousness and pervasiveness of cyberbullying as well as SSWs' responses to it. Questions 13 to 15 asked specifically about online posts by students.The responses for questions 1 to 15 were captured on an ordinal fivepoint scale: 1 = strongly disagree, 2 = disagree, 3 = agree, 4 = strongly agree, and 5 = do not know. Question 16 asked respondents whether their district had a policy on cyberbullying and if it was effective. Demographic data included gender, race, age, years of school social work practice, grade level of their practice, and location of work (rural, suburban, urban).

Survey questions were developed through consultation with an expert in the field as well as a review of the literature. Fifty SSWs from the MSSWC were randomly selected to participate in a pilot study to refine the survey instrument. Through the pilot study, we were able to gauge how long the survey took to complete so we could approximate this for participants in the cover letter, determine if more categories of responses were needed, and assess the clarity of questions. The pilot study responses suggested that the survey would take no more than 10 minutes to complete. Respondents indicated that no additional response categories were needed. Through the pilot study, it became clear that many of the SSWs worked at more than one school, and this compelled us to add an important question clarifying which level of school they were using as a basis for responses. We also revised some cyberbullying questions for clarity, deleted ones that caused confusion, and added new questions in response to pilot study participant feedback.

Procedure

The MSSWC was approached in the summer of 2007 to provide the sample for this survey. The MSSWC is a consortium of 1 1 member state organizations aimed at supporting and promoting the professional development of school social work practice. There are approximately 3,500 members among these 1 1 states, and each scace was individually contacted to participate in the current study. Nine of the 1 1 states agreed to release their mailing list, for a total of 3,141 people.To reach a confidence level of 95 percent and a confidence interval of 5 percent, 342 usable surveys were needed for analysis. Although this indicates that approximately 1,000 surveys needed to be mailed to accrue this number given che estimated response rate, a random sample of 1 ,500 were sene che survey to account for mailing list flaws and errors in addresses. We received 399 responses, a response rate of 26.2 percent. The process of obtaining consent included a cover letter indicating the purpose of the study, its voluntary nature, and the anonymity of participation. Return of the completed survey implied consent.

Analyses

Because one goal of the analyses was to document descriptive data on all respondents' perceptions and attitudes toward cyberbullying on the 15 questions across school levels (elementary, middle, high), we used univariate analyses to explore characteristics of the entire sample. We then used two data analysis techniques to explore SSWs' perceptions about cyberbullying. First, we used a data reduction technique called principal component analysis (PCA) to identify chemes.We used PCA rather than principal axis factoring (PAF) because PCA is the preferred method for data reduction, whereas PAF is the preferred method for dececting data structure or causal modeling. We ran a PCA with listwise deletion because PCA is sensitive to missing data. We coded any "do not know" response as missing data because it is conceptually different from the disagree/agree responses, leaving 178 respondents for our analyses. A visual examination of the scree plot suggested that the 12 perception questions (1 to 12) clustered around three factors. We then created composite scores for each of the three faccors Co creace three new dependent variables for use in our multivariate analysis of variance (MANOVA).

The second data analysis procedure was a MANOVA with SPSS (version 16.0). We used the three factors identified in the PCA (seriousness, pervasiveness, and responses) as three dependent variables and looked at how SSWs' perceptions about these cyberbullying factors varied by school level (elementary, middle, or high school) of the respondent To identify differences between school level, we performed post-hoc comparisons using the Bonferroni adjustment for multiple comparisons to Control familywise alpha at .05.

RESULTS

To address both of our research goals, we first conducted univariate analyses to describe the perceptions of cyberbullying of the 399 SSWs who responded to our survey, and then we conducted multivariate analyses on a subsample of 178 SSWs to explore possible differences between grade levels.

Univariate Analyses

School Policy. Respondents were asked if their school had a policy on cyberbullying and if they thought that policy was effective. Respondents were evenly divided among"no" (34.2 percent), "yes" (32.9 percent), and "don't know" (32.9 percent). Among those who responded yes, 62 percent reported that their schools policy was effective, and 38 percenc reported that their school's policy was ineffective.

Respondents' Perceptions about Cyberbullying. As reported in Table 1 , SSWs varied in their perceptions about cyberbullying. SSWs most strongly agreed with statements 1 and 3 chat cyberbullying is problem at. our school(s) (46.7 percent agreed, 17.7 percent strongly agreed) and that cyberbullying occurs outside of school (37.9 percent agreed, 47 percent strongly agreed) . Other strong trends in SSWs' responses of agree and strongly agree were seen in statements that cyberbullying is an issue that merits increased attention, can cause psychological harm, is more harmful than traditional bullying, should be addressed by school social workers, and can contribute to suicide among students and in the two questions relating to posting material online that threatens other students and acts of self-harm.

SSWs provided mixed responses to three questions. For example, response to is being reported by students was almost spit between disagree (37.2 percent) and agree (38.5 percent). They were also only slightly more likely to report that they agreed (43.6 percent) rather than disagreed (38.3 percent) with is something I feel equipped to deal with.The trend was the same for the question asking SSWs if they beh eve students are posting material online that threatens teachers or other staff, with an almost even split between responses.

SSWs were more likely to disagree with statements that cyberbullying occurs during school (37.3 percent disagreed, 17.8 percent strongly disagreed), is actively addressed in my school (48,2 percent disagreed, 13.4 percent strongly disagreed), and occurs more than traditional bullying (32.5 percent disagreed, 3.3 percent strongly disagreed).

Multivariate Analyses

Respondents 'Perceptions of Cyberbullying across Grade Levels. Because research has suggested that the incidence and character of cyberbullying may change as students get older, we wanted to find out if SSWs' perceptions of cyberbullying varied by grade level. To test this hypothesis, we conducted a MANOVA with grade level (elementary, middle, high) as the independent variable on the three constructs created by the factor analysis (seriousness, pervasiveness, and responses).

Factor Analysis. A PCA using listwise deletion and orthogonal varimax rotation was performed on the 12 questions about SSWs perceptions about cyberbullying.This resulted in a subsample of 178 for analysis. A visual examination of the scree plot suggested that the 12 questions clustered around three themes, accounting for 56.8 percent of the variance, which we labeled seriousness (a = .59), pervasiveness (a = .78), and responses (a = .71). As shown in Table 2, the PCA suggested a simple structure, with only one cross-loading item (number 4) .We included item 4 with the second factor, because it fit conceptually and was the lowest loading item on the third factor.

MANOVA. Means and standard deviations for seriousness, pervasiveness, and response variables are provided inTable 3. MANOVA analyses with post-hoc comparisons, explained further in the next section, indicated that seriousness and pervasiveness, but not responses, varied signifi candy across school levels,

To examine school-level effects on seriousness, pervasiveness, and responses, we conducted a MANOVA with an alpha level of .05. A significant main effect was found for school level [Wilkss ? = .71 1, F(6, 332) = 10.30,p < .001]. The effect size was small (0.16). There were significant differences among school levels for seriousness |F(2, 175) = 3.586, p= .03, partial η^sup 2^ = .039] and pervasiveness [F(2, 175) = 31 .259, p < .001, partial η^sup 2^ = .263], but not responses [F(2, 175) - 2.595, p = .077, partial η^sup 2^ = .029]. Post-hoc comparisons with the Bonferroni adjustment indicated that SSWs at the middle school level (M = 3.28, SD = 0.34) were significantly more likely (p - .032) to believe in the seriousness of cyberbullying than SSWs at the elementary level (M = 3.12, SD = 0.44). They also indicated that SSWs at che middle school (M = 3.01, SD - 0.46) and high school (M = 3.05, SD = 0.49) levels were significantly more likely (p < .001) to believe in the pervasiveness of cyberbullying than were SSWs at the elementary school level (M = 2.39, SD = 0.56).

DISCUSSION

Cyberbullying is a phenomenon that can involve students at all grade levels and is therefore a pertinent issue for all SSWs. However, because of the recent nature of the phenomenon and the lack of research focusing on SSWs, there are almost no data on SSWs' perceptions of seriousness, pervasiveness, and responses to the issue. The current exploratory study provides a first step in this direction by documenting descriptive data on SSWs' perceptions of cyberbullying and determining if they varied by school level.

SSWs and Cyberbullying

Through univariate analyses, we found that nearly all SSWs agreed or strongly agreed that cyberbullying can cause psychological harm, including suicide, and that it is art issue that merits increased attention. Approximately half of SSWs (51.3 percent) agreed or strongly agreed that cyberbullying is more harmful than traditional bullying. Although 93 percent of SSWs in this study agreed or strongly agreed that cyberbullying should be addressed by SSWs, they were ambivalent about their ability to intervene. Just over half of the SSWs (54.8 percent) agreed or strongly agreed that cyberbullying was something they were equipped to deal with. In contrast, research on traditional bullying suggests that 86 percent of school staffare confident in their ability to deal with bullying (Bradshaw et al., 2007).

The high percentage of "do not know" responses to some questions indicated that SSWs were uncertain about specific aspects of the cyberbullying phenomenon. One in four SSWs reported that they did not know if students were posting materials online that appeared to threaten teachers or staff or suggest acts of selfharm. SSWs should be aware that students can post harmful statements about themselves and that prosuicide Web sites and chat rooms exist where visitors are encouraged to kill themselves (for example, alt.suicide.methods). The relationship between cyberbullying and traditional bullying appeared to be a source of uncertainty for respondents in the present study. When asked to compare traditional bullying with cyberbullying, one in three SSWs reported not knowing which occurred more often, and one in five reported not knowing which was more harmful. It is possible that SSWs' uncertainty simply reflects the fact that cyberbullying is a relatively recent phenomenon and respondents have had little personal experience with or training about cyberbullying. And SSWs may be less likely to be proactive in the prevention of and intervention regarding cyberbullying if they are uncertain about the danger or frequency of the phenomenon.

SSWs' Perceptions of Cyberbullying, by School Level

As reviewed earlier, previous studies have suggested that cyberbullying and perceptions of the phenomenon may be affected by school level. The results of our multivariate analyses suggested that there were significant differences between SSWs at the elementary school level and those at the middle and high school levels regarding perceptions of the seriousness and pervasiveness of cyberbullying.This is consistent with research that has reported a marked increase in students' use of technology and cyberbullying activity starting in middle school (Smith et al., 2008). We did not find a difference in SSWs' responses to cyberbullying.

Seriousness. In the present study, SSWs at the middle school level were significantly more likely than SSWs at the elementary or high school level to believe that cyberbullying was a serious problem. This is consistent with research suggesting that middle school students and staff report the greatest exposure to and concern about bullying (Bradshaw et al., 2007). Since cyberbullying is a significant problem for the first time in middle school, SSWs at that level may have a heightened perception of the seriousness. Although research suggests that the rates of cyberbullying level offin high school (Smith et al., 2008; Williams & Guerra, 2007; Worthen, 2007), SSWs might perceive it as slightly less serious, both because it has been present since middle school and because it may be relatively less serious than other behaviors that peak in high school, such as suicide, criminal activity, drug use, pregnancy, and dropping out.

Pervasiveness. SSWs' perceptions about the pervasiveness of cyberbullying in the present study also support prior research. Both middle and high school SSW reports of pervasiveness were significantly higher than those of elementary school SSWs, which again mirrors prior research documenting an increase in cyberbullying from elementary to middle and high school. First exposed to the problem in middle school, SSWs may experience an initial jump in cyberbullying behaviors that levels off through high school. Research has noted that social forms of aggression increase during the transition from childhood to adolescence (Craig & Pepler, 2003), which can include bullying via electronic media. An additional explanation for the increase in cyberbullying behaviors at this point may be that Internet use and access to electronic media increase when students reach middle school age.This rapid diffusion of technological media and electronic interfaces used for socializing among youths is further attested to by estimates from 2005 that 45 percent öf teenagers owned their own cell phone, and 87 percent of those ages 12 to 17 now use the Internet (Lenhart, Madden, & Hitlin, 2005). We can only surmise that these numbers are becoming greater each year. Lastly, students are reporting cyberbullying more often at the middle and high school levels, according to this study. Thus, SSWs may have a better grasp on the true incidence of the issue and be better able to report it at these school levels.

Responses. The current study found no differences between SSWs' responses to cyberbullying. It is possible that because cyberbullying is a relatively recent phenomenon, SSWs have liirrited training on, or experience with, cyberbullying and, therefore, might be unclear about how to respond. Bradshaw et al. (2007) reported that school staff members were less likely to follow up on student reports of traditional bullying if they did not witness the bullying activity. If staff are trained to respond to bullying on the basis of firsthand knowledge and are less willing to intervene on the basis of student reports, then cyberbullyings low profile (due to most events occurring outside of school and laws preventing monitoring of cell phone and Internet activity) would result in significantly less intervention, simply because staff were not seeing it themselves.

Furthermore, the SSWs in our study reported that their school's policies on cyberbullying were not helpful, and almost half reported not feeling equipped to handle cyberbullying. An alternate interpretation of these findings is that although middle and high school SSWs are more aware of the seriousness and pervasiveness of cyberbullying, responses to cyberbullying activity would not differ between grade levels. For example, SSWs' responses to threats over e-mail or via text might be the same, regardless of the perpetrator s grade.

Practice Implications

In addressing the issue of cyberbullying in the schools, Mason (2008) underscored the importance of prevention and intervention efforts cutting across ail systems. Thus, SSWs are poised to function as key personnel in addressing cyberbullying issues within the home-school-community network given their role within and across these systems. Addressing issues of cyberbullying in the schools is a difficult practice issue for SSWs. Although most respondents agreed or strongly agreed that cyberbullying is a problem that should be addressed by SSWs, they also indicated ambivalence about responding to cyberbullying. This ambivalence might reflect limited training (Hinduja & Patchin, 2009), lack of awareness that cyberbullying is occurring (Beran & Li, 2005), or confusion about the role of SSWs in preventing and intervening in cyberbullying because of nonexistent or unclear school policies or unclear legal mandates (Mason, 2008) . Traditional approaches to preventing and intervening with bullying might not be applicable to cyberbullying. For example, school staff members make decisions about the presence and severity of traditional bullying based almost entirely on seeing or hearing the bullying take place. However, cyberbullying takes place in a virtual world where school staff members are rarely present. Consequently, school staff members rely on student informants to become aware of cyberbullying activity. Almost half of the SSWs in our study thought that students were not reporting cyberbullying.

The current study suggests the need for trainings and clear practice guidelines that could improve SSWs' knowledge and skills in this arena. SSWs at all levels reported being unsure about how to respond and found their school's policies and guidelines lacking. This suggests that SSWs at all levels would benefit from trainings and clear practice guidelines. However, the findings that SSWs at the middle and high school levels perceived cyberbullying as more serious and pervasive compared with elementary school SSWs suggests that trainings should differ in content and approach on the basis of school level. Middle- and high school- level trainings might be more beneficial as stand-alone trainings rather than part of a more general "bullying training" that would include both traditional and cyberbullying, which might better fit with elementary programming. In turn, this knowledge will increase SSWs' ability, particularly at the middle and high school levels, to create a school atmosphere conducive to students' disclosure and the likelihood that youths will use adults as resources at all school levels. Recent research found that fewer than 9 percent of victims of online bullying in their study told an adult (Patchin & Hinduja, 2006; Smith et al., 2008). Although students may not be reporting cyberbullying, youths in one study indicated that speaking to someone would have made a difference for them (National Children's Home, 2005). Furthermore, Smith et al. (2008) found that 73.7 percent of youths believed that the best way to stop both traditional and cyberbullying was to tell an adult or a teacher. Clearly, students see adults as resources but also perceive barriers to accessing them.

Across all trainings on cyberbullying, Smith et al. (2008) recommended addressing basic content such as identifying and explaining the technologies that could be used for cyberbullying, because most school staff members are unfamiliar with them. These trainings should also cover confusing legal and policy issues as well as highlight existing intervention programs and other system-level interventions (Mason, 2008). Once SSWs are equipped with this knowledge, they have a responsibility to conduct trainings for teachers, staff, administrators, parents, and students on relevant issues. As one study suggested, educators do not feel comfortable with the topics and students receive little to no training on topics related to Internet issues such as cybersafety or cybersecurity (Pruitt-Mentle, 2008). Particular attention should be given to increasing awareness that the cyberbullying appears to markedly escalate from elementary to middle and high school.

Policy Implications

Although cyberbullying, like traditional bullying, is something that the courts and the public hold staff accountable for addressing (Mason, 2008), only one in five SSWs in this study believed that their school had an effective policy on cyberbullying. Research on traditional bullying has found that school staff members are less effective interveners when schools do not have supportive policies (Bradshaw et al., 2007). If the same is true for cyberbullying, the 80 percent of SSWs nom our study who are "flying solo" would be more effective interveners with appropriate policies. Because SSWs are on the frontline of addressing conflict and providing psychosocial intervention, they need to work with teachers and administrators to develop policies and establish norms regarding cyberbullying so that there are supportive and effective guidelines for both prevention and intervention activities (see Mason, 2008, for a sample school cyberbullying policy). In addition, SSWs, along with other school staff, need to be advised of current legal standings on their ability to intervene regarding cyberbullying that may have occurred outside of school but wreaks havoc within the school.

Research Implications

Although there may be a need to further study other variables that affect SSWs' perceptions of cyberbullying and their ability to respond, such as demographic and employment variables, future research should focus on the broader issue of cyberbullying. Particular attention should be geared toward documenting and discerning differences with respect to variables such as grade and gender. In addition, an important finding from this study is that SSWs were ambivalent about their ability to intervene. Therefore, future research should include the influence of training on school personnel, including SSWs. This type of research can provide information on how best to influence knowledge, attitudes, and willingness to intervene in cyberbullying issues. Research could also attend to barriers that prevent help-seeking behaviors among youths faced with cyberbullying. Lastly, exploring the role of cyberbullying policies in and outside of schools could offer information on how to best develop and implement approaches toward this issue.

Limitations

Although the present study provides insight into SSWs' perceptions of cyberbullying and highlights comparisons across school levels, there are important limitations to be noted. First, this study is not generalizable to all SSWs because the sample was limited to the MSSWC membership. Next, the lower response rate holds the possibility of nonrespondent bias, meaning the significant number that did not respond may hold different views than those who did respond. Lasdy, although these exploratory questions were based on prior studies and developed with expert consultation, they were developed specifically for this study, thus precluding comparative analysis with existing studies.

CONCLUSION

Although cyberbullying has quickly emerged as a new form of harm among youths, we are still in the initial stages of understanding the role of SSWs in this phenomenon. We hope that this study helps facilitate awareness of cyberbullying as an important issue for SSWs. It was clear across all levels of SSWs that the problem warrants increased attention, that it can cause psychological harm, and that it can contribute to suicides and should be addressed by SSWs.This study also found that school level of respondent (middle and high school) had a significant impact on how respondents perceive the problem in terms of seriousness and pervasiveness. In addition, reports of schools lacking a policy or lacking an effective cyberbullying policy speak to the importance of SSWs' work on the macro level. These findings support existing research on cyberbullying, provide a first look at SSWs and cyberbullying, and Bigblight areas for further training and research.

[Sidebar]

Cyberbullying was defined for respondents as "sending orposttng harmful or cruel text or images using the Internet or other digital communication devices. "

[Sidebar]

School social workers may be less likely to be proactive in the prevention of and intervention regarding cyberbullying if they are uncertain about the danger or frequency of the phenomenon.

[Reference]

REFERENCES

Astor, R., Marachi, R., & Benbenishty, R.. (2007). Violence in the schools. In E Allen-Meares (EdOi Social work services in schools (5th ed., pp. 145-181), Boston: Ailyn & Bacon.

Beran,T., " Li, Q. (2005). Cyber-harassment: A new method for an old behavior. Journal of Educational Computing Research, 32, 265-277.

Blair, J. (2003). New breed of bullies torment rheir peers on the Internet. Education Week, 22, 6.

Bradshaw, C. P., Sawyer, A. L., & O'Brerman, L. M. (2007). Bullying and peer victimization at school: Perceptual differences between students and school staff. School Psychology Review, 36, 361-382.

CbibbarO, J. S. (2007). School counselors and the cyberbully: Interventions and implications. Professional School Counseling, ? /(1), 65-68.

Craig, W., & Pepler, D. (2003). Identifying and targeting risk for involvement in bullying and victimization. Canadian Journal of Psychiatry, 48, 577-582.

Dehue, E, Bolman, C, & Voliink,T. (2008). Cyberbullying: Youngsters' experiences and parental perception. CyberPsycltology & Behavior, 1 1, 217-223. doi: I0.1089/cpb.2007,0008

Finkelhor, D., Mitchell. K., & Wolak.J. (2000). Online victimization: A report on the nation's youth. Retrieved from http://www.missingkids.com/en_US/ publications/N C62.pdf

Franklin, C, ¿fc Harris. M. (2007). The delivery of school social work services. In P. Allen-Meares (Ed.), Sodai work services in schoob (5th ed., pp. 317-360). Boston: Allyn & Bacon.

Hinduja, S., & Patchin.J.W. (2009). Bullying beyond the schoolyard: Preventing and responding to cyberbullying. Thousand Oaks, CA: Sage Publications.

Hinduja, S., & Patchin.j.W. (2010). Cyberbullying by gender. Retrieved from http://www.cyberbullying. us/research. php

Klein, J. (2002, March). School violence: Public and professional policies. NASlV News, p. 6.

Kowalski, R.. & Limber, S. (2007). Electronic bullying among middle school students.JoUrnäl of Adolescent Health, 41, S22-S3Q.

Lenhart. A.. Madden M., & Hitlin, P. (2005,JuIy 27). Teens and teàmology.Youth are leading the transition to afidly wired and mobile nation. Retrieved from http://www.pewinternet.arg/~/media/Files/ Reports/2005/PIP_Tcens Tech_July2005web.pdf. pdf

Li, Q. (2007). New bottle but old wine: A research of cyberbullying in schools. Computers in Hitman Behavior, 23, 1777-1791.

Mason, K. (2008). Cyberbullying: A preliminary assessment for the school personnel. Psychology in the Schools, 45, 323-348.

Mason, K. (n.d.). Cyberbullying: Legal issues. Retrieved from http:// www.cbasedprevention.org/ toolbox/ bullying/cyberbullying-legal-issues

Nansel.T-, Overpeck, M., Pilla, R., Rúan. J., SirnónsMorton, B., Si Scheüdt, R (2001). Bullying behaviors among US youth: Prevalence and association with psychosocial adjustnient JAMA, 285, 2094-2100.

National Children's Home. (2005). Putting U in the picture: Mobile bullying survey. Retrieved from http:// www.nch.org.uk/uploads/documents/Mobile_ bullying_%20report.pdf

Olweus, D. (t 993). Bullying at school: What we know and what we can do. Oxford, England: Blackwell.

Patchin, J., & Hinduja, S. (2006). Bullies move beyond the schoolyard: A preliminary look at cyberbullying; Youth Violence and Juvenile Justice, 4, 148-169.

Proitt-Mentle, D. (2008). 200» National Cybcreihics, Cybersafety, Cybersecurity Baseline Study. Retrieved from http://staysafeonline.mediaroom.com/index. php?s=67

Raskauskas.J., & Stoltz.A. (2007). Involvement in traditional and electronic bullying among adolescents. Developmental Psychology, 43, 564-575.

Slonje, R., & Smith, P. K. (2008). Cyberbullying: Another main type of buflying? Scandinavian Journal of Psychology, 49, 1 47-1 54. doi: 1 0. ? ? /j. 1 4679450.2007.00611.x

Smith, R K., Madsen, K., & Moody, J. (1999). What causes the age decline in reports of being bullied in school? Towards a developmental analysis of risks of being bullied. Educational Research, 41, 267-285,

Smith, P. K., Mandavi, J., Carvalho, M., Fisher, S., Russell, S., &Tippett, N. (2008). Cyberbullying: Its nature and impact in secondary school pupils. Journal of Child Psychology and Psychiatry, 49, 376-385. doi: 10.1 1 1 l/j.1469-76 10.2007.01 846.x

Whitted. K., & Dupper. D. (2005). Best practices for preventing or reducing bullying in schools. Children & Schools, 21, 167-175.

Willard, N, (2007). Educator's guide to cyberbullying and cyberthreats. Retrieved from http://cyberbuuy.org/ cyberbully/docs/cbcteducator.pdf

Williams, K., & Guerra.. N. (2007). Prevalence and predictors of Internet bullying._/onmii/ of Adolescent Health, 41, S14-S21.

Wolak.J., Mitchell, K., & Finkelhor, D. (2006). OmÍí'mí· victimization of youth: 5 years later. Retrieved from htro://www.unh.edu/ccrc/pdf/CVi 38.pdf

Wolak, j., Mitchell, K., & Finkelhor, D. (2007). Does online harassment constitute bullying? An exploration of online harassment by known peers and online-only contacts. Journal of Adolescent Health, 4 1, S51-S58.

Worthen, M. (2007). Education policy implications from the expert panel on electronic media and youth violence. Journal of Adolescent Health, 41, S61-S63.

Ybarra, M. (2004). Linkages between depressive symptomology and Internet harassment among young tegular Internet users. Cyber Psyciiology & Behavior, 7, 247-257.

[Author Affiliation]

Karen Slovak, PhD, MSSA, is associate professor, Department of Social Work, Ohio University Zanesville, 1425 Newark Road, Zanesville, OH; e-mail: Slavak@ohio.edu. Jonathan B. Singer, PhD, LCSWÏ is assistant professor, School of Social Work, Temple University, Philadelphia.

Original manuscript received August 29, 2008

Final revision received August 31, 2009

Accepted September 8, 2009