Cyber Bullying Final Paper

Testing Assumptions About Cyberbullying: Perceived Distress Associated With Acts of Conventional and Cyber Bullying

Sheri Bauman University of Arizona

Matthew L. Newman Arizona State University

Objective: Cyberbullying has received considerable attention, and experts have made several assumptions about this phenomenon. In particular, experts have speculated that the potential harm from cyberbullying is greater than that from conventional bullying, but this assumption has not been confirmed empirically. Method: In this study we tested this assumption by using a questionnaire with pairs of items describing similar experiences, one occurring in “traditional” ways and the other using digital technology. Respondents indicated the degree to which they would be upset by the incident on a scale from 1 (not at all upset) to 7 (extremely upset). Results: Findings from this study suggest that the distress associated with an incident of bullying is related to the nature of the bullying incident rather than the form. When comparing the parallel items, we discovered that although cyber- actions and conventional actions were significantly different for most pairs, the form that was more upsetting varied across items, providing further evidence that the form is not the distinguishing feature. Finally, we found significant gender differences on all subscales, with females reporting more distress than males. Conclusion: We close with a discussion of implications for both typologies of bullying and interventions designed to reduce bullying. Because cyberbullying may not be uniformly more harmful than other types of bullying, strategies to assist victims may be implemented with regard to the context and severity of the bullying, rather than its method of delivery.

Keywords: bullying, cyberbullying, factor analysis, distress

Cyberbullying— using information and communication technology (ICT) to inten- tionally harm a target by affecting his or her social status, relationships, and reputation— has garnered considerable attention from the popular media, largely in the form of reports of cases with extreme personal and/or legal consequences. Scholars have also begun to focus on this problem, but because the line of inquiry is so recent, there are variations in definitions, measures, and methodology that make it difficult to generalize across studies.

Experts have speculated on the consequences of being victimized by technology (e.g., Campbell, 2005), and a few empirical studies offer some clues. Hinduja and Patchin (2008) found that many victims of cyberbullying in their sample of online adolescents reported reactions such as frustration, anger, and sad- ness, although 35% indicated that they were not affected by their experience. Other re- searchers (Beran & Li, 2007; Cassidy, Jack- son, & Brown, 2009; Raskauskas & Stoltz, 2007; Tokunaga, 2010; Ybarra & Mitchell, 2004) revealed that victims reported fear of attending school, diminished concentration at school, disrupted school friendship, and even suicidal thoughts, as a response to cyberbul- lying. The relative psychosocial impact of cyberbullying compared with “traditional” bullying has been the subject of speculation, but no empirical studies to date have directly tested this question. The objective of the cur- rent study was to compare the degree of distress

This article was published Online First September 17, 2012. Sheri Bauman, College of Education, University of Ari-

zona; Matthew L. Newman, School of Social and Behav- ioral Sciences, Arizona State University.

We thank anonymous reviewers for comments on earlier drafts of this article.

Correspondence concerning this article should be ad- dressed to Sheri Bauman, P.O. Box 210069, College of Education, Tucson, AZ 85721-0069. E-mail: sherib@u .arizona.edu

Psychology of Violence © 2012 American Psychological Association 2013, Vol. 3, No. 1, 27–38 2152-0828/12/$12.00 DOI: 10.1037/a0029867

27

experienced by incidents of traditional and cy- berbullying that were similar in all respects other than the method of perpetration.

Various experts (Campbell, 2005; Hinduja & Patchin, 2009; Kowalski, Limber & Agatston, 2008; Slonje & Smith, 2008; Willard, 2007) have suggested that unique characteristics of cyberbullying magnify the potential harm to victims. The perception of anonymity may cre- ate an online disinhibition effect (Suler, 2004), which reduces the usual social sanctions against cruelty, and results in more hurtful comments. Hinduja and Patchin (2008) reported that 37% of teenage survey respondents admitted to say- ing things electronically they would not say in person. In addition to the potential for increased hurtfulness, being the recipient of an anony- mous attack may undermine the victim’s trust in others, because anyone (including friends) could be the attacker. Another characteristic that may increase the harm is the absence of time and space restrictions on the bully. Elec- tronic aggression can be perpetrated at any time from any place, denying the victim a safe sanc- tuary. Finally, the size of the audience can am- plify the humiliation. One well-known case (The “Star Wars Kid”; Star Wars Kid Files Lawsuit, 2003), involved posting (and eventu- ally “enhancing”) an embarrassing video not intended to be seen by anyone other than the person who made it. It was posted by others on YouTube and Internet forums, shared via e- mail, and viewed 900 million times (BBC News, 2006).

Although there has been a focus on K–12 in cyberbullying research, several studies have at- tempted to quantify the problem at the college level. Selwyn (2008) reported that 90% of 1,222 undergraduates in the United Kingdom ac- knowledged online misbehavior. In the United States, Finn (2004) found that 10%–15% of students at one university in the northeast expe- rienced repeated threatening, insulting, or ha- rassing electronic messages, with 7% receiving unwanted pornography. A study conducted at the University of Northern Iowa found that 34% of the (mostly White freshman female) students in the sample reported being victims of cyberbullying, with 19% of the 191 participants admitting to cyberbullying others and 64% re- porting they had observed incidents of cyber- bullying (Tegeler, 2010). Research by Allison Schenk at West Virginia University, with a

sample of 799 undergraduates (72% female), found that just under 9% of participants had been victimized by cyberbullying more than once; four of those indicated that they had made a suicide attempt. Those who had been victim- ized were more likely to report depression, anx- iety, and paranoia than those who were not (wvu today, 2011). The findings of yet another study of 439 college students (MacDonald & Roberts-Pittman, 2010) showed that about 22% had been cyberbullied, 9% had cyberbullied others, and 38% had observed cyberbullying. Englander (2007) reported that 8% of college students had been cyberbullied via instant mes- saging in college, and 3% admitted to being cyberbullies. Taken together, these studies sug- gest that the phenomenon of cyberbullying per- sists across all levels of schooling and into young adulthood.

Several typologies of conventional bullying have been widely adopted by social science researchers. A number of researchers have in- vestigated the differences between physical, verbal, and social/relational types of bullying/ victimization (e.g., Bauman & Del Rio, 2006; Craig, Henderson, & Murphy, 2000; Yoon & Kerber, 2003), whereas others researchers focus on the distinction between direct and indirect forms of bullying/victimization (e.g., Woods & Wolke, 2003). Differential responses by type of bullying have been detected in numerous stud- ies that have found that indirect bullying is more harmful than direct bullying (e.g., Baldry, 2004; Bauman, 2010; Bauman & Summers, 2009; Card, Stucky, Sawalani, & Little, 2008; Hawker & Boulton, 2000; Sharp, 1995; van der Wal, de Wit, & Hirasing, 2003). When cyberbullying first appeared on the radar, researchers assumed it was yet another distinct type of bullying. However, there is scant empirical evidence that cyberbullying is in fact a distinct construct from the more conventional forms (physical, verbal, and relational). Researchers often do not report psychometric properties of measures, but often analyze items about electronic forms of bully- ing as though they were separate constructs (see e.g., Wang, Nansel, & Ianotti, 2009).

Two previous studies have utilized a factor analytic approach in order to determine whether cyberbullying/victimization is a separate type. In both cases, items were added to an existing measure to reflect the technological context. Dempsey, Sulkowski, Nichols, and Storch

28 BAUMAN AND NEWMAN

(2009) added four items to the Revised Peer Experiences Questionnaire (Prinstein, Boergers, & Vernberg, 2001), which has overt and rela- tional bullying scales. The overt scale includes four items, three of which describe physical victimization and one of which involves being threatened with physical violence. The rela- tional items all reflect social exclusion. Three of the new cyber items refer to a technological action that “was mean or threatened me” and the fourth describes a situation in which a webpage was created that included “mean or embarrass- ing information and/or photos” about the vic- tim. Because none of the overt or relational items described “mean behavior” or the use of images, the fact that the scale was found to be a separate factor is not surprising. That is, the overt items all related to physical actions that had either occurred or been threatened, and re- lational items all described instances of social exclusion, whereas the cyber items described receiving communications (electronic, in these examples) that were mean or threatening (sim- ilar to verbal bullying, which was not included in the measure) or involved the creation of an embarrassing or mean webpage. Because the behaviors were so qualitatively different from those in the other scales—and because form of delivery was confounded with the level of de- tail—we cannot determine from this factor anal- ysis whether cyberbullying/victimization is re- ally a separate construct.

In a similar study conducted by Griezel, Cra- ven, Yeung, and Finger (2008), researchers added items to an existing scale (The Adoles- cent Peer Relations Instrument–Bully & Target; Parada, 2000) to reflect the technological envi- ronment. In this case, the existing measure had three subscales: Physical, Verbal, and Social. The six Verbal items involved name-calling, teasing, making rude comments, and the six Social items described social exclusion, ignor- ing, and persuading friends to reject the target. The result of a factor analysis revealed that the new items comprise two scales: Bully Visual and Bully Text. The five items in the Visual scale describe taking and/or sending embarrass- ing or hurtful video or images of the target, and the eight Text items include several items that closely parallel those in the Verbal scale, with the comments conveyed electronically. Some items, however, were different in important ways. For example, the Verbal item reads,

“Made jokes about a student,” whereas the Text item states, “Made nasty jokes about a student to my friends in an instant chat message [italics added].” The inclusion of the term “nasty” im- plies a more obviously harmful intent; a joke could be light-hearted and not hurtful. The word “nasty” is used in another item as well. One item refers to creating a fake profile on a web- site, and two items refer to using someone’s online accounts without permission, to which there is no parallel in other scales. As with the Dempsey et al. (2009) study, it is not clear whether the factors emerged because of the cyber-context or because the actions described are qualitatively different from those included in the other scales on the measure.

To avoid this problem, we developed a scale consisting of eight pairs of items, in which the actions are parallel but the context differs. For example, one item refers to showing naked pho- tos of the target to others at school, and the parallel item describes sending the photo to others using a cell phone. With this strategy, we hoped to avoid the possibility that factors would reflect different behaviors rather than the differ- ent environments (conventional or cyber) in which the actions were perpetrated.

Although self-report measures are widely used in bullying research, there are concerns about social desirability in responses. Research- ers debate whether global items are sufficient, whether definitions should be provided, and whether lists of behaviors are more useful for identifying offenders and targets. Researchers also prefer to have multiple informants when participants are to be classified into bullying status groups (bully, victim, bully/victim, by- stander). To avoid these problems, instead of reporting on their own experiences, we asked participants to imagine they were involved a variety of situations and to indicate how upset they would be in those situations. Vignettes (brief descriptions of scenarios that exemplify the concepts being studied) have been widely used in social science research (e.g., Brody, 1984; Crick, 1995; Flaskerud, 1979; Marsh, 1982; Mize & Ladd, 1988; Nelson & Crick, 1999; Slaby & Guerra, 1988) and in identifying emotional responses and attributions. The use of scenarios or vignettes in research has numerous advantages for bullying research (Bauman & Del Rio, 2006). In a study of the use of vignettes in qualitative research with children and young

29SPECIAL ISSUE: TESTING ASSUMPTIONS ABOUT CYBERBULLYING

people, this technique was found to be engaging for participants and allowed them to explore sensitive issues, such as sexual harassment and bullying, which may not be tapped using stan- dard methods (Barter & Renhold, 2000). Vi- gnettes allow for depersonalization (Schoen- berg & Ravdal, 2000), which is important for sensitive topics such as bullying and cyberbul- lying. Responding to hypothetical scenarios al- lows the respondent to remain safe from any personal threat. Poulou (2001) speculated that participants are less likely to be influenced by social desirability because they are not reveal- ing their own personal experiences. Poulou noted that not only do vignettes tend to engage participants’ interest and imagination, but this method also increases the internal validity of studies by increasing researchers’ control over variables.

Research on cyberbullying to date has been guided by the assumption that its conse- quences are worse than those of conventional bullying (e.g., see Campbell, 2005). In the current study, our goal was to test this as- sumption by assessing differences in distress caused by victimization via conventional and cyberbullying incidents. We examined tradi- tional-age college students’ perceptions of the relative distress caused by conventional ver- sus cyberbullying experiences and investi- gated variations in the subscale scores by gender. Thus, our hypothesis, based on as- sumptions in the extant literature, was that victimization by cyberbullying will be more upsetting to participants than comparable conventional victimization experiences.

In addition, we explored the association between prior victimization history and per- ceived distress resulting from both forms of bullying. Previous research has demonstrated biases in social information processing (cf. Crick & Dodge, 1994) among victims of bul- lying, such that victims are more prone to making hostile attributions about ambiguous b e h a v i o r s ( e . g . , C a m o d e c a , G o o s s e n s , Meerum Terwogt, & Schuengel, 2002; Salmivalli & Nieminen, 2002). For the pur- poses of the present study, prior victimization history might lead participants to interpret all of the bullying behaviors as more harmful, so it was included as a potential covariate in our hypothesis testing.

Method

Participants

Participants were 588 students (76% female) at a large southwestern university. Age ranged from 17 to 25, with a mean of 19.8 (SD � 1.41); participants included freshman through seniors. About 45% of participants were born in the city where the university is located, 10% were born elsewhere in the state, 35% were natives of other U.S. states, and 9% reported other na- tional origin. The majority of participants self- identified as White (66%), with the remainder being African American/Black (7%), Hispanic/ Latino (17%), Native American (4%), and Pa- cific Islander/Asians (1%). Five percent of par- ticipants indicated “other” ethnicity. This distri- bution mirrors that of the campus as a whole.

Procedure

Students in psychology classes at a large ur- ban southwestern university were recruited via an online subject pool system, which provided several options for fulfilling students’ course research requirement. Importantly, this survey was advertised as a study of “social attitudes” and deliberately did not mention bullying until the debriefing. Participants completed the mea- sures online, with each person taking about 15 min to do the entire survey. Informed consent was obtained at the beginning of the survey, and all procedures and measures were approved by the Institutional Review Board at the research- er’s university.

Measures

A questionnaire was developed by the au- thors for this study. In developing the items for the questionnaire, our goal was to capture a broad range of behaviors that (a) represented familiar forms of bullying for this age group and (b) could be delivered through either “conven- tional” or “cyber” means. A group of under- graduate students in the second author’s psy- chology lab reviewed a pilot version of the questionnaire, and changes in wording were made in response to their suggestions. The final scale contained eight pairs of items, each de- scribing an incident in which the participant was bullied. Eight items described an incident of conventional bullying, and eight described par-

30 BAUMAN AND NEWMAN

allel scenarios in which the same bullying be- havior was inflicted via communications tech- nology. The order of items was randomized in the online survey, so that the pairs of items did not appear together on the scale. The seven response options ranged from 1 (not at all up- set) to 7 (extremely upset), with only the end points labeled. (See the Appendix for the full text of the eight pairs of items.)

Historical victimization experiences were as- sessed by asking participants whether they were bullied during junior high, high school, and college, by choosing from: never; occasionally; or frequently at each time period. This measure has been used in previous studies and is reliable up to 6 weeks later (r � .81; Hamilton, New- man, Delville, & Delville, 2008). Approxi- mately 57% of the current sample reported never being victimized, whereas almost 3% in- dicated that were victimized at least occasion- ally at all three levels. Only one participant indicated that he or she had been victimized only at college, whereas 16% were victimized at both junior high and high school, 19% were victimized only in junior high, and 5% only in high school. There were no gender differences in these victimization histories; the results of a chi-squared analysis were not significant.

Analysis

Predictive Analysis Software (2010; PASW 19.0) was used for all analyses. Missing data were MCAR (missing completely at ran- dom) (Little’s MCAR test: �2 � 626.44, df � 598, p � .21). However, because the rate of missingness was less than 1% for all variables and the sample was relatively large, we did not impute missing values, but used listwise dele- tion for all analyses.

Results

Preliminary Analyses

If cyberbullying represents a consistently more harmful form of bullying, we would expect two factors to emerge on the scale: one capturing conventional victimization and the other capturing cyber-victimization. Be- cause we had this a priori theoretical basis for the factors, we conducted a Confirmatory Factor Analysis to test the fit of the data to

this model. The fit statistics were uniformly poor (e.g., normed fit index and non-normed fit index .59 and .52, respectively), indicating that this particular two-factor model was not a good fit to the data.

We then conducted a principal components analysis with 16 items of the questionnaire to explore the underlying structure. Given the poor fit of the Confirmatory Factor Analysis model, we used an oblique rotation for the analyses, allowing the components to correlate with one another. The Kaiser-Meyer-Olkin Measure of Sampling Adequacy was .86, considered “mer- itorious” (Kaiser, 1970), and the Bartlett’s Test of Sphericity was significant, indicating the data were suitable for this analysis. The scree test suggested a three- or four-component solution, Parallel Analysis (O’Connor, 2000) suggested a three-component solution, and Velicer’s Mini- mum Average Partial test indicated that the smallest number of factors present in the data was one. We considered these findings, along with interpretability of the structure and deter- mined that the three-component structure was most useful. The rotated component matrix (Di- rect Oblimin rotation with Kaiser normaliza- tion) is shown in Table 1. The total variance explained by the three components was 65%.

After an examination of the component load- ings, we named the Component 1 “General Vic- timization,” Component 2 “Explicit Visual,” and Component 3 “Name Calling.” We then created scale scores for each component by computing the mean score of scale items. Internal consistency coefficients for these variables were .90 for Gen- eral (10 items), .92 for Explicit Visual (two items), and .83 for Name Calling (four items). We used mean scores for each scale so that the scores on the scales would be comparable. The overall mean was 4.61 (SD � 1.51) for General Victimiza- tion, 6.17 (SD � 1.45) for Explicit Visual, and 3.37 (SD � 1.61) for Name Calling. A one- way repeated measures ANOVA was conducted to compare the distress scores on the three scales. There was a significant effect by scales, Wilks’ � � .28, F(2, 559) � 707.35, p � .0005, multi- variate partial eta squared � .72. Pairwise com- parisons indicated that each scale was signifi- cantly different from the other two. The most distressful was Explicit Visual, followed by Gen- eral Victimization, and lastly Name Calling.

31SPECIAL ISSUE: TESTING ASSUMPTIONS ABOUT CYBERBULLYING

Is Cyberbullying Worse?

To examine the perceived distress associ- ated with cyber- versus conventional bully- ing, we conducted paired t tests with each of the eight pairs of items, using a Bonferroni correction. The difference between the paired items was significant for all pairs except “Teacher’s Pet” and “Naked.” The mean score of the conventional item was signifi- cantly higher in three of the pairs (“cartoon,” “exclusion,” and “slut”), and the cyber ver- sion was higher in three pairs (“lice,” “video,” and “personal ad”). Table 2 presents these analyses. The effect sizes indicate that the

magnitude of these differences were in the “small” range.

Gender. We investigated gender differ- ences among participants’ distress scores on the subscales. The differences between males and females were significant on all subscales: Gen- eral, t � 5.42, df � 573, p � .0005, �2 � .05 (a small effect); Explicit Visual (t � 7.6, df � 171.35 [adjusted due to violation of Levene’s test for equality of variances], p � .0005, �2 � .09 (a moderate effect); and Name Calling, t � 5.88, df � 571, p � .0005, �2 � .06 (a small effect. For all scales, females had significantly higher distress scores than males. See Table 3 for means and standard deviations. We also

Table 1 Rotated Component Matrix for Two-Component Solution

Item Component 1

(General) Component 2

(Explicit visual) Component 3

(Name calling)

Lice (conventional) .82 .27 .49 Lice (cyber) .80 .31 .42 Video (conventional) .80 .21 .35 Cartoon (cyber) .79 .28 .48 Cartoon (conventional) .78 .36 .54 Video (cyber) .77 .23 .35 Personal ad (cyber) .69 .50 .19 Personal ad (conventional) .68 .52 .20 Unfriended online .68 .28 .38 Not invited (conventional) .59 .30 .32 Naked (conventional) .43 .89 .14 Naked b (cyber) .45 .87 .14 Teacher’s pet (conventional) .52 -.03 .86 Teacher’s pet (cyber) .52 .01 .86 Slut (cyber) .52 .54 .73 Slut (conventional) .49 .59 .69

Note. Bolded values show component on which item loads.

Table 2 Descriptive Data and Paired t Test Results for Individual Items

Items Conventional

mean (SD) Cyber

mean (SD) Paired t p �2

Cartoon 4.8 (1.89) 4.44 (1.94) 2.65 .008 .01 Lice 4.56 (1.98) 4.75 (1.91) �4.10 .0005� .03 Naked 6.14 (1.51) 6.19 (1.52) �1.26 .26 .003 Exclusion 4.97 (1.79) 4.80 (1.99) 2.37 .02 .01 Video 4.33 (1.97) 4.47 (1.91) �2.70 .007� .01 Teacher’s pet 2.65 (1.78) 2.70 (1.78) �1.28 .20 .002 Personal ad 4.85 (1.95) 5.0 (1.94) �3.77 .0001� .02 Slut 4.17 (2.145 3.98 (2.13) 3.57 .0005� .03

Note. N � 558. � Significant at a Bonferroni-adjusted criterion of p � .005.

32 BAUMAN AND NEWMAN

examined gender differences in ratings of the individual scenarios. In every case, females rated the scenarios as more upsetting. These differences were significant in all but three sce- narios: the conventional “video” scenario ( p � .064) and both forms of the “teacher’s pet” scenario ( p � .45).

Victimization history. First, we calcu- lated bivariate correlation coefficients among the total Victimization score (sum of responses to the three items) and mean scores on the three scales. We observed that Victimization was sig- nificantly, but weakly, correlated with only the General Victimization scale (r � .10, p � .01). All three scales, as expected, were significantly correlated with each other (see Table 4).

To examine this question further, we con- ducted t tests comparing the distress scores on the three scales by victimization history (yes or no). Those with a history of victimization (M � 4.88, SD � 1.28) reported significantly more distress on the General Victimization items (t � �3.26, df � 567.88, p � .001, �2 � .02) than participants who reported no such history (M � 4.50, SD � 1.52). This pattern also held for the Explicit Visual items: M (yes) � 6.33, SD � 1.30; M (no) � 6.04, SD � 1.56); t � �2.40, df � 576.05, p � .02, �2 � .01. However, victims and nonvictims did not differ in their ratings of the Name Calling items, p � .30.

Discussion

This study adds to the literature by directly addressing a key assumption among researchers in the field of cyberbullying. Using a question- naire created for this study, we found three forms of bullying based on ratings of perceived distress: general bullying, which leads to hu- miliation or embarrassment of the target; name calling; and bullying that uses explicitly sexual images. It is noteworthy that in every case, the pairs of items loaded on the same factor. That is, the cyber and conventional forms of individual incidents were seen as more similar than differ- ent.

This study is the first to directly address the differences in perceived distress among victims of conventional bullying and cyberbullying. The findings are important because (a) we found no overall differences in distress by form (con- ventional or cyber) of victimization, contrary to expert expectations and (b) principal compo- nents analysis identified a three-component structure that was based not on form of victim- ization, but on the nature of the incident. Sec- ond, when comparing the parallel items, we discovered that although cyber-actions and con- ventional actions were significantly different for most pairs, the form that was more upsetting varied with items, providing further evidence that the form is not the distinguishing feature.

Table 3 t Tests for Gender Differences on the Two Components

Subscale

Male

n

Female

n t df p �2M SD M SD

General 4.03 1.46 138 4.80 1.47 437 5.42 573 .0005 .05 Explicit image 5.20 1.87 139 6.47 1.14 442 7.61 171.35 .0005 .10 Name calling 2.69 1.41 138 3.59 1.61 435 5.88 571 .0005 .06

Note. df � degrees of freedom.

Table 4 Correlations Among Subscales

General victimization Explicit visual Name calling

General victimization 1 .471�� .648��

Explicit visual 1 .341��

Name calling 1

�� p � 0.01.

33SPECIAL ISSUE: TESTING ASSUMPTIONS ABOUT CYBERBULLYING

Finally, we found significant gender differences on all subscales, with females reporting more distress than males.

Because the three scales were strongly corre- lated (see Table 4), it is not surprising that there were secondary loadings on some scales. The personal ad (both cyber and traditional) items were strongest on the general victimization component, which might reflect the perception that this action was a kind of prank. The sec- ondary loading is on the Explicit Visual scale, which has a sexual connotation. This is consis- tent with the nature of the advertisement in the vignette. The other set of items (calling some- one a “slut” in person or online) also had a secondary loading on the Explicit Visual, which is consistent with the sexual connotation of that term. The primary loading was on the Name Calling factor. In both cases, it could be that the feature of these items that is more salient to most respondents reflected the primary loading, but the strong relationship to other scales re- flects the other aspect of the item.

Research Implications

Taken together, these findings have several implications. First, it appears that the emotional distress caused by victimization is a function of the nature of the specific incident, rather than the method of its delivery. This demonstrates that expert opinions and theories need to be tested empirically; what seems logical based on extant work may not find support in data. Sec- ond, considerable research has focused on the types of conventional bullying (physical, ver- bal, relational) and the differences in distress and psychological outcomes for victims of the different types. Echoing the findings of Varjas, Henrich, and Meyers (2009), who concluded that the distinctiveness of the types of bullying is questionable, our findings suggest that it may not be the type of bullying, per se, that explains the differences in emotional responses, but rather the context of the particular incident and the victim’s gender. Although Eslea (2010) noted that several studies reported that indirect or relational bullying is more strongly related to such psychosocial outcomes as depression, he observed that more recent work has found that type of bullying was not a factor in perceived severity.

Eslea’s (2010) study is of particular relevance here because it involved participants at both the university level and at the secondary level (where students were ages 11–15). He investi- gated the levels of distress from bullying (or imagined bullying) by conventional means. He reported that adult females found their victim- ization to be more distressing than males, but no gender difference was detected in the secondary school sample. Eslea also found that those who had not actually experienced being physically bullied found the imagined situations to be more distressing than those who had, but this pattern was reversed for indirect bullying (so- cial exclusion and rumor spreading)—those who had personal experience being indirectly bullied reported greater distress than those who had not. In the present study, participants with a history of being victimized rated most of the scenarios as more distressing than nonvictims. Note that most of the items in our scale would be characterized as indirect bullying, so our results are similar to those of Eslea. One possi- ble explanation for these findings is that a his- tory of victimization leads people to interpret all bullying behaviors as more hostile and therefore more harmful (cf. (e.g., Camodeca, Goossens, Meerum Terwogt, & Schuengel, 2002; Salmivalli & Nieminen, 2002).

In the present study, the items on the “Gen- eral” victimization scale range from ambiguous items with the potential to be seen as playful to deliberate actions to hurt or humiliate the vic- tim. The fact that these items cluster on a single factor suggests that victims perceive compara- ble amounts of distress, regardless of whether the intent of the bully is direct harm or more subtle humiliation.

The items perceived as most distressing of all, regardless of method, was the sharing of naked pictures (Explicit Visual) that the partic- ipant had provided to a romantic partner, who then sent them to others than the intended re- cipient. When we consider the elements of this type of victimization, there appear to be two prominent features: betrayal of trust (the victim- ized person was assured by someone with whom he or she was in a romantic relationship that the photo was for their personal use only) and humiliation (not only was the photo pub- licly viewed, but the fact that the victim had taken such a self-portrait is also public). It may be the case that such characteristics magnify the

34 BAUMAN AND NEWMAN

severity and therefore the distress. Smith et al. (2008) reported that adolescents in their study considered bullying that used images to have the most negative impact on the target. In addi- tion, it is important to note that this practice of sending nude or seminude photos of oneself via picture messaging, known as “sexting,” is not uncommon. A January 2009 national report re- vealed that 20% of teens and young adults said they engaged in sexting, and 11% of females ages 13–16 said they had done so (National Campaign to Prevent Teen and Unplanned Pregnancy, 2009).

Limitations

One potential limitation of the present study is that the sample was composed of college students, and some of the scenarios might be more common in middle or high school settings. However, be- cause all the students in the sample had been middle and high school students in the not-too- distant past and because bullying (including cy- berbullying) occurs in college, we believe the sample provided useful information on this impor- tant subject. Other researchers (e.g., Englander, 2011) asked college students about their high school experiences with bullying and cyberbully- ing. In addition, not inquiring about personal ex- periences, but hypothetical ones, in the question- naire also may prevent social desirability from influencing responses. Finally, it is worth noting that the scenarios used in this study represent single incidents of bullying. Bullying typically involves repeated harm done by more powerful peers. Thus, future research should examine rat- ings of perceived distress following more sus- tained acts of cyberbullying and conventional bul- lying. Future research should attempt to refine the questionnaire, as noted above, and to study this question in middle and high school students to see whether findings are similar to those in this study.

Our measure of previous victimization experi- ences showed weak association with participants’ ratings of perceived distress. However, it is worth noting that the victimization measure assessed global perceptions of having been bullied, rather than the specific form of bullying (cyber vs. con- ventional). Thus, another important question for future research is the relationship between specific past experiences and responses to specific acts of bullying. Is it more distressing to encounter the same type of victimization or to encounter a novel

form? Does a history of cyber or conventional victimization have differential effects on percep- tions of specific acts? Future studies could address these questions with a more fine-grained measure of victimization history. We suspect, however, that the cumulative impact of bullying on psycho- logical symptoms depends more on the frequency of its occurrence than on the method of its delivery.

Clinical and Policy Implications

To summarize, we demonstrated that cyber- bullying may not be uniformly more harmful than other types of bullying. This suggests that strategies to assist victims may be implemented with regard to the context and severity of the bullying, rather than its method of delivery. Interestingly, Salmivalli, Kärnä, and Poskiparta (2011) assessed the effects of a bullying inter- vention program that did not include cyberbul- lying and found that cyberbullying also de- creased after the intervention. This suggests that existing effective antibullying programs could be effective in reducing cyberbullying as well. We also found that bullying with sexual mate- rial, whether conventionally or by technological methods, is the most upsetting kind of incident to targets. Technology makes it easier to perpe- trate this kind of behavior, which may explain the perception that technology is the problem. Given that these incidents of sexual victimiza- tion have different connotations for men and women, prevention efforts may need to be structured in gender-specific ways for maxi- mum benefit. Prevention and intervention pro- grams should also make it clear that the harm inflicted with such actions is serious and can severely damage someone’s reputation. In our society, in which sexualized images and themes are ubiquitous, that is a difficult message to convey.

References

Anonymous. (2006, November 27). Star wars kid is top video. BBC News. Retrieved from http:// news.bbc.co.uk/2/hi/entertainment/6187554.stm

Baldry, A. C. (2004). The impact of direct and indi- rect bullying on the mental and physical health of Italian youngsters. Aggressive Behavior, 30, 343– 355. doi:10.1002/ab.20043

Barter, C., & Renhold, E. (2000). “I wanna tell you a story’: Exploring the application of vignettes in

35SPECIAL ISSUE: TESTING ASSUMPTIONS ABOUT CYBERBULLYING

qualitative research with children and young peo- ple. International Journal of Social Research Methodology, 3, 307–323. doi:10.1080/ 13645570050178594

Bauman, S. (2010). Cyberbullying in a rural interme- diate school: An exploratory study. Journal of Early Adolescence, 30, 803– 833. doi:10.1177/ 0272431609350927

Bauman, S., & Del Rio, A. (2006). Pre-service teach- ers’ response to bullying scenarios: Comparing physical, verbal, and relational bullying. Journal of Educational Psychology, 98, 219 –231. doi: 10.1037/0022-0663.98.1.219

Bauman, S., & Summers, J. (2009). Victimization and depression in Mexican American middle school students: Including acculturation as a vari- able of interest. Hispanic Journal for the Behav- ioral Sciences, 31, 515–535. doi:10/1177/ 0739986309346694

Beran, T., & Li, Q. (2007). The relationship between cyberbullying and school bullying. Journal of Stu- dent Wellbeing, 1, 15–33.

Brody, L. R. (1984). Sex and age variations in the quality and intensity of children’s emotional attri- butions to hypothetical situations. Sex Roles, 11, 51–59. doi:10.1007/BF00287440

Camodeca, M., Goossens, F. A., Meerum Terwogt, M., & Schuengel, C. (2002). Bullying and victim- ization among school-age children: Stability and links to proactive and reactive aggression. Social Development, 11, 332–345. doi:10.1111/1467- 9507.00203

Campbell, M. (2005). Cyber-bullying: An old prob- lem in a new guise? Australian Journal of Guid- ance and Counseling, 15, 68 –76. doi:10.1375/ ajgc.15.1.68

Card, N. A., Stucky, B. D., Sawalani, G. M., & Little, T. D. (2008). Direct and indirect aggression during childhood and adolescence: A meta-analytic review of gender differences, intercorrelations, and relations to maladjustment. Child Development, 1185–1229. doi:10.1111/j.1467-8624.2008.01184.x

Cassidy, W., Jackson, M., & Brown, K. N. (2009). Sticks and stones can break my bones, but how can pixels hurt me?: Students’ experiences with cyber- bullying. School Psychology International, 30, 383– 402. doi:10.1177/0143034309106948

Craig, W. M., Henderson, K., & Murphy, J. G. (2000). Prospective teachers’ attitudes toward bullying and victimization. School Psychology International, 21, 5–21. doi:10.1177/0143034300211001

Crick, N. R. (1995). Relational aggression: The role of intent attributions, feelings of distress, and provoca- tion type. Development and Psychopathology, 7, 313–322. doi:10.1017/S0954579400006520

Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social information-processing mechanisms in children’s social adjustment. Psy-

chological Bulletin, 115, 74 –101. doi:10.1037/ 0033-2909.115.1.74

Dempsey, A. G., Sulkowski, M. L., & Nichols, R. (2009). Differences between peer victimization in cyber and physical settings and associated psycho- social adjustment in early adolescence. Psychology in the Schools, 46, 962–972. doi:10.1002/ pits.20437

Englander, E. K. (2007). Understanding violence. Mahwah, NJ: Erlbaum.

Englander, E. K. (2011). MARC Freshman Study 2011: Bullying, cyberbullying, risk factors, and reporting. Retrieved from http://webhost.bridgew.edu/marc/ MARC%20Freshman%20Study%202011.pdf

Eslea, M. (2010). Direct and indirect bullying: Which is more distressing? In K. Österman (Ed.), Indirect and direct aggression (pp. 69 – 84). Frankfurt am Main, Germany: Peter Lang.

Finn, J. (2004). A survey of online harassment at a university campus. Journal of Interpersonal Vio- lence 19, 468 – 483.

Flaskerud, J. H. (1979). Use of vignettes to elicit responses toward broad concepts. Nursing Re- search, 28, 210 –212. doi:10.1097/00006199- 197907000-00004

Griezel, L., Craven, R. G., Yeung, A. S., & Finger, L. R. (2008, December). The development of a multi-dimensional measure of cyberbullying. Pa- per presented at the Australian Association for Research in Education conference, Brisbane, Australia.

Hamilton, L. D., Newman, M. L., Delville, C., & Delville, Y. (2008). Physiological stress response of young adults exposed to bullying during ado- lescence. Physiology & Behavior, 95, 617– 624. doi:10.1016/j.physbeh.2008.09.001

Hawker, D. S. J., & Boulton, M. J. (2000). Twenty years’ research on peer victimization and psycho- logical maladjustment: A meta-analytic review of cross-sectional studies. Journal of Child Psychol- ogy and Psychiatry, 41, 441– 455. doi:10.1111/ 1469-7610.00629

Hinduja, S., & Patchin, J. W. (2009). Bullying beyond the schoolyard: Preventing and responding to cy- berbullying. Thousand Oaks, CA: Corwin Press.

Kaiser, H. F. (1970). A second generation little jiffy. Psychometrika, 35, 401– 415. doi:10.1007/ BF02291817

Kowalski, R., Limber, S. P., & Agatston, P. W. (2008). Cyberbullying. Malden, MA: Blackwell.

MacDonald, C., & Roberts-Pittman, B. (2010). Cy- berbullying among college students: Prevalence and demographic differences. Procedia: Social and Behavioral Sciences, 9, 2003–2009.

Marsh, D. T. (1982). The development of interper- sonal problem solving among elementary school children. Journal of Genetic Psychology, 140, 107–118.

36 BAUMAN AND NEWMAN

Mize, J., & Ladd, G. W. (1988). Predicting pre- schoolers’ per behavior and status from their in- terpersonal strategies: A comparison of verbal and enactive responses to hypothethical social dilem- mas. Developmental Psychology, 24, 782–788. doi:10.1037/0012-1649.24.6.782

National Campaign to Prevent Teen and Unplanned Pregnancy. (2009). Sex and tech: Results from a survey of teens and young adults. Retrieved from http://www.thenationalcampaign.org/sextech/ PDF/SexTech_Summary.pdf

Nelson, D. A., & Crick, N. R. (1999). Rose-colored glasses: Examining the social information process- ing of prosocial young adolescents. Journal of Early Adolescence, 19, 17–38. doi:10.1177/ 0272431699019001002

O’Connor, B. P. (2000). SPSS and SAS program for determining the number of components using par- allel analysis and Velicer’s MAP test. Behavior Research Methods, Instruments, & Computers, 3, 396 – 402.

Parada, R. (2000). Adolescent Peer Relations In- strument: A theoretical and empirical basis for the measurement of participant roles in bullying and victimisation of adolescence: An interim test manual and a research monograph: A test manual. Publication Unit, Self-concept En- hancement and Learning Facilitation (SELF) Research Centre, University of Western Sydney, Australia.

Prinstein, M. J., Boergers, J., & Vernberg, E. M. (2001). Overt and relational aggression in adoles- cents: Social psychological adjustment of aggres- sors and victims. Journal of Clinical Child & Ad- olescent Psychology, 30, 479 – 491. doi:10.1207/ S15374424JCCP3004_05

Salmivalli, C., Kärnä, A., & Poskiparta, E. (2011). Counteracting bullying in Finland: The KiVa pro- gram and its effects on different forms of being bullied. International Journal of Behavioral De- velopment, 25, 405– 411. doi: 10.1177/ 0165025411407457

Salmivalli, C., & Nieminen, E. (2002). Proactive and reactive aggression among school bullies, victims, and bully-victims. Aggressive Behavior, 28, 30 – 44. doi:10.1002/ab.90004

Schoenberg, N. E., & Ravdal, H. (2000). Using vi- gnettes in awareness and attitudinal research. Jour- nal of Social Research Methodology, 3, 63–74. doi:10.1080/136455700294932

Selwyn, N. (2008). Developing the technological imagination: theorising the social shaping and con- sequences of new technologies. In S. Livingstone (Ed.), Theorising the benefits of new technology for youth: Controversies of learning and develop- ment (pp. 18 –29). Department of Education, Uni- versity of Oxford, Oxford, UK.

Sharp, S. (1995). How much does bullying hurt? The effects of bullying on the personal wellbeing and educational progress of secondary aged students. Educational and Child Psychology, 12, 81– 88.

Slaby, R. G., & Guerra, N. G. (1988). Cognitive mediators of aggression in adolescent offenders: 1. Assessment. Developmental Psychology, 24, 580 – 588. doi:10.1037/0012-1649.24.4.580

Slonje, R., & Smith, P. K. (2008). Cyberbullying: Another main type of bullying? Scandinavian Journal of Psychology, 49, 147–154. doi:10.1111/ j.1467-9450.2007.00611.x

Smith, P. K., Mahdavi, J., Carvalho, M., Fisher, S., Russell, S., & Tippett, N. (2008). Cyberbullying: It’s nature and impact in secondary school pupils. Journal of Child Psychology and Psychiatry, 49, 376 –385. doi:10.1111/j.1469-7610.2007.01846.x

Star wars kid files lawsuit. (2003, July 24, ). Wired. Retrieved from http://www.wired.com/culture/ lifestyle/news/2003/07/59757

Suler, J. (2004). The online disinhibition effect. Cy- berPsychology & Behavior, 7, 321–326.

Tegeler, C. (2010, August 8). Text harassment, cyberbul- lying a concern even for college students. Retrieved from http://wcfcourier.com/news/local/article_ 4e33f555-24bf-5154-af4d-f0aa734d46f8.html

Tokunaga, R. S. (2010). Following you home from school: A critical review and synthesis of research on cyberbullying victimization. Computers in Hu- man Behavior, 26, 277–287. doi:10.1016/ j.chb.2009.11.014

van der Wal, M. F., de Wit, C. A. M., & Hirasing, R. A. (2003). Psychosocial health among young victims and offenders of direct and indirect bully- ing. Pediatrics, 111, 1312–1317. doi:10.1542/ peds.111.6.1312

Varjas, K., Henrich, C. C., & Meyers, J. (2009). Urban middle school students’ perceptions of bul- lying, cyberbullying, and school safety. Journal of School Violence, 8, 159 –176. doi:10.1080/ 15388220802074165

Wang, J., Ianotti, R., & Nansel, T. R. (2009). School bullying among US adolescents: Physical, verbal, relational, and cyber. Journal of Adolescent Health, 45, 368 –375. doi:10.1016/j.jadohealth .2009.03.021

Willard, N. (2007). Cyberbullying and cyberthreats: Responding to the challenge of online social ag- gression, threats, and distress. Champaign, IL: Research Press.

Woods, S., & Wolke, D. (2003). Does the content of anti-bullying policies inform us about the preva- lence of direct and relational bullying behavior in primary schools? Educational Psychology, 23, 382– 401. doi:10.1080/0144341032000096265

wvu today. Researchers look at cyberbullying victim- ization among college students.(2011, March 30). Retrieved from http://wvutoday.wvu.edu/n/2011/

37SPECIAL ISSUE: TESTING ASSUMPTIONS ABOUT CYBERBULLYING

3/30/wvu-researchers-look-at-cyber-bullying- victimization-among-college-students

Ybarra, M. L., & Mitchell, K. J. (2004). Youth en- gaging in online harassment: Associations with caregiver-child relationships, Internet use, and per-

sonal characteristics. Journal of Adolescence, 27, 319 –336. doi:10.1016/j.adolescence.2004.03.007

Yoon, J. S., & Kerber, K. (2003). Bullying: Elemen- tary teachers’ attitudes and intervention strategies. Research in Education, 69, 27–34.

Appendix

Bullying Scenarios

Bullying Incidents

Harm Conventional version Cyber version

Cartoon You discover a cartoon making fun of you on the bulletin board in one of your classrooms.

You find a website that has a cartoon making fun of you on it.

Lice You were absent from school because you were sick. Someone started a rumor that you missed school because the nurse sent you home for having lice in your hair. When you come into the classroom, everyone starts scratching their heads and saying “Eewww, I got lice from you.”

Someone sends a text or e-mail to everyone that says you were absent from school because you were sent home by the nurse for having lice in your hair, and says that they got lice from you.

Naked Photo You sent your boyfriend or girlfriend a picture of you naked that they promised would be private. He/she showed it to a bunch of friends.

You sent your boyfriend or girlfriend a picture of you naked that they promised would be private. Then he/ she sent it to a bunch of friends.

Exclusion You find out that one of your friends is having a party this weekend and did not invite you.

You find out that several of your friends unfriended you on MySpace or Facebook.

Video You make a video of yourself in a silly costume, just for laughs, and you don’t expect anyone to see it. You realize it is missing after one of your friends visits you. You are pretty sure he/she will show it to some other friends.

You make a video of yourself in a silly costume, just for laughs, and you don’t expect anyone to see it. You realize it is missing after one of your friends visits you. You are pretty sure he/she will put it on YouTube.

Teacher’s Pet Someone writes “teacher’s pet” in your yearbook.

Someone puts a comment on your MySpace or Facebook page that says “Teacher’s pet.”

Personal Ad Someone puts an ad in the newspaper saying you are looking for love and giving your phone number. You start to get calls.

Someone put a personal ad on Craigslist saying you are looking for love and giving your phone number. You start to get calls.

Slut Being called a “slut” in front of a group of your friends.

Getting an e-mail or text message calling you a “slut.”

Received February 7, 2012 Revision received July 23, 2012

Accepted July 25, 2012 �

38 BAUMAN AND NEWMAN