Computers in Human Behavior 49 (2015) 288–295

Contents lists available at ScienceDirect

Computers in Human Behavior

journal homepage: www.elsevier.com/locate/comphumbeh

The impact and response to electronic bullying and traditional bullying

among adolescents
Stacy Horner a, Yvonne Asher b, Gary D. Fireman b,⇑
a
Metrowest Neuropsychology, Westborough, MA, United States
b
Department of Psychology, Suffolk University, United States

a r t i c l e i n f o a b s t r a c t

Article history: With adolescents’ frequent use of social media, electronic bullying has emerged as a powerful platform
for peer victimization. The present two studies explore how adolescents perceive electronic vs. tradi-
tional bullying in emotional impact and strategic responses. In Study 1, 97 adolescents (mean
Keywords: age = 15) viewed hypothetical peer victimization scenarios, in parallel electronic and traditional forms,
Peer victimization with female characters experiencing indirect relational aggression and direct verbal aggression. In
Electronic bullying Study 2, 47 adolescents (mean age = 14) viewed the direct verbal aggression scenario from Study 1,
Adolescence
and a new scenario, involving male characters in the context of direct verbal aggression. Participants
were asked to imagine themselves as the victim in all scenarios and then rate their emotional reactions,
strategic responses, and goals for the outcome. Adolescents reported significant negative emotions and
disruptions in typical daily activities as the victim across divergent bullying scenarios. In both studies
few differences emerged when comparing electronic to traditional bullying, suggesting that online and
off-line bullying are subtypes of peer victimization. There were expected differences in strategic
responses that fit the medium of the bullying. Results also suggested that embarrassment is a common
and highly relevant negative experience in both indirect relational and direct verbal aggression among
adolescents.
Ó 2015 Published by Elsevier Ltd.

1. Introduction that involves an observed or perceived power imbalance and is

repeated multiple times or is highly likely to be repeated.
Concern about school based bullying has recently grown to Bullying may inflict harm or distress on the targeted youth
include peer aggression that occurs through social media. With including physical, psychological, social, or educational harm
the dramatic increase in adolescents’ use of social media and the (Gladden, Vivolo-Kantor, Hamburger, & Lumpkin, 2014, p. 7).
time spent in unsupervised interaction on social media, there is a
concurrent rise in opportunity for bullying within this medium. Bullying that occurs through social media popularly referred to
Prior to the rise of social media, bullying was typically concep- as ‘‘cyberbullying’’ is labeled ‘‘electronic bullying’’ by the CDC.
tualized as the result of an exchange between at least two people, Social media is an often-used form of interaction among adoles-
a bully and a victim, in direct face-to-face contact with each other. cents with 95% of 12–17 year olds reporting utilization of the
Given the scope of bullying and the potential damage to healthy Internet (Lenhart et al., 2011). It should be noted that in electronic
development, the Center for Disease Control and Prevention and bullying the unwanted aggressive behavior may be perceived by
the Department of Education have recognized it as a significant the victim as repeated not due to repeated acts of the bully but
public health concern. The CDC has provided a uniform definition rather, due to the enduring nature of electronic content with
to support effective scholarship, prevention and intervention: repeated viewings and potential for being shared widely. Given
Any unwanted aggressive behavior(s) by another youth or the distinct qualities of electronic bullying, questions arise regard-
group of youths who are not siblings or current dating partners ing the nature and intensity of emotional and strategic responses
of youth to electronic vs. traditional bullying.
Being a victim of traditional bullying is connected to a number
⇑ Corresponding author at: Department of Psychology, Suffolk University, Boston, of negative academic, social, and psychological consequences (e.g.,
MA 02114, United States. Nansel et al., 2001; Rigby & Slee, 1991). Similar negative conse-
E-mail address: gfireman@suffolk.edu (G.D. Fireman). quences have been reported through survey research for victims

http://dx.doi.org/10.1016/j.chb.2015.03.007
0747-5632/Ó 2015 Published by Elsevier Ltd.
 S. Horner et al. / Computers in Human Behavior 49 (2015) 288–295 289

of electronic bullying, including school avoidance, truancy, eating 2007). Youth were presented with 12 possible strategies to cope
disorders, depression, poor self-esteem, and suicidal ideation or with peer victimization (e.g., telling parents, telling other students,
suicide (e.g., Schneider, O’Donnell, Stueve, & Coulter, 2012). standing up to the bully, distraction, ignoring bullying), and asked
However, despite the tragic examples of electronic bullying to endorse which strategies they had utilized. The most frequent
reflected in the media, little experimental research distinguishing response, endorsed by approximately 50% of the respondents,
electronic bullying from traditional bullying has been completed was that they had tried to ignore the bullying.
(for exceptions, see Pieschl, Porsch, Kahl, & Klockenbusch, 2013; Smith et al. (2008) utilized large-scale surveys and focus groups
Sticca & Perren, 2013). The present research is one of a few initial to assess 11–16 year-olds’ beliefs regarding the best ways to stop
efforts to move beyond correlational and retrospective designs by electronic bullying. When provided with a list of possible ways to
providing adolescents with specific hypothetical scenarios in order cope with electronic bullying, the most commonly endorsed strate-
to examine how adolescents conceptualize and strategize about gies were to block messages/identities (75%), tell someone (63%),
bullying and electronic bullying. change email address/phone number (57%), and ignoring the behav-
The negative behaviors involved in bullying are often identified ior (41%). In traditional bullying prevention programs, telling an
as direct, including overtly aggressive behaviors such as physical adult is often a recommended strategy. However, there is evidence
acts (e.g., pushing, punching) and verbal acts (e.g., insults, teasing), to suggest that traditional victims are more likely to tell others than
or indirect, which includes relationally aggressive acts designed to victims of electronic bullying (Smith et al., 2008). In research
damage social relationships (e.g., spreading rumors, social exclu- examining who victims tell about electronic bullying, the most fre-
sion). Similar to traditional bullying, most electronic bullying quent responses are to tell ‘‘no one’’ or ‘‘friends’’, with parents and
requires an intentional act committed by an individual or group teachers reported far less often (Aricak et al., 2008; Dehue et al.,
in order to cause harm or distress to another individual or group 2008; National Children’s Home, 2005; Slonje & Smith, 2008). It is
but done through electronic communication technologies (Beran possible that adolescents are more willing to tell friends (as opposed
& Li, 2005; Mason, 2008). In one recent study, respondents were to parents or teachers) about both electronic and traditional bully-
asked about different types of electronic bullying they were ing, as friends are an important source of social and emotional sup-
exposed to (Aricak et al., 2008) with 19% reporting threats while port during this developmental period. In addition, a substantial
81% reported some form of embarrassment (e.g., teasing, insults, number of adolescents do not think school staff would or could do
rumors, or pictures displayed by others without consent). anything to stop electronic bullying (47%). These findings highlight
Although electronic bullying also requires a power differential the need to better understand if adolescents would respond differ-
between the bully and the victim (Mason, 2008), across online ently to comparable bullying situations across off-line and online
and off-line contexts the source of a bully’s strength and the rea- contexts in order to provide parents and professionals with guidance
sons why a victim feels defenseless may vary considerably. As on how best to help those being bullied.
opposed to traditional bullying in which bullies often rely on a
combination of attractiveness, local popularity, and physical
strength as a source of power, it has been hypothesized that in 2. The present studies
electronic bullying the power is based more exclusively on a bully’s
social connectedness and prestige all of which are visible on social The present two studies build on the existing research by
networking site profiles (Patchin & Hinduja, 2006). For adolescents, examining 9th and 10th grade adolescents’ perceptions of the
who are especially concerned about what others think of them and impact and ways to respond to electronic bullying vs. traditional
are more likely to believe others are always watching (Elkind, bullying. As mentioned above, the majority of research to date
1967), electronic bullying that interrupts relationships, can be seen has used self-report survey methodologies, which have notable
by many, and remains visible over time may be particularly upset- limitations (Raskauskas & Stoltz, 2007). The current research uses
ting. In the present research, the social media of Facebook was a design that allows adolescents to respond to specific hypothetical
approximated and the numbers of ‘‘liked’’, tagged photos, and scenarios one focused on a situation involving threat of overt
‘‘friends’’ were all considered with the bully appearing more popu- aggression (OA), and the other focused on a situation involving
lar than the victim. relational aggression (RA). These were chosen as prior research
As with traditional bullying, a number of negative consequences has suggested that threat of OA and RA (specifically, embarrass-
have been associated with electronic bullying. Self-report data indi- ment) situations are the most commonly experienced types of bul-
cate that adolescents who are victims of electronic bulling experi- lying for adolescents (e.g., Aricak et al., 2008; Huang & Chou, 2010;
ence a number of emotional consequences (e.g., feelings of Lenhart, 2007). Adolescents viewed the same scenarios in both tra-
sadness, depression, anxiety, fear), as well as behavioral difficulties ditional and online formats to make direct comparisons possible.
(e.g., missing school, difficulty concentrating, losing trust in peers, Since December 2007, Facebook has become one of the most
and falling grades; Beran & Li, 2005; Dehue, Bolman, & Völlink, frequently used social networking sites (Smith, 2009b). Because
2008; Raskauskas & Stoltz, 2007) that are similar to traditional bul- of Facebook’s popularity among adolescents, the present research
lying (Bauman & Newman, 2013). Utilizing a retrospective self-re- was designed to mimic this social networking site. The present
port methodology with college students, Bauman and Newman research is divided into two distinct studies: the first study exam-
(2013) found that perceived distress at a bullying situation was ined male and female adolescents’ responses to two scenarios, OA
not consistently related to the form of bullying (i.e., electronic or tra- and RA, with female characters only. The second study replicates
ditional), but rather related to the factors involved in the specific the OA scenario and adds the examination of male and female ado-
situation. Similarly, Sticca and Perren (2013) note that ‘‘cyber- lescents’ responses to an OA scenario with male characters (as
bullying is not a priori perceived as worse than traditional bullying’’ opposed to female characters). Analyses focus on adolescents’ rat-
(p. 739), but that certain factors (e.g., publicity of the bullying) deter- ings of their emotional reactions and problem solving strategies in
mine adolescents’ perceived level of distress. Therefore, the present response to the bullying scenarios. We hypothesize that the bully-
two studies examine adolescents’ emotional and strategic responses ing scenarios involving OA will elicit more negative emotion and
to bullying in distinct online and off-line scenarios. distinct strategies in traditional vs. electronic platforms. In con-
In a large self-report study, 1852 adolescents between the ages trast, with RA, we hypothesize a more negative emotional impact
of 4 and 19 who had been bullied participated in an online survey online. We also predict that males will be more distressed by the
about responses to traditional bullying (Craig, Pepler, & Blais, OA bullying scenarios and females by the RA scenarios.
290 S. Horner et al. / Computers in Human Behavior 49 (2015) 288–295

3. Study 1 bullying victims experience, respondents’ answers most frequently

involve being teased, the subject of rumors, threatened, pictures
3.1. Methods displayed by others without consent, physically hurt, intentionally
excluded, or ignored (Aricak et al., 2008; Patchin & Hinduja, 2006;
3.1.1. Participants Seals & Young, 2003). Of these categories of bullying, the scenarios
Adolescents were selected from a large public high school near constructed needed to be presentable in comparative formats
a major Northeast city. Adolescents were recruited through across mediums. Behavior that involved threat of OA and RA
announcements in their study hall classrooms and via email. A (specifically, teasing/embarrassing) was selected to allow for direct
total of 97 students, of approximately 450 eligible students, had comparisons on- and off-line.
parental consent, provided assent, and completed the study. All scenarios were taken from the popular media to ensure high
Adolescents’ responses were first examined to ensure they were production quality and relevancy of the content to adolescents. In
attending to the stimuli and identified the correct person as the the RA scenario, the off-line bullying example consisted of a video
victim of bullying. Eight students were not used in the following clip where a group of school-aged adolescents distribute an embar-
analysis (n = 89) because they failed to identify the correct victim. rassing picture of the victim during the school day. At lunch, they
Adolescents were asked to provide demographic information, sum- show the female victim’s photograph to her in front of her peers.
marized in Table 1. Her peers laugh at her, and she becomes visibly upset and runs
out of the cafeteria. The electronic bullying example consisted of
3.1.2. Development of stimuli a snapshot of a Facebook page where the bully posted an embar-
The selection of specific scenarios was based on two key factors: rassing picture of the female victim on the victim’s wall and many
representativeness of typical bullying experiences for adolescents people are commenting about it and making hurtful and insulting
and comparability between on- and off-line contexts. A review of statements.
relevant research revealed that when asked about the types of In the threat of OA scenario, the off-line bullying example
showed the female victim mildly flirting with the bully’s boyfriend.
Table 1 The victim then found the word ‘‘slut’’ written on her school locker.
Demographics. The bully confronted the victim and threatened to hit her after
school. The electronic bullying example showed snapshots of two
Characteristic Study 1 (%) Study 2 (%)
profiles and the wall-to-wall dialog between the two profiles. On
Age
the female victim’s profile, some flirtatious pictures of the victim
14 20.2 69.4
15 69.7 30.6
and the bully’s boyfriend were included. The wall posts between
16 9.1 0 the bully and victim showed the bully calling the victim a slut
17 1.1 0 and threatening to beat her up the next day.
Gender
Female 50.6 63.8 3.1.3. Procedures
Male 49.4 36.2 During the study, adolescents were seated in front of a com-
Grade puter with headphones. They were directed to a webpage that
9th 56.2 100 briefly described the study and asked demographic questions.
10th 43.8 0
Each adolescent viewed the two bullying conditions (threat of OA
Ethnicity and RA) in both off-line and online forms. Presentation of the four
Caucasian/White 73 83.3
bullying scenarios was counterbalanced. After each scenario was
African-American/Black 5.6 2.8
Asian 13.5 8.3 presented, the adolescents were first asked to identify who was
Other 5.6 5.6 the victim. This served as a manipulation check, and ensured that
When use SNS the adolescent was paying attention to the stimuli before answer-
Before school 21.3 13.9 ing the study questions.
During school 6.7 22.2 The primary study questions were designed to examine how
Early afternoon 37.1 25
adolescents problem solve in a bullying situation. Adolescents
Evening 79.8 83.3
Bedtime 36 50
were asked to rate how they would feel in such a situation if they
were the victim on a number of emotions and how it might disrupt
Skill on SNS
Beginner 21.3 19.4
their typical activities. Adolescents then provided information on
Experienced 50.6 61.1 how they might respond if they were the victim, and how impor-
Expert 28.1 19.4 tant different outcomes or goals would be. The answer choices pro-
Daily hours on SNS vided were selected from several prior studies based on the
0 11.2 13.9 responses that were included and endorsed most often (Craig
1 47.2 52.8 et al., 2007; Dehue et al., 2008; Patchin & Hinduja, 2006).
2 24.7 16.7
Adolescents were also given several open-ended prompts (e.g.,
3 9 8.3
4+ 7.8 5.6 after being asked to rate the general impact of traditional vs.
Facebook bullying, adolescents were asked ‘‘Why?’’ and given
Use Internet at least once a day for
Facebook 79.8 80.6 space to respond). These data were analyzed qualitatively.
MySpace 0 0
Twitter 2.2 21.2 3.2. Results and discussion
YouTube 52.8 47.1
Formspring 4.5 0
Email 73.1 70.6
A total of 89 students were used in the analysis. A repeated
AIM 4.5 6.7 measures multivariate analysis of variance (MANOVA) was used
GChat 5.6 3.3 to analyze each set of dependent variables (emotions and disrup-
Class assignments 52.8 67.6 tion in daily activity; responses to the bullying situation; desired
Chat rooms 1.1 3.2
outcomes or goals when responding to the situation). The
Note. SNS = Social Networking Sites. within-subjects factors used were type of bullying (threat of overt
 S. Horner et al. / Computers in Human Behavior 49 (2015) 288–295 291

aggression or relational aggression) and medium (electronic or off- Table 3

line). Gender of the participant was used as a between-subjects Repeated measures ANOVAs comparing responses to electronic and traditional
bullying, for OA and RA scenarios.
variable, and no systematic differences in responses by male vs.
female adolescents were found. Statistical test assumptions were Item Main effect of Main effect of OA Interaction
not violated and no effect due to counterbalancing was found. electronic vs. vs. RA
traditional
Ratings of negative emotions and disruption in daily activities indi-
cated that video clips and Facebook snapshots impacted adoles- F(1,88) Partial F(1,88) Partial F(1,88) Partial
eta2 eta2 eta2
cents in the expected manner, consistent with prior literature
(see Table 2). Tell friend 0.02 .000 1.29 .014 3.18 .035
Tell adult 11.00* .111 3.17 .035 0.95 .011
Students were asked to rate how they would feel if they were
Do nothing 3.29 .036 2.53 .028 2.86 .031
the victim in the bullying situation on a scale from 1 (much less Ask bully to 1.14 .013 0.10 .001 0.72 .008
than usual) to 5 (much more than usual), with 3 indicating ‘‘same stop
as usual’’ to allow for directionality, and adolescents to serve as Bully the bully 0.33 .004 3.24 .036 0.26 .003
Bully others 0.02 .000 0.31 .004 0.01 .000
their own controls. No significant effects from the MANOVA were
Avoid 6.56* .071 10.76* .111 7.67* .082
noted, suggesting that neither the medium (traditional vs. Facebook
Facebook) nor the type of bullying (OA vs. RA) had a consistent Avoid school 0.75 .009 1.42 .016 8.36* .088
effect on which emotions participants noted they would experi- Joke about it 0.94 .011 0.20 .002 0.11 .001
ence. As can be seen in Table 2, participants reported that they Plot revenge 4.69* .052 0.38 .541 0.00 .000
Report to 0.19 .001 3.65 .040 1.40 .016
would experience negative emotions in response to bullying
authorities
regardless of medium or type.
*
p < .05.
3.2.1. Analysis of the ways to respond to Facebook and traditional
bullying
students rated themselves more likely to seek adult assistance
Students were then asked how they would respond if they were
when bullying occurs in an off-line setting. The significant
the victim in the bullying situation on a scale from 1 (would not do
ANOVA interaction effect for this dependent variable indicates that
it) to 5 (would do it). MANOVA results suggested two interaction
adolescents were most likely to seek adult assistance when embar-
effects. A significant interaction between medium (traditional vs.
rassment bullying occurred in a traditional setting.
Facebook) and response (Wilks’ Lambda = .74, F(10, 68) = 2.45,
Additionally, follow up ANOVA results indicate that adolescents
p < .02, partial eta squared = .265) suggests that students are more
reported an even greater likelihood of avoiding Facebook after
likely to choose certain responses when bullying happens in a tra-
experiencing relational aggression online (electronic OA condition
ditional setting as opposed to electronic. A significant interaction
M = 2.78, SD = 1.27 vs. off-line OA condition: M = 2.32, SD = 1.09;
between bullying type (OA vs. RA) and response suggests that stu-
electronic RA condition: M = 2.95, SD = 1.24 vs. off-line RA condi-
dents are more likely to choose certain responses when dealing
tion: M = 2.91, SD = 1.34). This may be related to the large audience
with OA vs. dealing with RA (Wilks’ Lambda = .77, F(10,68) = 2.05,
that can see bullying online, which is particularly distressing when
p < .05, partial eta squared = .23). Follow up ANOVA results (see
embarrassment is the goal of the relationally aggressive bullying
Table 3) indicate that students are more likely to tell an adult when
behavior. Unfortunately, given the high usage of electronic and
an off-line bullying situation occurred, as compared to an incident
social media among youth the strategy of avoidance is likely to
of electronic bullying (electronic OA condition M = 2.66, SD = 1.33
be isolating, ineffective, and not maintained.
vs. off-line OA condition: M = 2.84, SD = 1.34; electronic RA condi-
tion: M = 2.81, SD = 1.41 vs. off-line RA condition: M = 3.11,
SD = 1.39). Given adults’ comparatively limited experience with 3.2.2. Analysis of the desired outcomes in response to Facebook and
technology and online communication, it is not surprising that traditional bullying
Students were asked how important different outcomes would
Table 2 be if they were the victim in the bullying situation on a scale from
Means and standard deviations of emotions and disruption in typical activities for 1 (not important) to 5 (extremely important). An omnibus
Study 1.
MANOVA analysis indicates a three-way interaction effect for med-
Item OA RA ium (traditional vs. Facebook), bullying type (OA vs. RA), and the
Online Off-line Online Off-line goals that students endorsed as important (Wilks’ Lambda = .88,
M SD M SD M SD M SD
Table 4
Sad 3.59** 0.96 3.36** 0.93 3.90** 1.01 4.03** 0.99
Repeated measures ANOVAs comparing desired outcomes when responding to
Angry 4.38** 0.86 4.25** 0.84 4.19** 0.85 4.35** 0.84
electronic and traditional bullying, for OA and RA scenarios.
Embarrassed 3.72** 1.12 3.71** 1.01 4.49** 0.77 4.58** 0.80
Afraid 3.84** 1.13 3.79** 1.17 3.45** 1.07 3.63** 1.13 Item Main effect of Main effect of OA Interaction
Bad about 3.43** 1.06 3.35* 1.09 3.62** 1.17 3.73** 1.22 electronic vs. vs. RA
myself traditional
Trusting of 2.31** 1.04 2.30** 0.90 2.08** 1.14 2.08** 1.25
F(1,88) Partial F(1,88) Partial F(1,88) Partial
othersa
eta2 eta2 eta2
Alone or 3.19 1.12 3.00 0.95 3.46** 1.13 3.52** 1.33
lonely Bullying to 0.01 .000 1.05 .012 0.98 .011
Want to go to 1.93** 0.88 2.01** 0.86 1.90** 0.87 1.75** 0.98 stop
schoola Get along with 0.36 .004 2.99 .034 1.50 .017
Want to go 2.24** 0.97 2.51** 0.88 1.90** 1.06 2.02** 1.07 bully
onlinea Keep friends 3.18 .035 6.87 *
.072 10.53 *
.107
Feel better 0.19 .002 13.27** .131 8.11* .084
Note. Scale ranges from ‘‘1: Much less than usual’’ to ‘‘5: Much more than usual’’ to Maintain 1.00 .011 0.46 .005 0.73 .008
allow for directionality, with ‘‘3: Same as usual’’. reputation
a
Reverse coded.
* *
p < .05. p < .05.
** **
p < .01. p < .001.
292 S. Horner et al. / Computers in Human Behavior 49 (2015) 288–295

F(4, 82) = 2.72, p < .04, partial eta squared = .12). Follow up ANOVA on and laugh. The electronic bullying example shows the same
analyses (see Table 4) revealed two dependent variables (‘‘keep comments being made via Facebook ‘‘comments’’ about the victim
friends’’ and ‘‘feel better’’) that students indicated would be most students’ ‘‘Facebook wall’’. The threat was also written on the
important to them when the scenario involved RA that occurred ‘‘Facebook wall’’. The actually study questions were the same as
in an off-line format (Keep friends: electronic OA M = 4.18, Study 1. Procedures used were identical to Study 1.
SD = .972 vs. off-line OA: M = 3.91, SD = 1.11; electronic RA:
M = 4.19, SD = 1.09 vs. off-line RA: M = 4.27, SD = 1.00; Feel better: 4.2. Results and discussion
electronic OA M = 4.28, SD = .941 vs. off-line OA: M = 4.11, SD =
.994; electronic RA: M = 4.44, SD = .839 vs. off-line RA: M = 4.56, A total of 36 students were used in the analysis. A repeated
SD = .783). Overall, students indicated that getting the bullying measures multivariate analysis of variance (MANOVA) was used
behavior to stop and feeling better were the most important out- to analyze each set of dependent variables (emotions and disrup-
comes in a bullying situation. tion in daily activity; responses; desired outcomes or goals). The
within-subjects factors used were gender of character in bullying
scenario (female or male) and medium (electronic or off-line).
4. Study 2
Gender of the participant was used as a between-subjects variable,
in order to assess any gender differences in participant responses.
4.1. Introduction and method
Statistical test assumptions were not violated and no effect due to
counterbalancing was found. To ensure that the video clips and
OA and RA scenarios in Study 1 involved female adolescents
Facebook snapshots used in the study had an impact on the adoles-
only, and research to date has found mixed results regarding the
cents in the intended direction, ratings of negative emotions and
role of gender in both traditional and electronic bullying.
disruption in typical activities were examined. Students reported
Therefore, Study 2 seeks to examine the impact of gender of char-
negative emotions and disruption in typical activities equal to or
acters in a bullying scenario on adolescents’ perceptions of bully-
more than usual after watching a bullying scenario. All means were
ing. Some prior literature suggests that, consistent with gender
in hypothesized direction. Additionally, almost all of the students’
role expectations, girls were more likely to be victims of electronic
average ratings were significantly different from how they usually
bullying (Dehue et al., 2008; Lenhart; 2007; Li, 2007; Smith et al.,
feel if they were the victim in the bullying situation (‘‘same as
2008), and boys were more likely to be bullies online (Dehue
usual’’ coded as ‘‘3’’) based on a series of one sample t-tests. This
et al., 2008; Li, 2006). In other studies, however, no gender effects
suggests the stimuli impacted the s in the intended manner (see
are found, and males and females were equally likely to be online
Table 6).
bullies and victims (Beran & Li, 2005; Patchin & Hinduja, 2006;
Slonje & Smith, 2008; Ybarra & Mitchell, 2004). In the studies
4.2.1. Analysis of the impact of Facebook and traditional bullying
where no gender effects are found, the lack of a significant differ-
An omnibus MANOVA analysis indicated a significant interac-
ence between genders in electronic bullying suggests that females
tion between gender of the characters in the scenario (male or
participate in electronic bullying more than they do in traditional
female) and emotion (Wilks’ Lambda = .46, F(8,26) = 3.86, p < .01,
bullying (Smith et al., 2008). This finding stands in comparison to
partial eta squared = .54), indicating that adolescents frequently
traditional, off-line bullying, which studies suggest occurs more
viewed scenarios with male characters as inducing more negative
often among males (e.g., Smith et al., 2008). Given the potentially
emotions than scenarios with female characters (for means and
different experiences that males and females have with traditional
standard deviations, see Table 5). Given that past research indi-
bullying and electronic bullying, Study 2 replicates the on and off
cates that OA among males is more likely than among females
line OA scenario for Study 1, and adds an additional overt aggres-
(Pepler, Jiang, Craig, & Connolly, 2008), it is interesting to note that
sion scenario with male characters as bully and victim. Thus, there
both male and female students viewed the male OA bullying
was a direct comparison for the OA of threat by gender.

4.1.1. Participants Table 5

The students for the second study were selected from the same Means and standard deviations of emotions and disruption in typical activities for
Study 2.
large public high school near a major Northeast city. Students were
recruited through announcements in their study hall classrooms Item OA – Male characters OA – Female characters
and via email. A total of 47 students, of approximately 450 eligible Online Off-line Online Off-line
students, had parental consent, provided assent, and completed M SD M SD M SD M SD
the study. This was an independent sample, with no overlap in s
Sad 4.36** 0.72 4.36** 0.80 3.81** 0.82 3.72** 0.85
from Study 1. Students’ responses were first examined to ensure
Angry 4.47** 0.70 4.58** 0.60 4.39** 0.65 4.36** 0.93
they were attending to the stimuli and identified the correct per- Embarrassed 4.50** 0.78 4.42** 0.77 3.89** 0.92 4.03** 0.97
son as the victim of bullying. After this initial analysis, 11 adoles- Afraid 4.50** 0.66 4.58** 0.65 3.89** 1.09 3.92** 1.03
cents’ data were dropped from subsequent analyses, leaving a Bad about 4.33** 0.72 4.22** 0.80 3.81** 0.92 3.78** 0.87
myself
sample of 36. Adolescents were asked to provide demographic
Trusting of 3.56* 1.28 3.75** 1.27 3.33 1.12 3.64** 1.02
information, summarized in Table 1. othersa
** ** * *
Alone or 4.03 1.03 4.17 0.94 3.42 1.05 3.34 0.91
lonely
4.1.2. Development of stimuli and procedures
Want to go to 4.28** 0.88 4.36** 0.90 3.78** 1.05 4.00** 0.89
The same OA scenario with female characters was used. In addi- schoola
tion, a second OA scenario with male characters was taken from Want to go 4.33** 0.93 4.06** 1.07 3.94** 1.09 3.78** 0.87
the popular media and edited to be similar in length to the clip onlinea
with female characters. The video clip of the male traditional bul- Note. Scale ranges from ‘‘1: Much less than usual’’ to ‘‘5: Much more than usual’’ to
lying example showed four high school-age adolescents accosting allow for directionality, with ‘‘3: Same as usual’’.
a
a younger student in the hall and saying disrespectful things about Reverse coded.
*
his physical appearance. Then, one of the largest students threat- p < .05.
**
p < .01.
ens to beat up the victim student after school while the others look
 S. Horner et al. / Computers in Human Behavior 49 (2015) 288–295 293

Table 6 Table 7
Repeated measures ANOVAs comparing responses to electronic and traditional Repeated measures ANOVAs comparing desired outcomes when responding to
bullying, for scenarios with male and female characters. electronic and traditional bullying, for scenarios with male and female characters.

Item Main effect of Main effect of Interaction Item Main effect of Main effect of Interaction
electronic vs. male characters electronic vs. male characters
traditional vs. female traditional vs. female
characters characters
F(1,35) Partial F(1,35) Partial F(1,35) Partial F(1,35) Partial F(1,35) Partial F(1,35) Partial
eta2 eta2 eta2 eta2 eta2 eta2
Tell friend 1.76 .048 0.89 .025 1.86 .050 Bullying to 0.47 .013 8.45* .194 0.05 .001
Tell adult 5.85* .143 5.09* .127 0.00 .000 stop
Do nothing 0.01 .000 0.16 .005 0.01 .000 Get along with 0.12 .003 2.76 .073 0.00 .000
Ask bully to 3.05 .082 1.09 .031 0.05 .002 bully
stop Keep friends 0.02 .000 0.02 .000 0.04 .001
Bully the bully 1.00 .026 5.00* .125 0.00 .000 Feel better 0.53 .016 0.06 .002 2.06 .059
Bully others 0.36 .010 0.66 .019 0.55 .016 Maintain 0.00 .000 4.53* .118 0.14 .004
Avoid 2.46 .066 4.14* .106 1.09 .030 reputation
Facebook *
Avoid school 3.15 .083 4.53* .115 1.55 .042 p < .05.
Joke about it 0.64 .018 8.66* .198 4.58* .116
Plot revenge 1.29 .036 4.53* .115 0.15 .004
Report to 0.52 .015 5.08* .127 0.02 .001 comments suggest that although students report both on and off-
authorities line bullying are similar in the general level of distress, they
* appreciate of the structural differences in the medium of
p < .05.
communication.
scenarios as more emotionally upsetting. It is unclear if this higher
level of emotional distress is due to an expectation of greater 4.2.3. Analysis of the desired outcomes in response to Facebook and
physical harm with male vs. female OA or some other factors. traditional bullying
Students were asked how important different outcomes would
be if they were the victim in the bullying situation. MANOVA results
4.2.2. Analysis of the ways to respond to Facebook and traditional
reveal an interaction between gender of the characters in the sce-
bullying
nario (male or female) and goals the adolescent reports as being
Students were then asked how they would respond if they were
important (Wilks’ Lambda = .72, F(4,29) = 2.85, p < .05, partial eta
the victim in the bullying situation. MANOVA results suggest a
squared = .28), suggesting adolescents view different outcomes of
three-way interaction between gender of the characters in the sce-
a bullying situation as important when the situation involves
nario (male or female), response, and the adolescent’s gender
males bullying as opposed to females bullying. Follow up ANOVA
(Wilks’ Lambda = .45, F(10,23) = 2.85, p < .02, partial eta
analyses (see Table 7) indicate that maintaining one’s reputation
squared = .55). Follow up ANOVA analyses did not support differ-
is most important to adolescents when the scenario involved
ences related to participant’s gender. However, additional ANOVA
female characters, as opposed to male characters (online male
analyses (see Table 6) indicated that participants would be more
characters M = 3.20, SD = 1.23 vs. off-line male characters
likely to use the following strategies when a bullying situation
M = 3.17, SD = 1.29; online female characters M = 3.51, SD = 1.10
involved male characters, regardless of the medium in which the
vs. off-line female characters M = 3.54, SD = 1.04). This finding is
bullying occurred: avoid Facebook, avoid school, report to authori-
consistent with the high importance that adolescent females place
ties. In contrast, students reported that they would be more likely
on their social reputation. In addition, students indicated that the
to use the following strategies when a bullying situation involved
goal of having the bullying stop would be more important when
female characters, regardless of the medium in which the bullying
the bullying situation involved male characters, as opposed to
occurred: bully the bully, plot revenge. Interestingly, for bullying
female characters (Bullying to stop: online male characters
involving male characters the general preference is for the victim
M = 4.67, SD = .862 vs. off-line male characters M = 4.64,
to avoid peers, in contrast, for bullying involving female characters
SD = .899; online female characters M = 4.39, SD = .964 vs. off-line
the more likely strategy employed by the victim is to directly
female characters M = 4.33, SD = 1.15). It may be that male bullies
retaliate against the bully. One possible explanation for this finding
are perceived as more threatening than female bullies. Further
may be due to the differing nature of female vs. male aggression. In
research is needed to understand why this might be the case.
particular, females more often employ RA and among high school
students this has less potential social cost than OA. Thus, adoles-
cents’ willingness to employ retaliatory strategies when reacting 5. General discussion
to female bullies may reflect a belief that relationally aggressive
behavior is a potentially cost effective option. Further research is The current studies examined high-school adolescents’ percep-
needed to explore this greater likelihood for students to utilize tions of electronic and traditional bullying through having students
retaliation when responding to female bullies. observe hypothetical scenarios, imagine themselves as the victim,
Some notable differences between electronic and traditional and then describe the impact, responses, and goals in resolving
bullying did emerge through the open-ended questions. Several the bullying scenarios. As predicted, across all scenarios, students
students wrote that the ambiguity of threats delivered online were indicated a significant negative impact of bullying situations high-
a particular challenge. For example, one adolescent noted that, ‘‘on lighting the emotions of sadness, anger, embarrassment, fear, and
Facebook, it is much more difficult to read people’s feelings or reac- losing trust in others. However, there was a difference found for
tions. If someone was to threaten another person online, it would the strategies most endorsed in response to bullying as related to
not be easy to determine the seriousness of the situation like it the gender of the characters involved. When the characters were
would be if it happened in person’’. Another stated that threats male, participants of both genders were more likely to endorse
online are ‘‘worse because you cannot see the person’s reaction strategies of avoidance of the bully. In contrast, when the charac-
and you do not know the tone of how they are saying it’’. These ters involved (both bully and victim) were female, participants of
294 S. Horner et al. / Computers in Human Behavior 49 (2015) 288–295

both genders were more likely to endorse strategies of retaliation reported by adolescents to be both threatening (characteristic of
requiring increased contact between the protagonists. High-school OA) and embarrassing (characteristic of RA; in both Study 1 and
students also indicated the strategy of behavioral avoidance of the Study 2). Given how critical embarrassment is to adolescents, it
context (avoiding school or online activities) where they as the vic- is unclear if all bullying situations involving OA would be embar-
tim might encounter those who had witnessed the bullying. This rassing, or if this is specific to the scenario chosen, where there
suggests that the witness or bystander to bullying becomes a nega- was an audience present. Early experimental research on electronic
tive stimulus in addition to the actual bully. Both on- and off-line, bullying has focused on anger and sadness as responses to bullying,
the negative emotional reaction and disruption in typical daily leaving out embarrassment entirely (Pieschl et al., 2013). Thus, fur-
activities demonstrates the high personal cost of peer victimization ther research on the frequency and strength of embarrassment in
on adolescents across widely divergent bullying scenarios (e.g., response to electronic bullying is important.
Bonanno & Hymel, 2013; Sticca & Perren, 2013). Second, given the rapidly changing format of online communi-
Overall, across a range of bullying situations (RA or OA, male or cation, some emerging popular forms of communication were not
female protagonists) there were many similarities in the nature of assessed (e.g., Snap Chat, Twitter, Instagram). Third, due to the
student reactions to electronic and traditional bullying. As sug- small sample size in Study 2, it was not possible statistically to
gested by Bauman and Newman (2012), electronic bullying compare victim gender (in the bullying scenarios presented) with
appears to be a sub-type of the broader category of ‘‘bullying adolescents’ gender. Prior experimental and self-report studies
behaviors’’, with concomitant negative outcomes, not a separate yield mixed findings with regard to the role of gender in electronic
and unique phenomenon. This is consistent with preliminary bullying, particularly when directly compared with traditional bul-
research examining differences between equivalent bullying lying (e.g., Pieschl et al., 2013; Slonje et al., 2013). Therefore, future
behaviors on- and off-line among adolescents and young adults research should include more varied bullying scenarios (e.g., a
(Bauman & Newman, 2013; Sticca & Perren, 2013). It is likely that wider range of relationally aggressive behaviors), with the gender
adolescents experience online and social media activities as an of bullies and victims manipulated. This will allow for further
integrated part of their everyday lives. The negative impact of bul- exploration of gender differences in responses to a variety of elec-
lying seemed to vary more in response to the type of bullying, tronic and traditional bullying situations.
overt or relational aggression, or by the gender of the protagonists Overall, given the significant impact, negative emotional reac-
than by whether the bullying occurred on- vs. off-line. Consistent tion, and disruption in typical daily activities of both electronic
with the framework of egocentrism and heightened social sensitiv- and traditional bullying (e.g., Bauman & Newman, 2013; Bonanno
ity (Buhrmester & Furman, 1987), many of the adolescents’ & Hymel, 2013), adults in adolescents’ lives must recognize the
responses to the open-ended questions in Study 1 indicate the importance of addressing bullying when it occurs both on- and off-
importance of their peer group, and spectators’ potential judg- line. The high acceptance of avoidance strategies in response to bul-
ments or opinions. For example, students wrote in response to lying is particularly worrisome as it may lead to isolation and
the relational aggression scenario that it would be particularly reduces the opportunity for positive peer experiences. Given the
hurtful, ‘‘Because all of your friends can see it’’, ‘‘In this situation, high degree to which adolescents value the viewpoint of peers, plus
it was completely public for all of their friends to see, and they their heightened sensitivity to being publically embarrassed, victims
have a whole lot of friends. Also, it can’t be taken back or denied’’, may mistakenly avoid those students who could actually be a
‘‘Everyone can see everything on Facebook’’, and ‘‘Facebook bully- resource. Initial literature suggests that boys may be more likely to
ing only hurts your reputation, and maybe your feelings’’. These avoid seeking social support when confronted with electronic bully-
responses highlight that the victims’ awareness of an audience ing, and may particularly benefit from interventions designed to
has a powerful negative impact when being bullied. Given the dis- combat this tendency (Pieschl et al., 2013). Bystanders and friends
tinct nature of in-person vs. electronic audiences (e.g., number of are likely to be a critical protective factor if they are able to mitigate
witnesses, temporal stability, abstraction) the nature of the embar- the victim’s sense of embarrassment. Future research needs to
rassment and distress resulting from bullying in the two mediums examine how best to empower adolescents to address the
needs further exploration. The public nature of the bullying mat- developmental challenges associated with social egocentrism.
ters and bystanders are not viewed as benign observers but as
peers with the potential to judge the victim.
Although adolescents in Study 2 emphasized fear in the OA sce- References
narios, adolescents in both studies noted that embarrassment was
a powerful emotion connected to all bullying situations. The Aricak, T., Siyahhan, S., Uzunhasanoglu, A., Saribeyoglu, S., Ciplak, S., Yilmaz, N.,
et al. (2008). Cyberbullying among Turkish adolescents. CyberPsychology and
heightened concern about peers’ viewpoints is highly relevant to Behavior, 11(3), 253–261.
adolescents’ experience of being bullied both on- and off-line, Bauman, S., & Newman, M. L. (2013). Testing assumptions about cyberbullying:
and thus needs to be taken into consideration when designing pre- Perceived distress associated with acts of conventional and cyber bullying.
Psychology of Violence, 3, 27–38. http://dx.doi.org/10.1037/a002986.
vention and intervention programs. For example, strengthening Beran, T., & Li, Q. (2005). Cyber-harassment: A study of a new method for an old
adolescents’ positive friendships and building new relationships behavior. Journal of Educational Computing Research, 32(3), 265–277.
may help adolescents to diminish the disruptive power of the Bonanno, R. A., & Hymel, S. (2013). Cyber bullying and internalizing difficulties:
Above and beyond the impact of traditional forms of bullying. Journal of
imaginary audience. One intervention program, described by Youth and Adolescence, 42(5), 685–697. http://dx.doi.org/10.1007/s10964-013-
Slonje, Smith, and Frisén (2013) involves adolescents forming 9937-1.
small workgroups at school to deal with specific issues of cyber- Buhrmester, D., & Furman, W. (1987). The development of companionship and
intimacy. Child Development, 58, 1101–1113.
bullying. This may be a particularly effective intervention because Craig, W., Pepler, D., & Blais, J. (2007). Responding to bullying: What works? School
it not only places technology-savvy adolescents at the forefront of Psychology International, 28(4), 465–477.
problem solving, but also allows adolescents in these groups to Dehue, F., Bolman, C., & Völlink, T. (2008). Cyberbullying: Youngsters’ experiences
and parental perception. CyberPsychology and Behavior, 11(2), 217–223.
form positive peer relationships, thereby improving the school
Elkind, D. (1967). Egocentrism in adolescence. Child Development, 38(4), 1025–1034.
climate. Gladden, M. R., Vivolo-Kantor, A. M., Hamburger, M. E., & Lumpkin, C. D. (2014).
As the current study was an early step in understanding adoles- Bullying surveillance among youths: Uniform definitions for public health and
cents’ perceptions of electronic and traditional bullying using an recommended data elements. Centers for Disease Control and Prevention &
United States Department of Education, Version 1. Atlanta GA; National Center
experimental design, there are important limitations to be for Injury Prevention and Control, Centers for Disease Control and Prevention
addressed through future research. First, the OA scenario was and U.S. Department of Education.
 S. Horner et al. / Computers in Human Behavior 49 (2015) 288–295 295

Huang, Y., & Chou, C. (2010). An analysis of multiple factors of cyberbullying among Raskauskas, J., & Stoltz, A. D. (2007). Involvement in traditional and electronic
junior high school students in Taiwan. Computers in Human Behavior, 26, bullying among adolescents. Developmental Psychology, 43(3), 564–575.
1581–1590. Rigby, K., & Slee, P. T. (1991). Dimensions of interpersonal relation among
Lenhart, A. (2007). Cyberbullying and online teens. Pew Internet & American Life Australian children and implications for psychological well-being. Journal of
Project <http://www.pewinternet.org/~/media/Files/Reports/2007/PIP%20Cy Social Psychology, 133, 33–42.
berbullying%20Memo.pdf.pdf>. Schneider, S. K., O’Donnell, L., Stueve, A., & Coulter, R. W. (2012). Cyberbullying,
Lenhart, A., Madden, M., Smith, A., Purcell, K., Zickuhr, K., & Rainie, L. (2011). Teens, school bullying, and psychological distress: A regional census of high school
kindness and cruelty on social network Sites: How American teens navigate the students. American Journal of Public Health, 102(1), 171–177.
new world of "digital citizenship". Pew Internet & American Life Project. Seals, D., & Young, J. (2003). Bullying and victimization: Prevalence and relationship
Li, Q. (2006). Cyberbullying in schools: A research of gender differences. School to gender, grade level, ethnicity, self-esteem and depression. Adolescence, 38,
Psychology International, 27(2), 157–170. 735–747.
Li, Q. (2007). New bottle but old wine: A research of cyberbullying in schools. Slonje, R., & Smith, P. K. (2008). Cyberbullying: Another main type of bullying?
Computers in Human Behavior, 23(4), 1777–1791. Scandinavian Journal of Psychology, 49, 147–154.
Mason, K. L. (2008). Cyberbullying: A preliminary assessment for school personnel. Slonje, R., Smith, P. K., & Frisén, A. (2013). The nature of cyberbullying, and
Psychology in the Schools, 45(4), 323–348. strategies for prevention. Computers in Human Behavior, 29(1), 26–32.
Nansel, T. R., Overpeck, M., Pilla, R. S., June Ruan, W., Simons-Morton, B., & Scheidt, Smith, J. (2009b). Nielsen: Average US user now spends over 4.5 hours per month on
P. (2001). Bullying behaviors among US youth: Prevalence and association with Facebook <http://www.insidefacebook.com/2009/07/14/nielsen-average-us-
psychosocial adjustment. JAMA, 285(16), 2094–2100. user-now-spends-over-45-hours-per-month-on-facebook/>.
National Children’s Home. (2005). Putting U in the picture: Mobile bullying survey Smith, P. K., Mahdavi, J., Carvalho, M., Fisher, S., Russell, S., & Tippett, N. (2008).
2005 <http://www.filemaker.co.uk/educationcentre/downloads/articles/ Cyberbullying: Its nature and impact in secondary school pupils. Journal of Child
Mobile_bullying_report.pdf>. Psychology and Psychiatry, 49(4), 376–385.
Patchin, J. W., & Hinduja, S. (2006). Bullies move beyond the schoolyard: A Sticca, F., & Perren, S. (2013). Is cyberbullying worse than traditional bullying?
preliminary look at cyberbullying. Youth Violence and Juvenile Justice, 4(2), Examining the differential roles of medium, publicity, and anonymity for the
148–169. perceived severity of bullying. Journal of Youth and Adolescence, 42, 739–750.
Pepler, D., Jiang, D., Craig, W., & Connolly, J. (2008). Developmental trajectories of http://dx.doi.org/10.1007/s10964-012-9867-3.
bullying and associated factors. Child Development, 79(2), 325–338. http:// Ybarra, M. L., & Mitchell, K. J. (2004). Online aggressor/targets, aggressors, and
dx.doi.org/10.1111/j.1467-8624.2007.01128.x. targets: A comparison of associated youth characteristics. Journal of Child
Pieschl, S., Porsch, T., Kahl, T., & Klockenbusch, R. (2013). Relevant dimensions of Psychology and Psychiatry, 45(7), 1308–1316.
cyberbullying: Results from two experimental studies. Journal of Applied
Developmental Psychology, 34(5), 241–252.