What Constitutes Cyberbullying: Perspectives from Middle School Students

.


Defining the Characteristics of Cyberbullying
Due to cyberbullying characteristics similarity to traditional bullying (Olweus, 1997), researchers rely on three primary criteria posited by Vandebosch and Van Cleemput (2008; i.e., intended to harm, repetitive in nature, and representing an imbalance of power). Yet, cyberbullying encompasses additional factors not found in traditional bullying. For example, cyberbullying allows the ability for the perpetrator to not confront the individual in person, thus providing the illusion of anonymity. This may lead to an increase in disinhibition to post comments, pictures, videos, etc. that may not otherwise be shared while in a public setting. Finally, negative posts can be shared to a wider audience not inherent in traditional bullying which transpires in one place. It is imperative to also investigate how cyberbullying differs from traditional bullying according to the current, widely accepted definition to include: intent to harm, repetition, and a power differential (Patchin & Hinduja, 2015).

Intention
Intent to harm is a primary component when defining traditional bullying (Olweus, 1997). With cyberbullying, however, the intent is more subjective in nature as the intention of the perceived aggressor may be more ambiguous and easily misinterpreted by the recipient (Patchin & Hinduja, 2015). Younger social media users may be especially susceptible to misunderstanding the intentions of those they interact with online (Talwar et al., 2014). Additionally, words are more likely to be misinterpreted in online settings when someone intends a post to be a joke, yet is perceived as harmful by the recipient (Englander et al., 2017;Vandebosch & Van Cleemput, 2008). As a result, cyberbullying is particularly problematic because it is difficult for potential victims to accurately interpret the perpetrator's affect, tone, and demeanor.

Repetition
With traditional bullying, repetition is a necessitated component ( Hellström et al., 2015), meaning the negative behavior must be prevalent for more than one occurrence. With regards to cyberbullying, repetition is more ambiguous online. Even if a message, picture, or video is only posted one time, the post can be sent to more people or posted to a group site with an intent to harm (Slonje & Smith, 2008). Furthermore, repetition in relation to intent to harm can be dependent on whether the aggressive or harmful behavior is being administered directly by the perpetrator or indirectly by those re-posting what was original viewed (Englander, 2017). Therefore, unlike traditional bullying where the repetition is more clear-cut, online repetition is not confined to just one intentional perpetrator.

Imbalance of Power
In traditional forms of bullying, an imbalance of power involves one individual holding a physical or psychologically perceived advantage (the perpetrator) over another (the victim; Menesini & Nocentini, 2009). In the context of cyberbullying, some researchers have defined an imbalance in power as one person holding superior technological abilities over an intended victim (Pieschl et al., 2013). More recent data suggests this is not a factor when investigating whether a particular transgression was cyberbullying (Menin et al., 2021). However, when a recipient perceived the aggressor as holding more power, they were more likely to deem the experience as cyberbullying (Talwar et al., 2014). Therefore, further assessment of whether power imbalances play a crucial role in cyberbullying is warranted.

Anonymity
Anonymity is an important aspect of cyberbullying that contributes to the potential severity and frequency of these behaviors. For example, when a perpetrator believes that their actions are anonymous, there is the potential to experience disinhibition; this disengages the sender from feeling remorse or responsibility for their post (Raskauskas & Stoltz, 2007). There is growing concern that this ability might encourage people to say things online that they would not otherwise say in real life (Pabian et al., 2016). Furthermore, even when the sender's identity is available, the actual intention of the sender may be misinterpreted as in-person communication allows for inference of affect and tone (Slonje & Smith, 2008). Pieschl et al. (2013) warned that further research into disinhibition is important in order to better understand how this phenomenon occurs and what can be done to counter its effects. Moral disengagement may play a role in cyberbullying (Kowalski et al., 2014) as individuals are communicating behind a veil. More worrisome is a continuation of perpetration as the aggressor receives a visceral reward from harming others devoid of responsibility (Dooley et al., 2009). This aspect is troublesome as the victim may not be aware of who the perpetrator is, making the bullying sometimes difficult to stop or prevent (Barlett et al., 2016). Some researchers have found that in some cases, being unaware of the identity of the perpetrator can cause less angst for the victim, making them less inclined to experience detrimental outcomes (Menesini et al., 2012). Conversely, not knowing the perpetrator can potentially cause anxiety especially if the perpetrator is relentless (Smith, 2013). Anonymity gives the perpetrator a sense of being "invisible" to the victim.

Public versus Private Interactions
Interaction through online platforms have the potential to occur between two individuals or with broader audience Cyber interactions may occur privately between two people or publicly within a larger audience (Thomas et al., 2015). If a negative online interaction occurred repeatedly by the same individual, that would be considered cyberbullying. However, if the incident was posted once yet shared to a wider audience by that individual, that could also be considered cyberbullying (Nocentini et al., 2010). Slonje and Smith (2008) found participants stated when cyberbullying occurred within a large audience, it was more severe and detrimental than bullying or cyberbullying between two people (Chen & Cheng, 2017). Unfortunately, due to the ability to share posts with a wider audience without the need to take responsibility for one's actions, individuals may be more tempted to engage in this type of behavior (Thomas et al., 2015). Therefore, one public post may be considered "repetitive" depending on the breadth of the audience.

Perspective
According to Diazgranados et al. (2016), adequate perspective taking requires the ability to move beyond feeling what another might be experiencing, or intending, but also finding ways to problem-solve what is transpiring. This is particularly problematic for adolescents, as many are still learning to appreciate and understand the perspectives of others. Therefore, the ability to engage in adequate perspective taking when receiving an unwanted or aggressive post online, may be due to immaturity on the recipient's end (Diazgranados et al., 2016;Martin et al., 2008). Additionally, one member of the communication process may not see the harm in what is being posted, whereas the recipient does (Vandebosch & VanCleemput, 2009). Interestingly, Li's (2010) research indicated that many adolescents perceived cyberbullies to act out due to the behavior being viewed as fun. Additionally, expressing one's beliefs online was also reported as acceptable even if it hurt another as it fell under one's ability to speak freely. Therefore, there is the potential for the intentions of the sender and perceptions of the recipient to be grossly misaligned.
The additional factors related to cyberbullying increase the risk of potentially detrimental outcomes potentially causing distress and long-lasting harmful effects. The false sense of anonymity and the ability to reach a widespread audience contributes to these increased risks. In order to address these constraints, it is salient to first develop an agreed upon conceptualization of what constitutes cyberbullying, then operationalize the construct in a way so that consensus on findings aligns (Olweus & Limber, 2018). Once this has been established, researchers can investigate adaptive coping strategies and positive self-efficacious beliefs in order to diminish detrimental outcomes. Therefore, the purpose of this study is to better understand what constitutes cyberbullying behavior as viewed through the lens of a middle school student. The following research questions are addressed: 1) How do middle school students perceive potential cyberbullying situations?
2) How do middle school students describe the type of potential cyberbullying occurring and the reasons for the occurrence?

Method
The purpose of this explanatory mixed methods study is to better understand how middle school students perceive potential cyberbullying situations and what constitutes this type of behavior. To investigate the research questions, descriptive vignettes of potential cyberbullying situations (Mischel et al., 2018) were implemented to ascertain a deeper understanding of perceptions.

Participants
The convenience sample of students (N = 189) ranged in age from 11 to 14 and represented a diverse range of ethnicities and socioeconomic levels. Participants included males (n = 80), females (n = 106), and other (n = 3). Grade breakdown indicated that most of the participants (n = 122) were in seventh grade. When asked if the participant owned a cell phone, the majority (n = 159) responded yes, with almost all participants (n =177) stating they had access to a computer at home.
For the interview portion of the study, students (n = 10) indicating interest in participated in follow-up interviews randomly chosen from each grade level. Of those that were chosen, six brought back consent forms (M = 2; F = 4). None of the participants interviewed were in eighth grade but rather sixth or seventh grade (6 th = 4; 7 th = 2).

Setting
The study was implemented at two separate middle schools in the Mid-Atlantic region. One of the middle schools jedp.ccsenet.org Journal of Educational and Developmental Psychology Vol. 13, No. 1;2023 set up laptops in the cafeteria and in an auxiliary computer room which allowed students to take the survey in shifts, lasting a little over an hour in total duration. Students at the other middle school also took the survey in shifts using computers provided by the school located in one room. Qualitative interviews were conducted at one middle school only, and took place one week after the Vignettes as Vehicles for Defining and Assessing Cyberbullying (Mischel & Kitsantas, 2018) survey was administered. All participants and parents granted consent/assent and were audio-taped.

Data Collection Instruments
Participants were asked several demographic questions including gender, age, access to social media, social media sites frequented most often, and concern for cyberbullying. Participants were allowed to check off more than one response to this question and there were 285 responses in total. Respondents indicated use of Instagram (37.4%), Snapchat (33.7%), Twitter (7.7%), and Facebook (3.7%). Of those respondents who reported using other types of social media (74.0%), YouTube or none was reported most often. Other types of social media reported included Tik Tok, Musical.ly, and Houseparty.

Cyberbullying Vignettes
Vignettes as Vehicles for Defining and Assessing Cyberbullying (Mischel & Kitsantas, 2018). The instrument consisted of eight vignettes, 16 questions in total, on a 3-point Likert-type scale (1 = not cyberbullying, 2 = on the border, 3 = cyberbullying). Within each vignette, the scenarios escalated in severity to determine at which point middle school students perceived the situation to become cyberbullying. For this study, severity was interpreted as the degree of aggression perceived by the participant. (see example below).

Example
Gigi is a new student at a middle school. She is very pretty and many of the boys notice her. The other girls are jealous of the attention Gigi is getting. One of the girls posts that Gigi is a "slut" on social media. Another girl shares this post and asks who else agrees? Is this cyberbullying?
No On the Border Yes 1 2 3 The next day, one of the girls takes a picture of Gigi in the hallway and posts it on social media with the caption, "Slut," underneath. Quite a few of the boys at school, some she doesn't even know, post comments. Is this cyberbullying?

No
On the Border Yes 1 2 3

Semi-Structured Interview Protocol
In order to obtain a deeper understanding of why middle school participants stated a vignette accurately depicted cyberbullying and reason for this perspective, interviews were conducted. The interview questions were based on the vignettes presented in the former study. Participants were asked to, again, read through the vignettes and answer whether they believed the scenarios depicted cyberbullying. Either way, they were asked to explain their reasoning.
They were asked what type of cyberbullying was occurring and, finally, whether they felt the vignettes depicted were plausible.

Coding Procedures
After interviews were transcribed, pre-determined codes from prior research were used to investigate how middle school students identified a situation as cyberbullying as well as the reason. Codes used to identify cyberbullying situations included: harassment, denigration, exclusion, impersonation, outing/trickery, flaming, and cyberstalking (Li, 2010;Noncentini et al., 2010;Ortega et al., 2007). Initial codes to determine why it was considered cyberbullying included: intent to harm, ability to attack at any time, repetition, imbalance in power, and repeated victimization (Corcoran et al., 2015;Kowalski & Limber, 2008;Langos, 2012;O'Moore & Minton, 2009: Vandebosch & Van Cleemput, 2009). These codes were further broken down into subcodes codes.

Procedures
Prior to participation, all students were given a hard copy of the parental consent as well as an online copy and reminder to turn in the form. Those who turned in the parental consent were allowed to participate. Qualtrics (2017) was the platform of choice and the middle schools set up laptops in their school cafeteria and auxiliary computer room, rotating in shifts. The Vignettes as Vehicles for Defining and Assessing Cyberbullying survey lasted approximately 15 minutes and was comprised of questions based on vignettes piloted in a previous study (Mischel, 2018). Once participants completed the survey, they were entered into a drawing for a $20 Amazon gift card. The researcher did not have access to the participant names, but the principal's secretary chose 10 consent forms randomly and passed out gift cards to those students.
For the qualitative portion of the study, the researcher only contacted those participants who met the preliminary criteria of participating in the survey, expressed concern regarding cyberbullying behavior, and a willingness to participate in an interview. If interested, participants put their name on an index card. The researcher chose 10 random index cards for interviews. The names were sent to the principal's secretary, and she contacted the students. Students were then given another consent and assent form specifying the purpose of the interviews, the process, and the reward for participation (i.e., another $20 Amazon gift card). Similar to the first consent form, counseling services were again displayed to address any unforeseen harm participants might experience. If parents agreed and the student was still interested, the parents were asked to sign the consent form allowing their child to participate. Additionally, the consent forms indicated their child would be audio-recorded, so parents were asked to give consent for this as well. The student was only allowed to participate if the consent form was returned. Six students returned the signed consent forms. Interviews took place at one of the middle schools in a conference room in the main office, one week after the implementation of the survey.

Results
The primary purpose of this study was to explore middle school students' perceptions of what constituted cyberbullying based upon specific cyberbullying situations. Vignettes depicting potential cyberbullying situations were used to address these perceptions. Initially, a survey was administered followed by semi-structured interviews with six of the middle school students.

Results from Quantitative Data
Beginning scenarios of vignettes were developed to be less severe in nature, with the latter portion escalating.

Results from Qualitative Data
Follow-up interviews helped researchers better understand why the vignettes were rated as cyberbullying, on the border, or not cyberbullying.

Reasons for Perceptions of Cyberbullying
Two main themes for why participants reported a situation as being cyberbullying emerged from the interviews: repeated victimization and intent to harm. The most prominent of the two was repeated victimization.

Repeated Victimization.
Participants stated these entailed recipients experiencing victimization multiple times from the potential of others to re-post comments or images, or continue posting comments on such embarrassing posts or images. There was an overall consensus that when a post is re-posted, it encourages comments which can, at times, be detrimental to the recipient. For example, Todd shared his concern about posts being broadcast to a wider audience through his comments, "So more people from different schools can see it and post it over and over again which can get all over the place." Two other participants concurred both stating, "That would probably be cyberbullying because then everyone would see it." Ashley, a seventh-grade student, shared a similar perception, "I do think that it is because she's doing it more publicly because she's more popular." Similar to the definition of repetition of traditional bullying, Becca stated this would be harmful because, "That's multiple times that they called him that name and repetitive is considered bullying." When the situation was not shared with others, however, participants did not feel it escalated into cyberbullying. Two students stated: I think that since they just shared the text with one person, it's not really cyberbullying. (Joe, sixth-grader)

I don't think it's cyberbullying that much because it's on text and not social media or anything, and it was only shared to two people. (Ashley)
When the situation was later shared with others, most tended to consider this to be cyberbullying.

Intent to Harm.
Similar to the characteristic found in traditional bullying, intent to harm was considered to be cyberbullying by the interview participants. An example of intent to harm was encapsulated in Tina's comment, "Because they copied it, and they re-posted it without her consent. So, it was just [I feel] like they had a bad intention." Chelsea agreed stating, "Making others feel bad is cyberbullying." Becca shared that intent to harm was cruel stating, "Doing that can really hurt somebody and really harm them in the head because you could really feel unaware to anything after that. You could think that even something little could be cyberbullying because you're so scared after that." Interestingly, all interview participants, however, agreed that if the recipient was not aware of what was transpiring, then it would not be considered cyberbullying. Furthermore, there was the perception that the potential for harm was dependent on whether the posts were positive or negative.

Labeling of Cyberbullying
Research labels eight types of potential cyberbullying. These include: harassment, denigration, exclusion, impersonation, outing and trickery, flaming, and cyberstalking (Willard, 2007) (see Figure 1). The two types of cyberbullying reported most often during the interviews were denigration and outing and trickery (see Table 4). Denigration is described as an individual sending or posting messages that might damage one's reputation or friendships (Willard, 2007). When participants perceived the actions depicted in the vignettes as harmful toward the recipient, they considered the scenario to be cyberbullying. For example, Joe, a sixth-grade participant, commented on the potential to damage an individual's reputation when he shared that, "The statements being posted were not necessarily factual and could affect Peter's reputation (vignette 2)." Outing and trickery, sending or posting someone's secrets or embarrassing information or images (Willard, 2007), was stated when the participants felt the situation might embarrass the recipient. Outing and trickery can also include platforms for others to post derogatory comments through initial posts, that could be further commented on by others. This can include an individual sharing an embarrassing picture and then asking others to comment. For example, Chelsea, a seventh-grade student, commented, "It is actually cyberbullying because it says that it's shared and asking who else agrees."

Repeated Victimization and Denigration
Finally, an overall analysis of participant interviews, across all eight vignettes, was investigated to look for patterns and report conclusive findings. A majority of participants felt that all eight vignettes were plausible. Most believed the primary and secondary situations for Vignettes 1-4 depicted cyberbullying, whereas the tendency to only report cyberbullying for the secondary situations for Vignettes 5-8 was indicated. Vignettes 1-4 depicted situations on an individual level, whereas Vignettes 5-8 involved more than one individual. The theme that emerged most often was repeated victimization, followed by intent to harm. The type of cyberbullying described most often was outing/trickery followed closely by denigration.

Discussion
Understanding what constitutes cyberbullying is elusive. There continues to be a lack of how to appropriately conceptualize and operationalize the construct (Slonje et al., 2013), and as Patchin and Hinduja (2015) suggest, only extreme cases are typically identified. The findings of the present explanatory study helps further clarify what constitutes cyberbullying behavior according to middle school students. Interviews were particularly useful in understanding potential reasons online situations are perceived as cyberbullying, as well as the type of cyberbullying. Main reasons for perceiving online interactions as cyberbullying included intent to harm as perceived by the recipient and the potential for posts and/or pictures to be seen by a wider audience.

Perceptions of Cyberbullying
In analyzing the surveys, when participants perceived an act as being intentional, most stated the interaction indicated cyberbullying. The scenario did not need to escalate into being reposted or seen by many. Overall, survey participants reported that once the interactions on social media were perceived as mean-spirited, the situation was considered to be cyberbullying. This was evidenced with the first four vignettes. Each primary vignette depicted a situation in which the perpetrators were purposefully acting in a way to harm the recipient. For example, the first vignette describes two friends who hoped to make the cheerleading team. Only one was successful and the other less fortunate individual lashed out at the friend who made the team. The response by the less fortunate individual was perceived as intent to harm. This was further evidenced with vignettes 2-4. The primary scenarios for these included the element of an intent to harm the recipient. As intent to harm is one of the three criteria necessary for an act to be considered bullying, in the traditional sense, it is unsurprising participants would concur. Additionally, in a systematic review to define bullying, Younan (2019) found aggressive displays of behavior were noted most often when participants were asked what constitutes bullying. In another study, participants responded so strongly to instigation of intentional harm that even one incident could be considered bullying (Hellström et al., 2015).
Interviews indicated similar findings. Interestingly, most reported that when interactions between two individuals occurred, or situations in which the perpetrator did not seem to purposefully hurt the other individual, it did not constitute cyberbullying. For example, 6 th grade student Ashley shared that, "I don't think it's cyberbullying that much because it's on text and not social media or anything, and it was only shared to two people." However, when a hurtful interaction went beyond a two-way communication model or the actions were deliberately mean or included others, interview participants indicated the situation escalated into cyberbullying. This was shown through 6 th grader Tina who stated, "Because they copied it, and they reposted it without her consent, so I feel like they had a bad intention." Sixth grader Joe shared a similar perspective and included the potential for more individuals to be included saying, "People were posting mean comments, more than one person, about the text message." Yet, as Corcoran, McGuckin, and Prentice (2015) posit, how do we operationalize what "intent to harm," means in this era? Grigg (2010) states that when a cyber interaction is offensive, derogatory, harmful or unwanted, then this type of transgression constitutes intent to harm. As noted in prior research (Nocentini et al., 2010) and the present study, participants were less concerned about the type of messages posted and more so about how the recipient responded emotionally. In accordance with the literature (Corocoran et al., 2015), participants also complained that the recipients may not have the opportunity to defend themselves in real time. Thus, it seems reasonable to surmise that intent to harm online is more subjective in nature. This notion suggests helping middle school students interpret online messages without attaching an emotional component from the sender may potentially reduce this perception of intent to harm. However, this does not excuse blatant posts or messages that are clearly meant to harm the recipient.
Repetition was the other reported reason for stating an incident was cyberbullying which, again, aligns with the definition of traditional bullying (Olweus, 1997). Repetition as defined in traditional bullying is clear: repeated aggressive behavior (Wang & Iannotti, 2012). However, repetition online is difficult to conceptualize (Smith, del Barrio, & Tokunga, 2015) as it includes posts or messages being sent repeatedly, as well as a singular incident that is reposted or commented on repeatedly (Corcoran et al., 2015). Participants noted damage recipients compounds when this occurs. Adding to the difficulty in defining repetition, whether a post is created by a singular offender or multiple offenders, or re-posted, begins to fall under the title of power imbalance (Kazerooni, Taylor, Bazarova, & Whitlock, 2018), not illustrated in this particular study.
Interview participants agreed when posts go viral (i.e., posted to a widespread audience), or are reposted, damage to the recipient can escalate. Seventh grader, Chelsea, stated that when posts go viral, "It definitely is cyberbullying because it's sending it to even more people. They might send off more mean comments too." Becca, a 7 th grader, discussed the potential for long-term harm to the recipient stating, "I think that doing that can really hurt somebody and really harm them in the head because you could really feel unaware to anything after that. You could think that even something little could be cyberbullying because you're so scared after that." There are compounding factors related to this potential situation as once a post is shared by others, the ability to take down a derogatory post is out of the original perpetrator's control (Casas et al., 2020). This may, unfortunately, lead to exacerbated detrimental outcomes for the recipient.

Middle School Students Reports of the Type of Potential Cyberbullying Occurring and the Reasons for the Occurrence
The types of cyberbullying reported were also illuminating. Main types of cyberbullying include: harassment, denigration, exclusion, impersonation, outing/trickery, flaming, and cyberstalking (Li, 2010;Noncentini et al., 2010;Ortega et al., 2007). For this study, only two main types were repeatedly reported by participants: outing/trickery (i.e., sharing secrets or embarrassing information or images) and denigration (i.e., sending or posting gossip or rumors about a person to damage their reputation). It makes sense these types of cyberbullying were reported most often given survey and interview participants stated when online messages or posts are sent to intentionally harm the recipient, were reposted, or sent to a wider audience, the situation escalated into cyberbullying. This may be due to the importance of social relationships in adolescence and the innate desire to connect with others and avoid loneliness (Nixon, 2014). Additionally, research indicates adolescent feel a heightened fear of becoming cybervictimized by anyone, even a close friend (Raskauskas& Stoltz, 2007). Furthermore, unlike traditional bullying that tends to occur during the school day, cyberbullying can transpire after school with little supervision or accountability for the perpetrator (Dooley et al., 2009;Patchin & Hinduja, 2006). In addition to fears of being lonely or lacking emotional support, embarrassment from posts being spreading to a wider audience was reported through the interviews (Slonje et al., 2017). Dodging embarrassing situations is also crucial to middle school students and avoidance is considered a success (Rodkin et al., 2013).

Limitations and Future Research Directions
There were several limitations in this study. As there is the tendency to compare traditional forms of bullying to cyberbullying (Dooley et al., 2009;Perren & Gutzwiller-Helfenfinger, 2012;Slonje et al., 2013), participant responses may have been skewed depending on how they defined cyberbullying, as a definition was not provided. In addition, in order to protect participants, actual scenarios were not experienced, rather written situations were asked to be read and then interpreted. Responses may be different in an actual potential cyberbullying situation. Finally, interview participants were chosen at random and further screening may have yielded further insight. Furthermore, interviewing more participants may yield additional insight into perceptions of this behavior.
Further research is warranted in conceptualizing and operationalizing what constitutes cyberbullying. This study helped to further clarify what constitutes cyberbullying through the lens of middle school students: intent to harm and repetition. Yet, conceptualization of these components remains elusive due to subjectivity in defining intent to harm as well as whether repeated posts fall more appropriately under an imbalance in power than repetition.
Teasing out each of these constructs separately may help to better clarify these differences. Once a clear definition is agreed upon, researchers and educators will be better able to develop programs addressing cyberbullying behavior.
Furthermore, due to school closures in the past two years, an increased reliance on electronic devices to communicate with peers has arisen. Coupled with immaturity when interacting through electronic platforms, it jedp.ccsenet.org Journal of Educational and Developmental Psychology Vol. 13, No. 1;2023 seems imperative to additionally focus on strategies to help middle school students better understand how to communicate effectively through social media platforms. Even more salient, however, is helping adolescents realize that messages received do not necessarily have nefarious intent. For example, a quick reply received which negates certain punctuation marks may set some middle schoolers into a tailspin. Therefore, it might be helpful to remind adolescents to refrain from inferring what the sender meant until they can speak with them directly. Hence, finding effective strategies to remind middle school students of these misinterpretations may go far in reducing the perception of cyberbullying incidences.