Abstract
While research has examined bystander responses in a traditional sense, there is a dearth of research regarding responses of cyberbystanders in a real-time situation, such as observing a cyberbullying scenario. This article describes a novel protocol designed to develop a series of images to be used to undertake research that aims to examine cyberbystander responses. A total of 24 scenarios (12 negative (cyberbullying) and 12 neutral) were created by the researchers and designed to mimic the way such scenarios would appear on a social networking site. The negative (cyberbullying) stimuli were rated in terms of level of severity, and the scenarios were compared to a selection of images from the International Affective Picture System using the Self-Assessment Manikin. These stimuli were compiled to form the Cyberbullying Picture Series (CyPicS). Through the development of the CyPicS, this protocol will aid future researchers in examining responses to real-life scenarios, as it is the first of its kind to develop these scenarios and test and evaluate them. CyPicS will provide researchers with the means to systematically evaluate responses to validated, real-life cyberbullying scenarios. More specifically, future researchers can utilize CyPicS to investigate how cyberbystanders respond when observing cyberbullying stimuli compared to neutral stimuli, as well as to measure and understand reactions or perceptions of cyberbullying. CyPicS can be used in any form of cyberbullying research (including electroencephalography and eye-tracking studies, psychological research, and functional magnetic resonance imaging studies) that may utilize bystander reactions and behaviors. Findings from research that utilizes CyPicS will greatly increase our understanding of bystander responses, and with variations in study design, researchers can further examine past or future associations with cyber-victim/bully status and mental health outcomes.
Introduction
Cyberbullying is as an aggressive, repeated, intentional act carried out on an individual using electronic forms (Smith et al., 2008). While prevalence estimates vary, it is well documented to have serious impacts on mental health (Fahy et al., 2016; Le et al., 2017; L. McLoughlin et al., 2018, 2019). To date, there has been limited research specifically examining potential links between cyberbullying and its impacts on the brain. While research shows the brain experiences peer victimization in a similar way to physical pain (Vaillancourt et al., 2013), research regarding the behaviors and reactions of cyberbystanders and the brain remains vastly underexplored. More specifically, little research has used real-time scenarios to measure how young people respond or react to a cyberbullying incident and even less has examined how the brain responds to witnessing a cyberbullying incident.
The present study aimed to develop a series of images to be used to examine cyberbystander responses to witnessing cyberbullying. This article will begin by briefly reviewing literature regarding cyberbystanders, specifically regarding the use of cyberbullying scenarios when researching cyberbystanders, and what factors may impact how a cyberbystander responds to cyberbullying. In addition, this article will propose the use of such scenarios for many different types of research, before outlining the methodology of the current study and discussing the results.
Cyberbystanders
Traditionally, bystanders are described as including four main participant roles: (i) assistants who join in the bullying, (ii) reinforcers who give positive feedback to the bully, (iii) outsiders who do not get involved and stay away, and (iv) defenders who aim to support and defend the victim (Jenkins & Nickerson, 2017; Salmivalli et al., 1996). Bystanders online, also referred to as “cyberbystanders,” have the power to contribute to bullying others by forwarding cyberbullying posts to their friends, or others. In addition, cyberbystanders could be with the cyberbully when the post is made or with the cybervictim when it is received, or witness the sharing and forwarding of bullying posts (Kowalski et al., 2012; Menesini & Nocentini, 2009; Smith, 2014). This behavior of sharing cyberbullying posts implies that a single cyberbullying act can be more than an isolated incident, and the negative effects of it can be repeated through multiple means (Grigg, 2012). Furthermore, research by Smith et al. (2008) suggests that by not acting on witnessing cyberbullying behaviors, bystanders can prolong the effects of the cyberbullying.
Few studies have used cyberbullying scenarios as part of their research design to test cyberbystanders responses. Bastiaensens et al. (2014) examined the responses of 453 students of Flemish secondary schools, with an average age of 13 years and found that bystanders had higher behavioral intentions to help the victim when they witnessed a more severe incident. The research also found that bystanders had higher behavioral intentions to join in the bullying when other bystanders were good friends (Bastiaensens et al., 2014). Similarly, research with 206 students (mean age of 13 years) using video scenarios depicting different emotional responses of victims found that angry victims elicited more negative reactions, sad victims elicited greater intentions to act, while incidents involving confident victims were perceived as less serious (Sokol et al., 2015). In other words, the response of the victim can influence bystanders’ attitudes toward that victim. Other studies have used virtual environments to simulate cyberbullying scenarios, but with limited findings regarding how young people would react in real life (Wright et al., 2009). Price et al. (2014), using digital animations of typical cyberbullying scenarios to explore young people’s (aged 8–12 years, M = 15 years) views of cyberbystanders, reported that young people perceive cyberbystanders to have the capacity to morally engage in cyberbullying incidents. The researchers did explain, however, that there are various barriers to a given bystander’s active positive engagement, such as the fear of being victimized themselves, that should also be considered (Price et al., 2014).
Other cyberbystanders research based on personal experiences has found particular characteristics related to whether bystanders help a cybervictim or not. Pabian et al. (2016) found in their study of 1412 adolescents aged 10–13 years that exposure to cyberbullying acts predicts lower levels of empathic responsiveness over time. In a study of 2309 adolescents, aged 9 to 17, Erreygers et al. (2016) found that less impulsive adolescents were more likely to help a cybervictim. These aforementioned studies suggest that constant exposure to cyberbullying can reduce the level of empathy in adolescents. Those who are more impulsive are more likely to join in with the cyberbullying or ignore it suggesting that cyberbystanders are complex with a range of characteristics. Importantly, such findings also suggest that there may be different phenotypes with different brain responses and/or different neurobiological underpinnings in the various cyberbullying roles.
Little to no research has used realistic social media scenarios with cyberbullying comments to measure how a cyberbystander responds to witnessing cyberbullying. Some researchers, however, have addressed ratings of severity (Sticca & Perren, 2013) or perceived severity (Palladino et al., 2017) of cyberbullying or seriousness of cyberbullying (Chen & Cheng, 2016; Chen et al., 2012). Each of these studies have used five-point Likert scales rating bullying as either less severe/serious to more severe/serious. Here, we aim to expand on the previous literature by developing a set of cyberbullying scenarios that can be used to investigate how cyberbystanders respond when observing cyberbullying stimuli compared to neutral stimuli. Thus, the overall aim of this project was to validate and optimize a cyberbullying stimulus paradigm for use in future research that incorporates methodologies that measure brain response and/or activity (e.g., electroencephalography, eye-tracking studies, psychological research, and functional magnetic resonance imaging, fMRI, studies) in various populations (e.g., adolescents and young adults).
Method
This study was approved by the University of the Sunshine Coast (USC), Human Research Ethics Committee, approval number A181135.
Recruitment
Young adult participants (aged 18–25 years) were recruited through the University of the Sunshine Coast (USC). Information was shared about the study via social media posts, student newsletters, student support services, announcements in lectures, the university website, and word of mouth. Those who expressed interest in participating where given the information sheet and were directed to the online survey via the Qualtrics platform.
Participants
A total of 32 participants completed the survey. The mean age of the total sample was 21.5 ± 2.02 years and 56.3% (n = 18) were female. No participants identified as Aboriginal and Torres Strait Islander.
Scenarios
A total of 24 scenarios were created by the researchers and were designed to look as they would appear on a social networking site and were based on a scenario used in a study by Bastiaensens et al. (2014). Using photoshop software, the scenarios were made to appear as though they were “posts” on social networking sites (with no branding of any particular site), with comments associated with them to determine their stimulus condition (negative, i.e., cyberbullying, or neutral). A total of 12 scenarios were made in each condition. The scenarios contained a nonidentifiable photograph (provided with consent) with a neutral caption beneath it with “hashtags.” The photographs all had a single person in them, but no faces were shown. Beneath each photo and caption was either a single cyberbullying comment (negative stimulus) or neutral comment (neutral stimulus). The cyberbullying comments were based on real life comments obtained from various social media platforms. The neutral stimuli included the same images as the negative (cyberbullying) stimuli, with only the comment differing to determine the two conditions. Importantly, the cyberbullying comment was given more “likes” than the post itself, to portray the sense of a power imbalance between the victim and bully, in that the victim would feel a sense of powerlessness when witnessing the cyberbullying comment receiving so many “likes.” Similarly, the sense of repetition, another criterion for cyberbullying, would have been felt, when seeing these “likes” continuing to rise on the cyberbullying comment.
An example of one of the negative (cyberbullying) stimulus, with a neutral stimulus using the same photo alongside it, is depicted in Figure 1.

Example cyberbullying scenario (left) and neutral scenario (right) from CyPicS.
Measures
The cyberbullying scenarios were rated on their severity to ensure the scenarios would invoke an emotional response and would be perceived as bullying. This study adapted the Bullying Severity Scale (Chen & Cheng, 2016; Chen et al., 2012) to rate the severity of the cyberbullying scenarios, by using a visual analogue scale (i.e., 0 being not severe at all and 100 being the most severe). Any rated with a severity score of 50 or above would be considered for inclusion in the Cyberbullying Picture Series (CyPicS).
The 12 images used in the scenarios were assessed on the three dimensions of pleasure, arousal, and dominance using the Self-Assessment Manikin (SAM; Lang, 1980) and were compared to a similar selection of images from the International Affective Picture System (IAPS) with ranges of neutral to positive mean scores on the three dimensions (Lang et al., 2008). IAPS was developed to “provide ratings of affect for a large set of emotionally evocative, internationally accessible, color photographs that includes contents across a wide range of semantic categories,” and recruited approximately 100 college students to rate each picture (Lang et al., 2008, p. 2). By comparing our images to the IAPS images, we could determine that any responses to the negative (cyberbullying) stimulus will be because of the bullying comment, rather than as an emotional response to the pictures; the valence of the stimulus will be determined by the content of the comments. These scenarios will then form the Cyberbullying Picture Series (CyPicS). Importantly, the CyPicS has now been validated for use as an fMRI task, in a study to understand bystanders responses to witnessing cyberbullying (McLoughlin et al., 2020).
For ease of interpretation, participants in the current study will be from here on referred to as “CyPicS participants” and those involved in the original IAPS study (Lang et al., 2008) will be referred to as “IAPS participants.”
Data Analyses
Data were initially screened for the standard assumptions for bivariate correlation testing including independence, normality, linearity, and homoscedasticity. SPSS® Version 24 (SPSS Inc., Chicago, Illinois, USA) was used to perform independent samples t tests to investigate differences between IAPS participants mean SAM ratings of the IAPS images compared to CyPicS participants, CyPicS participants mean SAM ratings of the IAPS images compared to CyPicS images, CyPicS male and female ratings of CyPicS images of severity.
In this article, we report all measures, manipulations, and exclusions, and sample size was determined before any data analysis. Power analysis revealed that to achieve a significance level of p < .05 and 80% power, the sample size was sufficient.
Results
Validity and Reliability
Each scale used had strong reliability statistics: Severity scale (α = .95), valence (α = .86), arousal (α = .76), dominance (α = .78). The reported Cronbach’s alpha values of above .7, indicated good psychometric properties of the scales employed (Kline, 2013).
Intraclass correlation analyses indicated that our estimated reliability between participants of each scale was 95% (CI 0.91, 0.97) for the severity scale, 95% (0.92, 0.97) for valence, 77% (0.63, 0.87) for arousal, and 72% (.55, .84) for dominance.
SAM Ratings
Table 1 displays mean ratings (and standard deviations) for the CyPicS sample compared to the IAPS (Lang et al., 2008) samples on the valence, arousal, and dominance dimensions for each of the selected neutral IAPS images. While there were no significant differences between ratings of valence and arousal across the IAPS images between the groups, the IAPS sample had significantly higher ratings of dominance than the CyPicS sample, t(22) = 4.05, p = .001 (Table 1).
Mean scores (±standard deviation) of the three SAM dimensions (valence, arousal, and dominance) between the CyPicS sample and IAPS sample ratings.
IAPS: International Affective Picture System; CyPicS: Cyberbullying Picture Series.
When comparing the CyPicS sample ratings of the CyPicS versus IAPS (Lang et al., 2008) images, there were no significant differences in mean ratings of valence or dominance (p > .05); however, participants rated the IAPS images with significantly higher scores of arousal (M = 6.11, SD = 1.13) than the CyPicS images (M = 4.52, SD = 0.90), t(22) = 3.83, p = .001. Table 2 displays mean ratings (and standard deviations) according to the SAM (Lang, 1980) dimensions of valence, arousal, and dominance for each of the 12 CyPicS images in isolation, as well as ratings of severity of the CyPicS scenarios.
Mean scores (±standard deviation) of the three SAM dimensions (valence, arousal, and dominance of CyPicS images, and severity ratings of CyPicS scenarios) in order of severity.
CyPicS: Cyberbullying Picture Series.
aFemale-based image.
bMale-based image.
Severity
As evident in Table 2, the CyPicS scenarios received various ratings for severity; however, the scores are ordered from those rated as most severe to least severe. All scenarios, excluding image 10, received scores of severity above 50, suggesting they met the threshold to be included in CyPicS.
Discussion
This article has outlined the development of the Cyberbullying Picture Series (CyPicS), a world first suite of scenarios designed to be used in studies examining cyberbystander responses. While research has examined bystander responses in a distal sense (i.e., previous experience), there is a dearth of research regarding cyberbystander responses in a simulated real-world scenario, such as that offered by CyPicS. The value of a cyberbystander approach is that a larger pool of participants are potential cyberbystanders, and researchers are less constrained in terms of assigning cyberbully and/or cybervictim status to such participants.
CyPicS was developed to be used in any form of cyberbullying research (including electroencephalography and eye-tracking studies, psychological research, and fMRI studies) that may utilize bystander reactions and behaviors. Therefore, CyPicS was examined using the SAM (Lang, 1980) to measure feelings of valence, arousal, and dominance, when viewing the images alone (with no associated bullying comment), to ensure that any responses to observing the cyberbullying scenarios were because of the comment, not the image itself. Furthermore, the scenarios were rated on their severity to understand participants perceptions of the images.
Our results indicate that the CyPicS sample found the IAPS images to invoke less of a dominance emotion (compared to the IAPS sample; Lang et al., 2008), and the CyPicS images to invoke less of an arousal emotion than the IAPS images. These results are supportive of using the CyPicS scenarios, for example, as a fMRI task, as researchers can be confident that participant responses are more likely due to the associated cyberbullying comments, rather than an emotional response to the image. In addition, while all 12 scenarios could be used in research to determine responses to different types of cyberbullying comments, in studies with time constraints or more complex designs (such as fMRI or EEG studies), the top 6 scenarios (i.e., those with higher severity ratings) should be used, especially when used in addition to the neutral scenarios (thereby equaling 12 scenarios). Furthermore, the CyPicS has now been validated for use as an fMRI task, in a study to understand bystanders responses to witnessing cyberbullying (McLoughlin et al., 2020).
These findings are particularly important in regard to their relevance to past research but especially in regard to informing research going forward. As this research is the first of its kind, it is difficult to compare our results to other studies. However, it is important to explain these results in light of the aforementioned literature. For instance, research by Bastiaensens et al. (2014) found that bystanders had higher behavioral intentions to help the victim when they witnessed a more severe incident. This finding is particularly important in line with our findings, as it could indicate that the more severe scenarios may invoke a more empathetic and helpful response in cyberbystanders when used in future studies.
Similarly, other researchers have found that exposure to cyberbullying acts predicts lower levels of empathic responsiveness over time and that less impulsive adolescents are more likely to help a cybervictim (Erreygers et al., 2016; Pabian et al., 2016). In other words, constant exposure to cyberbullying can reduce the level of empathy in adolescents, with more impulsive adolescents being more likely to join in. This is particularly relevant in regard to the use of the CyPicS in future studies, as researchers could use these scenarios in longitudinal studies to determine if participants have constant exposure.
Limitations and Future Research
This study had some limitations. First, participants commented that some of the scenarios had too many “likes” on the bullying comment. Thus, future studies may consider reducing the number of “likes,” as our participants stated that it should be reasonable number for a member of the public (i.e., not a celebrity). In addition, future studies should confirm these ratings with younger samples and may wish to replicate the scenarios in a variety of social media platform styles.
An important next step in this research would be to include the response of the victim when using cyberbullying scenarios, as past research shows that this can influence a bystanders response (Sokol et al., 2015). It is important to note that although our protocol is initially being tested in a young adult (18–25 years old) sample, the aim is to implement this in adolescent samples (i.e., 12–17 years old). Furthermore, measures of prior cyberbullying experience and social cognition should be considered in order to better understand why cyberbystanders respond the way they do.
Conclusion
This article will be an important contribution to existing research and will lead to a greater understanding of how cyberbystanders may respond to real-world cyberbullying scenarios. By including these images in future studies, we will better understand how some cyberbystanders experience and develop difficulties concerning their mental health, while others remain resilient in terms of their wellbeing, particularly if used in brain imaging and psychological studies. The CyPicS will aid other researchers in examining responses to real-world scenarios, as this paper is the first of its kind to develop these scenarios and validate them. Through the development of the CyPicS, researchers can compare data obtained from the one, standardized stimulus set with further iterations to be made that include responses from the cybervictim and vary according to the appearance of various social media platforms.
