1 Introduction

Imagine someone coming too close to you. Usually, we find this unpleasant and always try to keep a certain distance to avoid the so-called violation of personal space. This is also true for victims of cyberbullying, with the difference that insults and slander are not as easy to keep at a distance. Digital devices are omnipresent nowadays and thus seemingly inescapable and close to personal (living) spaces. The digitization of bullying experiences exacerbates this all-encompassing intrusion. Digital media are used in multiple contexts, attacks are accessible to the public and can be difficult to erase (Was ist Cybermobbing 2018). Even offline, offenders in bullying situations often lack empathy (Pfetsch et al. 2014). Online, they are even more detached due to the effects of online communication coined as the online disinhibition effect (Suler 2004). The online disinhibition effect is the lack of restraint and empathy felt in online communication compared to face-to-face communication. Therefore, promoting empathy is proposed as an effective means of preventing (cyber)bullying (see, e.g., (Ang and Goh 2010; Pfetsch et al. 2014; Schultze-Krumbholz et al. 2018)). Psychologically, prevention measures to increase empathy often resort to the mechanism of cognitive dissonance. Dissonance-based prevention programs aim to have one person walk in the shoes of another and thereby better understand that person’s perceptions and attitudes (Borbé et al. 2009). Through this method, perpetrators and victims of cyberbullying can be reached equally. However, students and teachers note a decline in available prevention measures and programs aimed at adolescents (Beitzinger et al. 2020), emphasizing the need for new and innovative forms of prevention programs that address the digital reality of children’s and young people’s lives.

Virtual reality (VR) has already been shown in numerous studies to improve the understanding of others. In particular, VR interventions have been shown to reduce prejudice, whether toward people of color (Peck et al. 2013), homeless people (Herrera et al. 2018), people with chronic pain (Tong et al. 2020), or people with mental illness (Penn et al. 2010). Similarly, VR interventions can induce behavior change (Wienrich et al. 2021) by lowering psychological distances [refer to the Construal Level Theory (Liberman and Trope 1998)]. Although impact factors for VR have been extensively studied, there are still comparatively few VR prevention programs for bullying (Nocentini et al. 2015; Ingram et al. 2019; Gu et al. 2022; McEvoy 2015). A small number of studies provide first promising evaluation results but focus mainly on role playing and perspective-taking in school bullying situations in VR (McEvoy 2015; Gu et al. 2022; Ingram et al. 2019). In particular, we are unaware of any prevention intervention that is designed to solely address cyberbullying and can elicit increased empathy for victims through VR-specific mechanisms of action.

To bridge this gap, this study presents a VR prevention strategy for cyberbullying. We used a virtual personal space invasion (PSI) by cyberbullying-associated objects, targeting perspective-taking with victims. We investigated the efficacy of our program in two human-centered circles. First, we explored the design space for mechanisms of virtual space invasion considering the context of cyberbullying in a pre-study and implemented the most promising one. Then, we conducted an empirical laboratory study with 41 university students to examine the effect of our VR prevention on empathy and affect. The second circle involved the target group of children and adolescents. A focus group of five children evaluated the previous implementation and provided numerous suggestions for improvements. We changed the implementation according to the recommendations and then used the improved VR prevention in a half-day workshop at a school. Initial feasibility, as well as potentials and risks of the VR prevention, was examined with 92 children. In summary, our study contributes to the effective development of a VR-based prevention against cyberbullying. It directly addresses digital reality in the lives of children and adolescents and builds on psychologically sound findings for designing prevention programs.

2 Related work

2.1 Cyberbullying

Cyberbullying refers to the intentional insulting, threatening, shaming, or harassing of others over an extended period of time using internet and cell phone services (Beitzinger et al. 2020). Over 90% of children and adolescents have a smartphone and use it at least daily, especially to access social media and messenger services (Feierabend et al. 2022). Thus, they are very likely to encounter cyberbullying, which poses a threat to their mental health (Reed et al. 2016; Ossa et al. 2022). In 2016, 8% of young internet users (12-19 years) in Germany were victims of cyberbullying (Feierabend et al. 2016). During COVID-19, the digitization of hate attacks gained more importance and urgency. During the pandemic, 17.3% of school students in Germany have already reported being affected (Germany: (Schuler 2020); similar results for other countries: (Jain and Yuan 2020; Polanin et al. 2021)). Although other studies reported lower prevalence (7.6%), cyberbullying significantly affects victims (Ossa et al. 2022). In 2022, the average duration of internet use by minors has reached pre-COVID rates, but the number of children affected by cyberbullying has increased severely up to 17,6% now consider themselves victims (Feierabend et al. 2022; Nolden 2021). Several studies found an association between cyberbullying and various physical and mental health problems (Barzilay et al. 2017; Reed et al. 2016; Bauman et al. 2013; Schuler 2020; Ossa et al. 2022). Victims reported negative affects, psychological problems, and engaged in risky and self-harming behaviors, including suicidal thoughts and suicide attempts (e.g., (Ossa et al. 2022)). Cyberbullying is also considered to exacerbate the effects of offline bullying. Various studies gave several reasons for this (e.g., Ahn et al. 2016; Kowalski and Limber 2007; Smith et al. 2008; Ossa et al. 2022). Among them, the greater distance between victim and perpetrator leads to dehumanization and decreased empathy of the perpetrators (Bertrand et al. 2018). Victims of cyberbullying also report that almost all areas of life are affected by the digital omnipresence, making it very difficult to distance oneself from the attacks (Was ist Cybermobbing 2018). It is particularly alarming that younger schoolchildren (5th–9th grade) and girls are affected more often than older children and boys (Ossa et al. 2022). Studies also found that victims are particularly likely to become perpetrators themselves (e.g., (Marcum et al. 2014; Ossa et al. 2022)). All of this underscores the importance of early intervention in a relatively young target group. Several studies have found that perpetrators showed less empathy toward victims than bystanders, the victims themselves, or people who had not yet been exposed to cyberbullying (Casas et al. 2013; Pfetsch et al. 2014). Even if there are opposite effects (de Almeida Mendes et al. 2021; Schultze-Krumbholz and Scheithauer 2009), increasing empathy plays a crucial role in the prevention of cyberbullying (Ingram et al. 2019). Most research has focused on perspective-taking (cognitive empathy) or the feeling of being worried (affective empathy) (Davis 1983). In addition, media-based empathy (cognitive and affective) has attracted considerable attention (Happ and Pfetsch 2015). It describes empathy in a media context and is intended to address both receptive and interactive media behavior in computer-mediated communication. Therefore, some prevention programs have included methods to promote empathy, such as role-playing (e.g., (Schultze-Krumbholz et al. 2018; Gu et al. 2022)). However, Nocentini et al. (2015) found that most prevention programs are not adapted to the digital reality of children’s and adolescents’ lives (Nocentini et al. 2015). This could also explain the moderate success of participating in prevention measures against cyberbullying (Ingram et al. 2019). Additionally, Beitzinger et al. (2020) showed that few schools systematically implement preventive measures and that the number of programs has been decreasing since 2017 (Beitzinger et al. 2020). A recent review shows that interventions on cyberbullying are somewhat more successful, but have still very rarely been studied independently of bullying (Polanin et al. 2021). Only one of the 50 studies used interactive immersive technologies in addition to videos. Researchers have been criticizing this simplified approach of integrating cyberbullying prevention into offline bullying prevention and reviews suggest that the mechanisms and consequences of cyberbullying differ from those of traditional bullying (Ossa et al. 2022; Gaffney et al. 2019). Additionally, the internet has created new ways of bullying, one of which is the emancipation from any location and the consequent intrusion into the homes of children. Taking those scientific results into account, the current work explores a VR-based prevention of cyberbullying that focuses on increasing empathy and the digital reality among the young target group.

2.2 Virtual reality (VR) as empathy machine

A recent meta-analysis by Ventura and colleagues (2020) showed that VR interventions lead to positive changes in empathy (Ventura et al. 2020). They found a moderate and significant effect size related to perspective-taking (cognitive), whereas VR interventions did not significantly affect other outcomes for empathy (affective). However, Ventura et al. (2020) included only a small number of studies (N = 6) and neglected studies that examined the Proteus effect (Ventura et al. 2020). The term refers to the phenomenon of people adapting their behaviors and attitudes to the characteristics of their avatars (Yee and Bailenson 2007). A recent meta-analysis showed that small to medium effect sizes emerged relatively consistently across 46 experimental studies (Ratan et al. 2020). The results were shown in various contexts and were mainly based on the sense of virtual body ownership. Self-representation, however, serves as only one example of VR’s potential to support empathy, behavioral, and attitudinal change. Almost every factor can be controlled and systematically changed in VR—the representation of the environment, of others, and virtual objects—all of which can also directly address psychological barriers and aid in attitude and behavior change (Wienrich et al. 2021). There is also ample evidence for this in various other contexts (representation of the environment and objects: (Hein et al. 2021b, a) and representation of others: (Latoschik et al. 2019; Wienrich et al. 2018a)). In the context of bullying, so far, only a few have attempted to use VR (bystander perspective: (McEvoy 2015); switch roles: (Gu et al. 2022; Oyekoya et al. 2021); and general considerations: (Carmona et al. 2011; Ingram et al. 2019)). These studies show positive results concerning perspective-taking and interest in VR of the target group. Additionally, VR offers practical benefits like the ability to quickly adapt the intervention in various ways, making it portable and individual. It offers privacy while engaging in difficult psychological situations and is in general easier and faster to set up than other digital or real-life interventions that need to be personalized to this degree. Ingram et al. (2019), for example, showed 7th and 8th grade students various bullying situations using the Google daydream headset. Their path analyses found that the VR intervention increased empathy and impacted, mediated through empathy, other outcome variables (e.g., readiness to intervene) (Ingram et al. 2019). However, this study shows limitations in that it does not exploit most of the above-mentioned potentials of VR, as the participants could only observe the scenes. Furthermore, the authors themselves criticize that “[...] future research should more closely examine the unique contribution of the virtual reality component specifically, role of psychological distance in bullying, [...]” (Ingram et al. 2019, p. 81). The other studies also focused on role play in the context of school bullying but neglected specific aspects concerning cyberbullying (McEvoy 2015; Gu et al. 2022). They also do not address the psychological distance through which prevention programs could develop promising effectiveness (Borbé et al. 2009). Therefore, we would like to analyze the possibilities of addressing psychological distance with VR potentials in more detail.

2.3 Addressing psychological distance with the personal space invasion

As previously described, psychological distance plays a special role in cyberbullying because it is even greater between victims and perpetrators than in traditional bullying. Not only their physical locations can be different, but the cyberbullying can also be asynchronous while the previously mentioned online disinhibition effect further distances victim and perpetrator emotionally. The Construal Level Theory considers mental representations of persons, objects, and events as a function of psychological distance (Liberman and Trope 1998). It is a social psychology theory describing the relationship between psychological distance and the extent to which people think about objects and events. The core idea is that the farther an object is from the person, the more abstractly it is thought about, whereas the closer the object is, the more concretely it is thought about. Research on interpersonal interactions found that people being psychologically distant from one’s self were viewed in more abstract terms with tremendous consequences. A large psychological distance can lead to stereotyping, lower levels of empathy, or a decreased willingness to help (Loewenstein 1996; Pronin et al. 2008; Ingram et al. 2019). In interpersonal interaction in cyberspace, the psychological distance of offenders may be even greater. For example, research on toxic online disinhibition shows dissociative effects that can lead to more violence and depersonalization between sender and receiver (Suler 2004; Bertrand et al. 2018). Reducing the psychological distance of (potential) offenders can, therefore, be a goal and mechanism of action in the design of prevention measures (Ingram et al. 2019). Numerous studies have already impressively shown that VR can contribute to such a reduction in other contexts (for an overview, see, e.g., (Ventura et al. 2020; Ratan et al. 2020)). However, reducing psychological distance can also lead to cognitive dissonance in the short term. This refers to opposing cognitions or attitudes that can lead to discomfort and stress. Resolving this uncomfortable dissonance can trigger an attitude change (Borbé et al. 2009). This means that by experiencing the victim role, a perpetrator not only increases his or her empathy but also changes attitudes about cyberbullying and, by integrating the new cognition, stops bullying (Stice et al. 2015, 2019). In other words, if the change of perspective is a prevention component, then negative feelings and distress should rise directly during the intervention. Unfortunately, the previous studies do not measure negative emotions and distress in addition to empathy as an indicator of cognitive dissonance. Reports from victims, on the other hand, indicate that cyberbullying, in particular, makes it difficult to distance oneself from the attacks. Psychologically, one would speak here of an invasion of personal space. Therefore, PSI as a term is not only a metaphorical description of a victim’s feelings but also refers to a proven psychological concept which was coined by Felipe and Sommer (1966). It encompasses all forms of violation of assumed norms of distance (Felipe and Sommer 1966). Four zones are distinguished: intimate space (1.5 ft or 0.45 m), personal space (4 ft or 1.2 m), social space (12 ft or 3.7 m), and public space (25 ft or 7.6 m) (Hall 1963). Studies have shown that discomfort and stress occur when the intimate or personal space is violated (Bönsch et al. 2018; Hayduk 1978; Middlemist et al. 1976; Smith and Knowles 1978; Sundstrom and Altman 1976). This has also been replicated in VR (Wilcox et al. 2006; Hecht et al. 2019). Again, the degree of discomfort was determined as a function of proximity to virtual avatars and agents. Thus, PSI could be a promising measure to make the victim’s perspective accessible. Unfortunately, there is no research on whether and how PSI can be used in the context of cyberbullying. Therefore, we explore whether and how PSI can increase empathy for cyberbullying victims.

2.4 Summary and present work

The previous work on cyberbullying has shown that the number of victims is increasing, and the need for prevention is particularly high among younger school students (5th–9th grade) (Feierabend et al. 2022; Nolden 2021). However, there is a gap between the need and existence of innovative digital prevention programs. Our work is inspired by numerous studies showing that VR can facilitate perspective-taking, attitudinal, and behavioral changes in various domains. Initial positive findings for its use in increasing empathy in the area of bullying also indicate that VR can present a prevention tool tied to the digital reality of the target group. This paper explores the feasibility and effects of a unique VR-based cyberbullying prevention component and addresses the following research questions:

  1. 1.

    How to design a PSI component in virtual reality to be usefully applied in the context of cyberbullying prevention?

  2. 2.

    Does our VR-based component lead to cognitive dissonance, a reduction in psychological distance, and prevention-promoting factors such as empathy and willingness to intervene?

  3. 3.

    Do children (8th grade) understand the metaphor of PSI in the context of cyberbullying?

  4. 4.

    How do children evaluate the VR-based cyberbullying prevention component?

Our work differs crucially from previous research in that we investigate the specific effect of current and interactive VR components in the context of psychological distance in particular, explore the design space for PSI in the context of cyberbullying in a human-centered iterative approach, and provide empirical evidence from a laboratory experiment as well as a field study with a large sample of the target group of children and adolescents.

Our work is presented in three sections according to the two human-centered design circles mentioned above. The first section presents the pre-study exploring the design space. The second section presents the laboratory study. The third section represents the field study. The surveys in all studies are based on an ethics vote from the local university ethics committee (file number 130123) and on consent from participants or parents and the schools’ administration.

3 Pre-study: design space and implementation

First, we explored to what extent we can transfer the psychological phenomenon of PSI to the design space of virtual reality as well as to the context of cyberbullying. The basic idea was to design virtual objects in a way that they symbolize cyberbullying and threaten the personal space. Thus, in the following pre-study, we focused on investigating if and to what degree specific object configurations lead to the perception of a personal space invasion.

3.1 Method

3.1.1 Participants

To gain first insight, we conducted a qualitative study with ten participants. All participants are university students with little to no experience with VR systems and belong to the age group of 21–29 years.

3.1.2 Materials and procedure

Participants passed through a set of eight VR scenarios, developed to test certain influencing factors on the perception of PSI.

To intrude into personal space, the objects appeared at a distance of 0.5 virtual meters to the subjects, a distance that has shown to induce discomfort (Wilcox et al. 2006) and was originally defined as personal space by Hall (1963). In our pre-study, we tested whether we can replicate these results with objects common to online communication such as text boxes and emojis. In order to find the best fit, we tested several separate factors within the conditions that could have a potential impact on the perception of a personal space invasion. To make sure that the attributed discomfort was experienced due to the personal space invasion and not only due to offending texts, we tested the text boxes in a separate condition from the violating text messages. Those factors included different textures related to cyberbullying (emojis and text signs), coloring (emojis: white, red, and yellow; text boxes: black and white), valence (emojis: no face, angry, or sad face; text: empty boxes or offending messages); and sound or removeability (excerpt from the design space, see Fig.  1).

Some of the factors have already been proven to have an impact, like noise (Lewis et al. 2017) and human-likeness (Bailenson et al. (2001); assigned to the emoji objects), the rest we chose to test because of their potential matching with the cyberbullying context. The VR scenarios were embedded into a test application, developed using Unity3D in the version of 2019.4.1f (LTS), deployed on an HTC Vive Pro connected to a Gaming-PC MSI MS-7977. The application allowed one to rate the affective effect of each scene directly after exposure using an In-VR 7-point Likert scale, measuring the feeling of discomfort, feeling of tightness, feeling of personal dismay, and feeling of joy (used as a control value). Participants should name the value corresponding to the Likert scale during use (ranging from 0 not at all to 6 very strongly). The respective scenes were played in a randomized sequential order. Participants used the two standard HTC Vive Pro controllers to control the application. Additionally, participants were told to use the thinking-aloud method to collect qualitative data on how certain design decisions affected emotional responses. To further investigate how the metaphor of PSI was experienced and adopted, a subsequent semi-structured interview was performed. To analyze the generated data, a content analysis of the transcribed interview audio recordings, as well as an evaluation of the metric data of the In-VR ratings, was conducted afterward.

Fig. 1
figure 1

The VR application used to explore the design space of a virtual personal space invasion in the context of cyberbullying from the user’s perspective. Left: yellow and red balls with angry (red) or sad (yellow) emojis. The white sphere represents the neutral variant, middle: Empty text boxes designed neutral in black or white, and right: cyberbullying-related text messages. In the third picture, you can also see the different distances used later in the laboratory study. Either the objects appeared directly around the subjects (0.5 ms) or further away on the wall (3 ms) of the room

Full size image

3.2 Results and discussion of pre-study

Quantitative results showed that the scenarios with negative speech texts hovering in the personal space and bright non-removable message boxes in the personal space had the highest effect on feelings of discomfort. Overall speaking, the results showed that the intended personal space invasion outcomes, such as feelings of discomfort and tightness, reached a medium level on average for five and three of the designed scenarios, respectively. Drawing on the qualitative results of the pre-study, assumptions for the design of the personal space invasion in the context of cyberbullying were developed. Additionally, the improved personal space invasion component was sought to be integrated in a full prevention-oriented VR application. As a result of the pre-study, a VR prevention application is developed for the standalone VR device Oculus Quest using the Unity3D Version 2020.2.2f. To embed PSI into a full prevention application, a storyboard has been developed which divides the VR experience into three phases: setup, conflict, and resolution. This setup borrows aspects of classical film structure based on research of field (2005). In the setup phase, subjects were able to familiarize themselves with the VR environment. They were given the task to “decorate” the virtual room they were in by hanging individualized pictures on one wall. During the conflict phase, a potential invasion of the personal space is triggered by intrusion objects designed to match the most effective design factors resulting from the pre-study data. Those are the usage of mean text phrases confronting the user, big white and non-removable text boxes, sound effects, and the adding of social cues like abstract profile pictures. The objects appeared at a distance of 0.5 virtual meters to the subjects like in the pre-study. Subsequently, the resolution phase presents an opportunity for the users to reflect on what happened during the conflict phase and also offers different options for (re-)action for cyberbullying encounters. Subjects are offered five options for action as a cyberbully victim as well as a bystander on a virtual canvas, from which they can choose the one option which they would most likely adopt themselves. Based on the promising results of the pre-study and using the iteratively developed VR prevention application prototype called TooCloseVR mentioned above, a consecutive laboratory study is conducted with university students.

4 Laboratory study with university students

The laboratory study focuses on the evaluation of the application TooCloseVR due to its effects on indicators of PSI, cognitive dissonance, empathy, and willingness to intervene. In order to do this, it is investigated whether PSI is able to reduce psychological distance in the context of cyberbullying prevention and whether it, therefore, has a significantly higher effect on the desired outcome variables in comparison with a control condition without the usage of PSI. The measurement took place three times to assess pre-, post-, and long-time effects.

4.1 Method

4.1.1 Participants

In total, 41 participants aged between 18 and 38 years (M = 23.37, SD = 4.66), including 27 women and 14 men, completed the study. The participants, who are mostly university students (highest degrees of education: high school diploma: 29; bachelor’s degree: 1; and master’s degree: 8), have been recruited using the university’s internal participant recruitment system and have been credited with either one experiment hour necessary for a student course or a 10 euro-voucher for a local bookstore. Eight of the participants had no prior VR experience, 17 some previous experience, eight moderate previous experience, and five extensive previous experience.

4.1.2 Material

The implementation of the VR application is already described in Sect. 3.1. To test the personal space invasion component, intrusion objects described in the mentioned section are spawned in a 0.5 virtual meter radius around the subject. In the respective control condition, the general setting is adopted from the treatment condition except for spawning the intrusion objects outside of personal space (in a three virtual meter radii around the subject). Exposure time of the spawning cyberbullying messages lasts 4 min in both conditions. In total, the VR application is designed to last for about 10 min, depending on the individual interaction time of subjects during the set-up task (see 3.1.), interaction with action options, and the processing of the during-exposure measurements. All during-exposure measurements are collected via an In-VR questionnaire which could be filled in via raycasting selection of a 7-point or 9-point Likert scale. After-exposure measurements have been conducted using the online questionnaire tool SoSciSurvey. The debriefing material has been deposited in paper form, clearly visible in the laboratory room. Debriefing material has also been used as behavioral measurement as described below. We clustered our measures into five groups.

Measuring demographic data. Data on age, gender, the highest degree of education, current occupation, and previous experience with VR systems have been collected before the exposure. Additionally, self-compassion using the German short form of the Self-Compassion Scale (Hupfeld and Ruffieux 2011) and empathy as a personality trait using the German version of the Interpersonality Reaction Index (Paulus 2009) are evaluated.

Measuring quality and manipulation check. To ensure a safe interaction in VR, we controlled for simulator sickness by assessing the German version of the Simulator Sickness Questionnaire (Kennedy et al. 1993) as well as controlling for the sense of presence with the German version of the Igroup Presence QuestionnaireFootnote 1 using the subscales of “general presence,” “spatial presence,” and “involvement.” Similar to the pre-study, we measured the sense of tightness and intrusion (manipulation check) by two items on a 7-point Likert during exposure (t2) using the In-VR interactive survey. The scale measured how much “tightness” and “intrusion” have been felt.

Measuring cognitive dissonance. As indicators of cognitive dissonance, personal distress was assessed with seven items immediately after the VR intervention at all three measurement time points (see further descriptions below): (Batson et al. 1997). Furthermore, we captured discomfort (“At the moment I feel uneasiness.”) and joy (“At the moment I feel joy.”) during the VR sessions on the 7-point Likert scale. To assess affective feelings immediately after the intervention, subjects indicated which of the following emotions—and with which intensity (10-point Likert scale, ranging from 1 low to 10 high)—they were currently feeling: fear, sadness, loneliness, anger, shame, guilt, helplessness, or no emotion. Subjects provided this information only once immediately after the intervention (t2).

Measuring empathy. We assessed the cognitive and affective aspects of (media-based) empathy. Affective empathy was captured by the Empathic Response Scale (Batson et al. 1997) after subjects were handed out fictive gender-controlled (Veenstra et al. 2010) cyberbullying scenarios designed to fulfill criteria raised by Olweus (2003) and matching the situation shown in VR. Note that we used personal distress as an indicator of cognitive dissonance since the items reflect discomfort and stress. Empathetic concern was applied as an indicator of affective empathy. Items have been translated by one of the authors into German and translated back into English by another author. Affective media-based empathy as well as cognitive media-based empathy was measured by Media-Based Empathy (Happ and Pfetsch 2015). For completeness, we collected all three subscales (“media sympathy,” “affective,” and “cognitive media empathy”). However, only the subscale cognitive media empathy seems to apply to the cyberbullying context since it specifically asks to what extent the respondent thinks about the effect of their posts in chats and blogs and what they trigger in others.

Measuring attitudes and behavior. Willingness to intervene in cyberbullying situations is measured according to the Willingness to Intervene in Bullying Situations Scale (Chen and Jin 2020). To meet the goal of our study, the originally 12 bullying situations are shortened and transferred into three self-designed cyberbullying scenarios matching the criteria of school bullying (Olweus 2003) and are gender-controlled again. The scale ranges from 1 (not agreed) to 4 (totally agreed). To further assess the potential influence on bystander behavior, participants are asked to recall the actions shown for options in open text fields after the exposure and in the long term. Six and five options were presented concerning self-help and bystander behavior, respectively. Last but not least, functioning as behavioral measurement, the subjects’ decision to take one or two of the debriefing flyers was noted.

4.1.3 Procedure

The laboratory study is structured according to the pre-post-follow-up study design of Pissarek and Wild 2018. This study method is profoundly efficient for measuring the effectiveness of a certain treatment or training over time. The first measurement point of time (in the following t1) measures a baseline of the desired outcome variables of personal distress, empathy, and willingness to intervene, as well as the assessment of demographic data. Some days after assessing the pre-questionnaire, but at least 1 day after, participants are invited into one of the university’s laboratories to be exposed to the VR application. The second measurement point of time (in the following t2) assesses the change in outcome variables after the VR exposure. One week after exposure, the follow-up survey (in the following t3) is conducted to measure the long-term effectiveness of outcome variables. T1 and t3 are conducted online, only t2 is conducted in presence. Participants are randomly assigned to the PSI (intrusion objects within the personal space) and noPSI (intrusion objects outside the personal space) groups by the application itself (Fig. 2). After fulfilling the last questionnaire, all participants receive a debriefing document.

Fig. 2
figure 2

Screenshots of the developed prevention application TooCloseVR used in the laboratory study. Both shots are from the user’s perspective, showing the PSI Condition on the left and the NoPSI Condition on the right

Full size image

4.1.4 Design and analyzing strategy

We have a two-factorial design with the factors time point (t1, t2, and t3) and PSI manipulation (noPSI and PSI). Depending on how many measurement time points were available for which measure, either a mixed analysis of variance with repeated measures was performed for the within factor time point (t1 vs. t2 vs. t3 or t1 vs. t2) and the between factor PSI manipulation (noPSI vs. PSI) as well as the corresponding post hoc tests. Significant results were based on an alpha level of 0.05, and a trend was reported up to an alpha of 0.1. Note that non-significant results are not statistically described in more detail to facilitate the readability of the results. All statistical test results can be provided in supplemental material.

4.2 Results

4.2.1 Impact of demographical data

Within the sample, self-compassion showed a normal distribution, while the empathy trait is not normally distributed (\(W=0.92,\, p<.01\)) but left-skewed. However, demographic data were evenly distributed across groups and, therefore, are not used for further moderation analysis.

4.2.2 Quality and manipulation check

All participants included in data analyses reported low values of simulator sickness, and no statistically significant differences between the noPSI and the PSI group were found. However, the noPSI group reported more simulator sickness descriptively and showed higher variances. Furthermore, participants reported medium presence values (M = 3.71; SD = 0.59), differing by trend between the noPSI and the PSI group with higher values for the latter (\(F(1,42)=3.60,\, p=.07,\, p\eta ^2=0.08\)). As expected, the manipulation check revealed that the application created a sense of tightness (t1 vs. t2: \(F(1,42)=51.74,\, p<.01,\, p\eta ^2=0.55\)) and intrusion (t1 vs. t2: \(F(1,42)=78.35,\, p<.01,\, p\eta ^2=0.55\)). Although, no main effect of the PSI manipulation occurred, an interaction effect tends to emerge concerning tightness (\(F(1,42)=3.35,\, p=.07,\, p\eta ^2=0.07\)). The feeling of tightness immediately after the intervention (t2) is greater in the PSI than in the noPSI group (see Fig. 3).

Fig. 3
figure 3

Perceived tightness (left) and intrusion (right) before the intervention (t1), and directly after it (t2) for both conditions of personal space invasion (PSI)

Full size image

4.2.3 Signs of cognitive dissonance

All indicators of cognitive dissonance showed a significant increase between t1 and t2 (personal distress: \(F(1,42)=4.23,\, p=.02,\, p\eta ^2=0.10\); discomfort: \(F(1,42)=56.00,\, p<.01,\, p\eta ^2=0.67\); and absence of joy: \(F(1,42)=83.86,\, p<.01,\, p\eta ^2=0.67\)). The PSI manipulation had no significant influence on these measures, and no interaction effect occurred. Since personal distress was also captured at t3, the results showed that distress returns to the initial level with a time delay after the intervention (Fig. 4). In addition, subjects in both groups reported that helplessness was the most common emotion directly after the intervention, followed by sadness and anger. Here, mean intensity scores were reported in both groups (PSI: M = 4.35, SD = 2.27 and noPSI: M = 5.00, SD = 1.79). The PSI manipulation did not produce differences in this regard.

Fig. 4
figure 4

Personal distress (left) and cognitive media empathy (right) before the intervention (t1), directly after it (t2), and 1 h after it (t3) for both conditions of personal space invasion (PSI)

Full size image

4.2.4 Signs of empathy

As expected, the cognitive media-based empathy increased significantly and linearly from t1 to t2 and t2 to t3 (\(F(1,42)=17.24,\, p<.02,\, p\eta ^2=0.32\)). Although the PSI manipulation had no significant influence (\(F(1,42)=1.98,\, p=.17,\, p\eta ^2=0.05\)), descriptively higher values were measured in the PSI than in the noPSI group (Fig. 4). In contrast, the affective (media) empathy measures did not differ between time points or PSI conditions.

4.2.5 Signs of attitude and behavior change

Neither the intervention as a whole nor the PSI manipulation significantly affected willingness to intervene. The recall of intervention options showed no significant difference between groups or type of options for action (self-help or bystander behavior). Directly after the intervention, however, subjects showed a medium level of recall of the, respectively, six or five presented options in both intervention groups concerning self-help (PSI: M = 4,08, SD = 1,29 and NoPSI: M = 3.61, SD = 1.38) and bystander behavior (PSI: M = 3.20, SD = 1.12 and NoPSI: M = 3.22, SD = 0.81). After 1 week, subjects still remembered on average half of the given action options concerning self-help (PSI: M = 3.22, SD = 0.81 and NoPSI: M = 3.09, SD = 1.2) and bystander behavior (PSI: M = 2.44, SD = 1.04 and NoPSI: M = 2.57, SD = 1.24). Subjects generally took little informational material with them, regardless of their group.

4.3 Discussion of laboratory study

Overall, the application showed good quality. The slightly higher values of simulator sickness in the noPSI condition could indicate that the nearer objects served as anchors (Wienrich et al. 2018b, 2022). Generally, simulator sickness scores were similar, and no subject dropped out of the intervention. The intervention as a whole had the expected impact. A sense of tightness and intrusion was induced, and indicators of cognitive dissonance and cognitive empathy were produced. Thus, we could demonstrate for the first time that the basic idea of the VR-based intervention leads to the desired effects of emotional tension as a sign of cognitive dissonance and perspective-taking using cyberbullying specific intrusion objects. Aspects are important to elicit better understanding for victims. On the other hand, the VR intervention did not increase affective empathy. This is consistent with Ventura’s review showing that VR is more likely to improve perspective-taking (cognitive empathy (Ventura et al. 2020)). In another study using 360-degree video, cognitive and affective aspects were not separated (Ingram et al. 2019). Like Ingram et al. (2019), but in contrast with Gu et al. (2022), there was no direct influence on behavioral measures (Ingram et al. 2019; Gu et al. 2022). Contrary to expectations, the PSI manipulation hardly had any significant impact. Already the results of the manipulation check suggest that the noPSI group, too, found the messages to be harassing. Although they were shown outside the intimate or personal space (Hall 1963), they were shown inside a closed room, which was also personalized beforehand, so it is possible that the whole room was perceived as personal space. The qualitative data support the assumption. For example, one participant said: “[...] Because it’s gone up a little bit in discomfort. Although the sentences are now so far away from me. So before, they were so close around me, and now they are so far away, but somehow I almost have the feeling that it is more uncomfortable for me because they are just everywhere, and I don’t have the feeling that I can step out of this small circle with the sentences into the neutral space, but they have taken over the space, these statements, they have become the space, that is why the space doesn’t feel so safe anymore [...]”. Such and similar statements indicate that the messages in the room were perceived as intrusive, regardless of distance. From the therapeutic point of view, it is important for the future design of the intervention that the personal distress and negative emotions return to the initial level some time after the intervention (Borbé et al. 2009). This ensures that participants can sufficiently distance themselves from what they have experienced and are open to discussion and further engagement with the topic. This is also particularly important since the program is to be used with children and young people directly in schools, where rapid recovery is a key requirement.

5 Field study: feasibility for schoolchildren

Because the intervention showed initial promising results but was developed and tested with individuals slightly older than the target population, the second iteration was conducted directly with children and adolescents.

5.1 Focus group to improve the implementation

A focus group of five children and adolescents (three females, two diverse and age: M = 16.8, range from 13 to 19) reviewed the previous application’s fit for the target group. Feedback was used to adapt the application to the specific needs of the target group. A children’s/adolescent’s room was designed in which subjects were asked to watch an educational video on the topic of German history on a modulated screen in VR (see Fig. 5). They could also interact with a smartphone on which messages first appeared before before they were later popping up in the room among subjects as before (either in personal space or outside). The messages were adapted to the current youth’s vernacular and, in terms of content, to the reality of life of the target group (messages, see excerpt in Table 1 and all 90 in the supplemental material). During the intervention’s first half, users could click away the messages via raycasting. These messages did not disappear, however, but piled up on the bed to demonstrate the permanence and impact of digital messages. In the second half, the scope of action was further restricted, and the messages could no longer be clicked away. In addition, the Unity project was ported to the Pico UI (Version (PUI) 4.6.3) because Oculus’ terms of use in Europe had changed so that they were no longer privacy compliant. As before, the project was built in Unity (Version 2021.3.3f1) and uses various assets (VR Keyboard,Footnote 2 Teenage Bedroom,Footnote 3 VR SmartphoneFootnote 4 and UX Flat IconsFootnote 5). Furthermore, a task was implemented, so the children had something to focus on. It was a video about the German Empire—a topic that is part of the curriculum for 8th grades in Germany.

Table 1 Examples of messages (N = 90) used as virtual objects translated to English
Full size table
Fig. 5
figure 5

Screenshots of the improved VR component used in the field study. Both shots are from the user’s perspective, showing the adapted room with the VR smartphone and display for the video students have to watch and on the left and the intrusion with cyberbullying-related messages

Full size image

5.2 Field study with school students (8th grade)

5.2.1 Participants

To explore the initial feasibility of VR prevention, we conducted a half-day workshop with 92 8th grade pupils (female = 34, male = 35, diverse = 7, no answer = 16; age: M = 13.81; SD = 0.46; no answer = 19). The workshop took place over 5 consecutive days with a group of 20–30 scholars. Each day, the groups were divided into three small groups of 7–10 participants. A teacher was present as a supervisor at all times.

5.2.2 Materials

The technical equipment is described above. The queries were performed with Microsoft Forms.Footnote 6 In general, the workshop design was based on existing prevention programs for bullying or cyberbullying and therefore included elements of (media) education, perspective-taking, and agreement on joint class rules (Gaffney et al. 2021; Ansary 2020).

Qualitative Measurements. Since this is the first feasibility test and the workshop took place directly in the school during a project week, we mainly collected qualitative feedback during group discussions. Here, we asked semi-structured questions about the understanding of the principles (e.g., “What do you think we wanted to achieve with this VR experience?”), the experience in VR (e.g., “What did you experience in VR?”), expectations (e.g., “What do you think is coming up next?”), and the workshop in general (e.g., “Did you like the materials, were they appropriate?”).

For the other measures, we mostly created one item that children could understand on a 10-point Likert scale (ranging from 0 not at all to 10 very much). Data were anonymously assessed through Microsoft Forms.

Measuring demographic data. Data on age and gender have been collected at the end of the workshop. The students also gave an anonymous self-disclosure about their knowledge and personal experiences with cyberbullying (13 questions, e.g., “I’ve been a victim of cyberbullying before”; “How many people (in %) have been affected by cyberbullying?”).

Measuring quality and understanding of principles. To ensure a safe interaction in VR, we controlled for simulator sickness by assessing the fast motion sickness scale (Hecht et al. 2019), measuring the sense of presence with one item (“Did you feel like you were really in that teenage room?”) and assessing the mental strain with one item (“How mentally demanding was the task?” adapted from the NASA-TLX (Flägel et al. 2019)). To evaluate how much the students were distracted from the main task (watching an educational video), we asked how much they were able to focus on the video. The further understanding of the VR intervention principles was assessed by qualitative data (see above).

Measuring experience. Similar to the laboratory study, we measured the sense of tightness (manipulation check) by one item. As indicator of cognitive dissonance, the perceived tension was assessed (“How tense do you feel right now?”). Further, the students anonymously indicated how they were currently feeling. The words happy, angry, ashamed, sad, guilty, carefree, lonely, helpless, stressed, and relaxed were offered as support. Following the laboratory study, we collected one item on cognitive (“Can you imagine how it feels to be bullied online?”) and affective (“How close does the issue of cyberbullying get to you?”) empathy, respectively. Finally, the willingness to intervene in cyberbullying situations was measured (“Have you ever thought about helping someone who is experiencing cyberbullying?”).

Measuring quality of workshop and suggestions. In addition to the qualitative data (see above), workshop quality was assessed by questions about the satisfaction (“How happy were you with the workshop?”), usefulness (“Participating in the workshop was extremely useful for me.”), learning effects (“I now know a lot more than before about cyberbullying and how to protect myself from it.”), and transfer effects (“What I learned today I can use very well in everyday life.”).

5.2.3 Procedure

After the researchers’ introductions, they presented the agenda and established rules for the group together with the students. This was followed by an introduction to the topic of cyberbullying through a video of a case example (Lijana Kaggwa, a well-known former candidate of the TV-show Germany’s Next Top Model) and to the forms and consequences of cyberbullying. Depending on which group they belonged to, the students completed the following three program points in different order and separate rooms in the small groups. (1) They were familiarized with VR since most of them had no experience with VR. The goal was to mitigate the novelty effect. The students played “VR Beginner—The Escape Room.”Footnote 7 We assessed their experiences with the VR game, indicators of simulator sickness, and their expectations for the upcoming program with qualitative questions. (2) They gave an anonymous self-disclosure through Microsoft Forms about their knowledge and personal experiences with cyberbullying. (3) They talked about joint rules in the class regarding cyberbullying prevention and common conduct. After a short break, a baseline measurement (t1) was performed anonymously by assessing the experience (current feelings, perceived tension and tightness, empathy, and willingness to intervene). Then, the results of the anonymous self-reports were presented to the whole class and placed in the context of existing scientific findings. We also asked about their VR experience and expectations about the upcoming session. After this session, the VR prevention components continued in the small groups. In each small group, the intervention could be experienced in parallel. Trained researchers were present at all times. This was followed by a second measurement (t2). The students anonymously answered the same questions as in the pre-measurement. In addition, the small group discussed what they experienced in VR, how they interacted in VR, whether they could remember the content of the messages, and how well they had been able to concentrate on it during the intervention. Afterward, a group evaluation was conducted with the whole class. Here, the researchers once again asked more specifically why the students thought that the VR room had been designed the way it was, to what extent the content fit the theme of cyberbullying, whether they thought that it was an appropriate way to facilitate better understanding of victims, and whether the experience met their expectations. We also measured the perceived tension and tightness to ensure that the desired negative affects decline until the end of the workshop (t3). We then debriefed the students about the intention and sense of the VR prevention component. After a short break, the class rules that were worked out in the small groups were summarized, and the class agreed together on rules to prevent cyberbullying in their class. All students signed these rules. This was followed by a media literacy session in which privacy settings in particular were discussed. Finally, the steps that can be taken if one is affected or observes cyberbullying were discussed. After a final round of feedback regarding the entire workshop, we concluded the day. There was a trained children’s and adolescents’ psychotherapist on call throughout the workshop in case of complications. This contact, as well as the contact of special counseling instances, was distributed at the end of the workshop. There were no complications during or after the workshop.

5.2.4 Analyzing strategy

We formed mean values over the quantitative collected data and prepared the qualitative data in a content-analytical way. Responses were processed on a digital board.Footnote 8 Three raters grouped the responses inductively. From the results, the quality of the application and the understanding of the main principles, the emotional experiences, and suggestions for improvement can be derived.

5.3 Results of field study

5.3.1 Quality and understanding of principles

Similar to the laboratory study, the VR experience did not lead to an increase in simulator sickness. The sense of presence was in a suitable medium range (M = 5.64, SD = 2.46), and the perceived strain during the intervention was low to moderate (M = 3.71, SD = 2.71). The discussion with the students showed that they understood the basic idea of the VR intervention. They stated that the messages came very close to them and that they felt helpless because they could not escape them. They also understood that the messages were piled up on the bed to symbolize that they would not simply disappear. Figure 6 shows some statements of students reflecting on their understanding of the intervention’s principles. In the questionnaire, they also stated that they could not concentrate on the video (M = 1.74; SD = 2.52, scale ranged from 1 to 10).

Fig. 6
figure 6

Some statements from the final group discussion regarding emotions (top) and understanding of principles (bottom)

Full size image

5.3.2 Experience

Similarly to the results of the laboratory study, perceived tension and tightness were significantly greater immediately after the intervention than before and 1 h after the intervention (tension: \(F(1,182)9.10,\, p<.01,\, p\eta ^2=0.09\) and perceived tightness: \(F(1,180)56.84,\, p<.01,\, p\eta ^2=0.39\), Fig. 7).

Fig. 7
figure 7

Perceived tension (left) and tightness (right) before the intervention (t1), directly after it (t2), and 1 h after it (t3)

Full size image

Figure 8 illustrates the students’ emotions during the three measurement time points. The qualitative results from the final group discussion support the statements (see Fig. 6). In addition, they showed a temporal change in affect concerning the emotions. It changed from happy, good before the intervention, to distressed, helpless directly after it. After a while, the feelings became more positive again (relaxed, normal).

Fig. 8
figure 8

From left to right, students’ emotions before the intervention, shortly after, and 1 h after

Full size image

5.3.3 Evaluation of workshop and suggestions for improvement

The students were very satisfied with the workshop (M = 8.50, SD = 1.65, range of scale 1–10), stating that it was useful (M = 6.11, SD = 2.39), they knew more about cyberbullying than before (M = 5.03, SD = 2.54), they learned a lot of new things (M = 5.63, SD = 2.45), and they were able to apply the knowledge in real life (M = 5.21, SD = 2.51). However, the students also suggested numerous improvements. Most frequently, they wished for more opportunities for action. Even though they understood that part of the intervention was that you could not click away the messages in the second half, they would like to be able to click away the messages, reply to them or mute them. They also wished that they could pause the educational video so they would not be so distracted. They also stated that the messages were still too “lax” and that personalizing them, or at least adapting them to their gender, would have made an even bigger impact.

5.4 Brief discussion of field study

Our feasibility analysis shows that schoolchildren aged 13–14 also understand what effects the VR-based intervention and PSI are intended to produce. The results also appeared similar to that of the laboratory study with young adults. A perception of tension and tightness was elicited, as well as matching negative feelings. It also showed that tension and negative feelings increased as desired shortly after the intervention but then returned to an acceptable level. This is particularly important when working with children and adolescents to conduct a post-reflection. In general, the workshop was considered very positive and suitable, and numerous improvements were suggested.

6 General discussion

This article explores the design space, feasibility, and effects of a unique VR-based cyberbullying prevention tied to the digital reality of the target audience (5th-9th grade students). The central idea is to create a virtual invasion of personal space with virtual objects associated with cyberbullying, making the everyday intrusion of victims’ lives tangible.

6.1 Concerning RQ 1: design space of personal space invasion

Exploration of the design space revealed that real messages in the context of cyberbullying in white text boxes best represented the idea of intrusion. The feasibility study showed that the messages should be adapted to the target group’s language to achieve the appropriate effect. It also became clear that children and adolescents are quite used to “harsh” messages and that the threshold at which a message is really experienced as hurtful is relatively high. This provides important clues for the design of prevention measures that tie in with the reality of young people. None of the selected messages was perceived as too bad or led to negative after-effects. However, they were able to trigger stress and negative emotions shortly after the intervention as desired.

6.2 Concerning RQ 2: impact of the VR-based prevention

We could detect indicators of cognitive dissonance in both the young adult laboratory study and the feasibility study with children and adolescents. Cognitive dissonance is often used in dissonance-based prevention programs to indicate perspective-taking and attitude or behavior change (Stice et al. 2019). In this context, opposing cognitions or experiences may trigger stress and discomfort (Borbé et al. 2009; Stice et al. 2015). This leads to increased personal stress and tension during the VR intervention and decreases immediately afterward. This progression is very relevant to using the VR intervention in dissonance-based prevention programs (Stice et al. 2019). First, it ensures that no negative side effects occur, and second, the cognitions and emotions experienced can be discussed after the VR intervention. In addition to general signs of stress, we could also show that specific emotions and cognitions can be triggered. In both studies, perceived emotional tension and tightness—i.e., the feeling of being emotionally constrained and bothered by the news—increased. Participants also reported a sense of helplessness, which is often triggered in victims of bullying/cyberbullying. This confirms the impression in our data that, as noted in the review by Ventura and colleagues, VR causes increased cognitive empathy (Ventura et al. 2020). Our data suggest that this effect is related to the generation of cognitive dissonance and emotional tension. Therefore, the VR intervention possibly increased cognitive empathy, i.e., perspective-taking, but not affective empathy. However, it is known from the literature that changing emotions generally takes significantly more time than changing cognitions (Kahl et al. 2012). Still, the results indicate that the psychological distance was reduced (Liberman and Trope 1998). This is an important prerequisite for addressing the severe consequences of cyberbullying in prevention programs (Loewenstein 1996; Pronin et al. 2008). Thus, the current study demonstrates how VR can directly link to psychological principles of action in VR (Wienrich et al. 2021). In the laboratory study, we did not find a significant increase in either attitudinal or behavioral measures, which is in line with the previous studies (Ingram et al. 2019). However, the baseline scores were already relatively high, so that a ceiling effect may have occurred. Most importantly, qualitative data from the children and adolescents show that the VR experience led to a greater willingness to intervene and protect others from cyberbullying. These findings suggest that our VR intervention may be appropriate for a prevention program to reduce cyberbullying.

6.3 Concerning RQ 2: impact of the personal space invasion

To investigate the effect of personal space invasion by objects associated with cyberbullying, we show messages inside and outside the personal space (Hall 1963). Contrary to expectations, there were few effects of PSI manipulation even though the literature has already shown that PSI works in VR (Wilcox et al. 2006). So far, avatars or virtual agents and not objects have been used for invasion, like in our study. As we discussed above, however, the lack of effect may also occur because the entire room was personalized, and thus, even messages further away had the corresponding outcome. Additionally, some measures showed ceiling effects. Both studies show that the metaphor itself was understood and appropriately judged by the participants.

6.4 Concerning RQ 3 & 4: feasibility in schools

Our feasibility analysis shows that schoolchildren aged 13–14 understand what the VR-based intervention and PSI are intended to represent. The workshop was considered very positive and suitable. The implementation worked without any problems and with little use of technology. There were also no negative side effects. The setup, therefore, seems suitable for use in schools.

6.5 Limitation and future work

As discussed above, although the metaphor for the PSI was understood, the manipulation itself did not show any effect because, if applicable, the entire space was perceived as personal space. This needs to be further empirically tested. Future studies could first determine the personal space directly in VR per person and thus adapt the boundary for an individualized PSI. Furthermore, the effect of personalizing the space itself could be empirically investigated. Moreover, the question arises whether virtual objects can trigger the intrusion at all or if it is limited to avatars or virtual agents. Related to this, the present study cannot make any statement about how other impact factors of an immersive intervention would work (Wienrich et al. 2021). For example, questions about how virtual embodiment could enhance perspective-taking or even affective components, as often shown in Proteus studies (Ratan et al. 2020), remain unanswered. Gu et al. (2022) demonstrated recently that perspective-taking in VR has positive effects on empathy and willingness to intervene in the future situations (Gu et al. 2022). They investigated typical bullying situations so that it might be interesting to combine the perspective-taking with the intrusion of cyberbullying-related objects. There are not that many VR studies that include children and adolescents yet. However, recent work indicates that stress and emotions affect behavior directly after VR interventions (Lambe and Craig 2022). Further mechanisms should be explored in the future studies. In addition to testing feasibility, a controlled comparative study should be conducted compared to other measures such as the Bullying Questionnaire recently developed by Gu et al. 2022. Here, individual aspects such as previous experience might also be addressed. Concerning comparability of the measurements in the different test cycles of the current study, we could not repeat the entire test battery in the school setting, since study two was conducted as part of project weeks. For the field study, we, therefore, picked prototypical questions from the established questions. We aimed to provide at least a basic comparability of the concepts.

Finally, qualitative feedback from children and adolescents shows that they also believe that teachers and parents are not well informed and do not understand their digital reality well. Therefore, exploring appropriate modules for parents and teachers would also be necessary.

6.6 Conclusion and contribution

The current study addresses intervention and prevention needs in the context of cyberbullying by exploring a VR-based intervention in the laboratory (41 participants) and the field directly with the target group (92 8th grade students). The application shows promising results in terms of psychological impact and appropriateness for the target audience. This is particularly important given the enormous and concerning number of cyberbullying victims and particularly dire consequences among children and adolescents.

During the iterative development, the target group was involved in adapting the measure to the digital reality of the target group. The close collaboration between clinical psychologists for children and adolescents and HCI researchers ensures, on the one hand, the quality (e.g., low simulator sickness) and, on the other hand, a psychological foundation for the expected chain of effects. VR has great potential that goes beyond the initial “wow effect” which should be empirically proven to enable further acceptance of serious applications.