Introduction

While cyberbullying as a term has a long history with the first studies dating to the nineties [1], it remains to this day a topic of contest regarding its classification and importance. A recent review that was limited to the timeframe between 2015 and 2020 and to adolescent populations found a wide variability in findings, with prevalence rates of cyberbullying preparation ranging from 6.0 to 46.3%, and the rates of cyberbullying victimization ranging from 13.99 to 57.5% [2]. These incompatible findings point to the unresolved difficulties in researching the construct, difficulties stemming from a lack of a commonly agreed upon definition, the nearly complete employment of secondary education student samples in related research and the co-occurrence of traditional bullying.

The term “cyberbullying” itself is not standardized with researchers using a variety of alternate terms including “online bullying/harassment,” “cyber-aggression,” and others, leading to some confusion with actual instances of cyber stalking and cyber harassment [3]. While cyberbullying was viewed as an extension of bullying practices with different means, its definition has also been contested and revisited continuously, with a conciliatory proposal put forward by Tokunaga [4] who defined cyberbullying as “any behavior performed through electronic or digital media by individuals or groups that repeatedly communicates hostile or aggressive messages intended to inflict harm or discomfort on others,” emphasizing the intrusion into personal space.

A significant confounder with the perception of cyberbullying is that related research was heavily skewed toward primary and secondary education. In these younger ages cyberbullying more often than not is an extension of non-cyberbullying victimization. However, higher education is not devoid of cyberbullying, as shown in a recent review [5] that highlighted that cyberbullying involves social media and that undergraduate students are afraid to report it. Even among the first reports of this kind, it was established that a significant percentage of victims were cyberbullied for the first time while in college without a link to direct bullying [6]. A comparative study of bullying versus cyberbullying incidence published in 2016 estimated that 20–25% of students reported non-cyberbullying victimization in college and 10–15% reported cyberbullying victimization [7].

Although some high profile cases, that have led to the victim’s suicide, have propelled the term into the mainstream [8], results from prospective studies following students over a long time frame have led prominent researchers to conclude that cyberbullying was a low-prevalence phenomenon that cannot be viewed outside the context of traditional bullying [9] but rather as a subcategory or specific form of bullying[10], with a smaller incidence than the other forms and not as pronounced as was originally considered to be [11]. However, these reports date back to 10 years ago, a timeframe that in the context of our digital era appears very dated. A recent systematic review and meta-analysis [12] found that victims of cyberbullying were three times more likely to present with depressive symptomatology compared to controls and while the presence of traditional bullying moderated this relationship, it did not negate it.

Internet use disorder (IUD) is an umbrella concept that includes all aspects of problematic interaction with internet-related activities, and much as cyberbullying, it has been first described during the nineties [13] and has been mired in controversy ever since [14, 15]. Internet use involves many diverse activities, of which online gaming has received the most scrutiny, with a working definition of online gaming addiction offered in the latest version of the Diagnostics and Statistical Manual of the American Psychiatric Association, DSM-V [16]. This inclusion led to several dissenting views and stirred controversy [17], ultimately helping the field progress enough [18] for the World Health Organization to include Gaming Disorder (GD), either offline or online, as a separate disease entity in the latest version of the International Classification of Diseases, ICD-11 [19]. Other types of specific internet use disorders are identified and proposed as separate entities, including social media use disorder [20] and online pornography use disorder [21]. Prevalence of IUD in its various forms typically has a wide margin of error with a 2020 review [22] of studies worldwide, reporting a weighted average prevalence rate of 7.02% (95% confidence interval: 6.09–8.08%) and 2.47% (95% confidence interval: 1.46–4.16%) for IUD and GD respectively. These statistics are reportedly on the rise following the COVID-19 pandemic, as shown in a review of related studies [23], ascribed to a slew of factors that include financial hardship, isolation, problematic substance use, and mental health issues such as depression, anxiety, and stress.

IUD and cyberbullying now have two major staging environments in common: using social media and playing games online have become the most frequent choices of adolescents for communication and recreation [24], and this trend has increased during the COVID-19 pandemic to detrimental effect on their well-being [25]. Social media were recognized early on as frequent outlets for cyberbullying [26] while cyberbullying in online gaming communities appears to be an understudied issue. There is a small number of studies that point to the existence of cyberbullying within gaming communities [27,28,29,30] and this phenomenon is associated with the toxic culture prevalent in a number of gaming communities [31, 32] that persists despite the efforts from the game creators to reign it in [33]. This review aims to critically assess the published studies on the relationship between cyberbullying and IUD, and propose directions for further study.

Methods

Study Identification and Selection

Results from studies on cyberbullying and IUD published through March of 2022 were searched through the Scopus, ProQuest, and NLM/PubMed databases.

Because both the terms “cyberbullying” and IUD are not conclusively established, other interchangeable terms were added to the main keywords list. Accordingly, we screened studies through the combination of main keywords for both terms:

  1. (A)

    Main keywords for IUD: [Internet / online] and [Addiction / Problematic / Dependence / Excessive / abuse / compulsive / addictive / overindulgence / pathological / overuse / problem].

  2. (B)

    Main keywords for cyberbullying: cyberbullying / harassment / bulling / aggression / victimization

It is important to note in this point that the choice of keywords for cyberbullying did not necessarily relate to our understanding of the term but it was appropriate for a number of studies that did not necessarily agree with established terminology.

For the eligibility criteria, we set the following inclusion criteria: (1) studies should include a cross-examination of cyberbullying and IUD and not be limited to parallel reporting of incidence, (2) studies focused on either the cyberbullying perpetration, victimization, or both; (3) studies were journal articles in peer-reviewed publications; (4) studies were published in English, French, or German; (5) studies should have a clearly-defined research population; and (6) studies should describe original research work. Reviews, case studies, and case series were excluded from the search.

A total of 56 papers were identified electronically after duplicates were removed. Twenty-four were removed not adhering to the inclusion and exclusion criteria: one paper was a review, one paper presented a case report, one paper was a case series, eight papers focused on traditional bullying only, and eight papers were examining only IUD and not cyberbullying. The remaining 32 papers were included in the literature review. The procedures were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines[34]. The PRISMA flow diagram shows the detailed procedure (Fig. 1).

Fig. 1
figure 1

PRISMA flowchart of study selection

Full size image

Results

Included Studies

Table 1 presents the main points of the studies that were included in the review. Columns include study design, number of subjects, place, time frame of data collection (where available), age range of subjects, the measures that were employed to assess cyberbullying and IUD, and a brief outline of the findings.

Table 1 Results from studies carried out on cyberbullying correlated with excessive Internet use
Full size table

The very few prospective studies provide the most solid evidence. The very first survey of its kind by Yang et al. [38•] did not confirm any relationship between time spent online and cyberbullying. However, this survey was conducted in 2004 and did not include a valid measure of IUD, as it was not formally defined at the time. The survey by Floros et al. [35•] on 2008 found that the impulsivity subscale of the Online Cognitions Scale was a predictor of whether an adolescent victimized others online, although no associations were made with the severity or frequency of victimization. The third survey that was carried out on 2011 offered helpful results on both separate publications of its findings [36••, 41••]: cyberbullying victimization during the first point in time (T1) predicted depressive symptoms and IUD at the second point in time (T2) [36••] while IUD at T1 predicted an increase in the perpetration of cyberbullying and meeting strangers online at T2. The fourth prospective study that was carried out between 2018 and 2019 by Liu et al. [62••] added that the experience of cyberbullying victimization was positively related to IUD through the mediating variables of mindfulness and depression. Results from the cross-sectional studies confirm that there is a correlation between cyberbullying and IUD that may be mediated by a variety of factors; however, caution is required when treating results from cross-sectional surveys as indicative of causality, regardless of the statistical method that is employed to assess the data. In this instance, there are conflicting reports that are treated as conclusive findings despite the inability to assess directionality: depending on the viewpoint of the authors, IUD was either reported as being associated with cyberbullying [39, 40, 45, 47,48,49, 51, 52, 54, 55, 60, 61, 62••, 63, 66], moderating cyberbullying [42], mediating cyberbullying [57], predicting cyberbullying victimization [53, 65], or being predicted by cyberbullying [65].

Evaluation of Quality of Evidence So Far

There are several noteworthy findings when reviewing the relevant literature:

  1. a

    The vast majority of research is conducted on high-school students, with a handful of studies [43, 45, 48, 49, 56] expanding the scope to young adults, up to 25 years of age. There was a single clinical study of psychiatrically hospitalized adolescents [48]. Sampling high-school students has several advantages: it may be argued that the sample is representative of the population at large, since high school education is obligatory in all the countries where the surveys took place. Computing sample size and the potential responder number is simplified. If the survey is carried out during school time, then participation rates are high. However, the low number of clinical cases of either IUD or cyberbullying present in a population of this kind may lead to an underestimation of the severity of the negative impact. Surveys of this type presume a linear relationship between the studied variables (e.g., moderating variables in the relationship of cyberbullying and IUD). This presumption may not hold in clinical cases of IUD or serious forms of cyberbullying. This issue is amplified by the fact that there was no attempt in any study to quantify the severity of the impact of cyberbullying on the subject’s well-being.

  2. b

    The point in time of data collection was established for some after personal communication with the corresponding authors and remained unknown for a small number of studies [57, 59, 65]. Publication date may differ from data collection for as long as 7 years [55]. Data on occasion were collected as part of a larger survey with cyberbullying and/or IUD examined with a few items in a larger test battery [40, 54, 57].

  3. c

    The majority of studies employed ad hoc measures for cyberbullying, IUD, or both, despite the fact that validated measures for both constructs were available at the time that they were conducted [37, 38•, 39, 40, 42, 44, 46, 48, 52,53,54, 59, 64]. The specific items that were employed on those ad hoc measures are rarely mentioned. This renders study duplication, data aggregation, or comparison between studies impossible. Furthermore, cyberbullying was assessed in a large number of studies [35•, 37, 39, 40, 42, 44, 46, 48, 50, 54, 57, 59] with few items that could only relate to prevalence of its existence but not frequency or severity of its perpetration.

  4. d

    There were only four prospective surveys while the rest were cross-sectional. The prospective surveys were a survey carried out on 2004 by Yang et al. [38•], a survey carried out on 2008 by Floros et al. [35•], a survey carried out on 2011 and reported on two separate publications by Gámez-Guadix et al. [36••, 41••], and a survey carried out on 2018–9 and reported by Liu et al. [62••]. Unfortunately, despite the fact that some of the earliest publications correctly employed a prospective design, research has shifted to exploring associations between alternative psychological constructs with cross-sectional designs. These additional studies have very little new to offer, other than an additional correlation with a different psychological construct (e.g. peer affiliation [52], community bond [54], alexithymia [59], body self-esteem [65]). Furthermore, cross-sectional studies cannot ascertain the consequences of cyberbullying or IUD; this amplifies the issue stemming from the focus on community sampling mentioned above. Despite research data proving that the strongest associations with cyberbullying victimization were stress and suicidal ideation [67], no such parameters were researched.

Discussion

The small number of prospective studies have delivered the most robust findings, as expected. The usefulness of cross-sectional studies is very limited, especially as the phenomenon that they attempt to describe has already been conclusively confirmed and delineated. Unfortunately, as mentioned above, there are no studies that examine the relationship between cyberbullying and IUD in a relevant clinical sample. A single clinical study [48] carried out 10 years ago is misleading in that the population was not receiving help specifically for IUD or cyberbullying. Additionally, it employed a very basic measurement of cyberbullying with a single yes/no item and three items for IUD without delving deeper into any psychological correlates. Thus, we cannot assess the true impact of the relationship between cyberbullying and IUD on the well-being of the victim or any mental health correlates of the perpetrator. This would require a survey of patients seeking help for IUD or for the consequences of cyberbullying victimization on their mental health, or to address tendencies to victimize others.

Directions for further research

Study design stands to benefit from standardization of research instruments in future studies, with no studies so far sharing instruments and seventeen out of thirty-two using ad hoc measures. With the advent of a number of validated scales for cyberbullying and IUD, further usage of ad hoc measures should be discouraged. Cross-sectional studies have reached the limits of their usefulness as has the employment of non-clinical samples. There is a need for shedding more light in the complex interrelationship between cyberbullying, cybervictimization, and IUD, especially in gaming disorder, and causality cannot be adequately assessed with cross-sectional studies of community samples.

While a case can be made for measuring cyberbullying in a bullying context [10], a similar case should be made for measuring cyberbullying in a context outside the school environment and completely virtual. The overlap of cyberbullying and traditional bullying in a school environment may well be high but cyberbullying is not limited to this type of setting. Underage children and young adults no longer socialize exclusively within their school or their neighborhood. Social media widen the cycle of personal contacts to include total strangers in “real life.” Along with social media use, online gaming is a major pastime for most adolescents. However, there is no research that explored the toxic environment of certain online gaming communities. Ignoring this huge potential for victimization in the younger generation’s favorite pastime activities and demoting cyberbullying to a sub-category of bullying could lead to drastically underestimating its prevalence.

Future studies should include measures of well-being and psychological symptoms in order to quantify the relative impact of cyberbullying and IUD. Additionally, personality correlates should be studied in cyberbullying perpetrators. A prospective study of cyberbullying victims that could identify the factors that turn them to perpetrators themselves would be very helpful in elucidating the underlying psychological mechanisms. Finally, the studies so far have completely neglected mature adults, despite the fact that cyberbullying or IUD are not limited to younger age groups. College students in particular are an under-researched population with increased incidence of both cyberbullying and IUD.

Conclusions

The study of the relationship between cyberbullying and IUD is lacking studies with robust methodology, varied participant samples, and clinical measures of well-being and mental health. Future research should strive to employ samples more representative of the general online user population or focus on specific online activities and communities, employing clinical samples whenever possible.