Digital self-harm: an examination of the current literature with recommendations for future research

Digital self-harm (DSH) is a relatively new form of virtual self-harm that occurs mostly in young people, and has been defined as both the act of cyberbullying oneself (Fictitious Online Victimisation; FOV), and engaging with online spaces in a way that is detrimental to one’s psychological wellbeing. In our overview of DSH we aim to provide an understanding of this little-known subject by offering a clarification of extant terms, summarising current findings, and presenting recommendations for future research. The prevalence of DSH in teenagers appears to approach that of physical self-harm, and the behaviour is more common amongst young people who are ostracised and/or victimised by their peers. The motivations for DSH are heterogeneous, but there is evidence that, like physical self-harm, it may be a means of affect regulation. However, more research is required to untangle the different forms of DSH and gain a sense of their occurrence in different populations and age groups, their varying causes, their relationship to physical self-harm, and the means by which they should be measured.


Background
Recently, a form of digital behaviour has emerged amongst adolescents that is both surprising and concerning. 'Digital Self-harm' (DSH), as it is most commonly termed, involves the individual using digital platforms to present themselves as the target of interpersonal aggression, such as by pretending to be the victim of cyberbullying. This phenomenon was first described empirically by Englander [1], who conducted an early, non-peer reviewed survey of the high school behaviour of US college students. She found that 9% had anonymously posted a 'cruel remark' directed at themselves or had anonymously cyber-bullied themselves. A much larger study by Patchin and Hinduja [2] found that approximately 6% of 12-17-year-olds across the US had anonymously posted something online about themselves that 'was mean' , while 5.3% had 'anonymously cyber-bullied' themselves. Similarly, Pacheco et al. [3] found that 6% of New Zealand teenagers had posted or shared online mean or harmful content about themselves in the past year.
While DSH most often refers to the act of verbally abusing oneself through a digital persona, some would argue that the term encompasses behaviour that is traditionally harmful and is simply enacted virtually rather than physically. Pater and Mynatt [4] suggest that DSH is "online communication and activity that leads to, supports, or exacerbates, non-suicidal yet intentional harm or impairment of an individual's physical wellbeing. " The increasing accessibility of the internet has enabled the creation of online communities centred around the normalisation of specific psychopathologies, such as eating disorders [5]. Individuals can enact DSH by viewing the content circulated within these communities, which may encourage damaging behaviours (e.g., restricted eating) or contain deliberately triggering images and language. 2 Current research on digital self-harm 2.1 Defining digital self-harm: self-cyberbullying, fictitious online victimisation, and other concepts Due to its relative novelty, the phenomenon of DSH has not been well-defined. A contributor to the lack of consensus is the inherent vagueness of the name that is generally used for it. Deliberate self-harm in is defined as behaviours that are intended to cause direct or indirect physical harm to the self, with emphasis placed on intent-some 'normal' behaviours, such as getting a tattoo, can be classified as self-harm because the individual engages in them to hurt themselves [6]. 'Digital self-harm' implies the translation of traditional self-harm into the virtual sphere; however, in her exploratory survey Englander [1] links the name to the act of cyber-bullying oneself. While she also uses the term 'Digital Munchausen' , which directly references the existing psychological disorder wherein affected individuals pretend to have an illness, subsequent studies have generally adopted the former title [2,3,7].
The issue of nomenclature is important because DSH is also used to describe behaviours that are clearly a form of self-injury enacted through virtual means [4]. Recently, however, Erreygers et al. [8] conducted an exhaustive survey of self-cyberbullying in over 900 predominantly Belgian adolescents. Their study aimed to improve clarity within the field and to support further research by, firstly, proposing an alternate term for the behaviour-fictitious online victimisation (FOV)-and, secondly, developing an index for assessing the prevalence of FOV. The term 'fictitious online victimisation' ensures that the key features of the phenomenon-virtual nature, element of deception, hurtful content-are included without making misleading references to existing behaviour or disorders, e.g., cyberbullying, physical self-harm, and Munchausen by Proxy [8]. Henceforth, this paper will use FOV to describe online activity in which apparent interpersonal harm is actually self-created. Non-fictitious DSH will be used for online behaviour that cause deliberate, direct (i.e., not through a false identity) harm to the individual. DSH will be used as an umbrella term describing any online activity that is problematic and/or injurious to the individual's wellbeing, and thus includes both FOV and non-fictitious DSH.
It appears, then, that only DSH as characterised by Pater and Mynatt [4] meets the definition of self-harm used to evaluate behaviours occurring in the real world. However, as will be explored below, current research on FOV suggests that young peoples' motivation for-and risk of-engaging in this behaviour is associated with their social status, life experiences, and mental wellbeing, meaning the phenomenon of FOV may be as closely related to physical self-harm as DSH, as defined by Pater and Mynatt [4], is. It is possible that our understanding of self-harm in young people will need to expand in the future to account for the evolving ways in which youth engage with digital spaces.

Causes and risk factors of digital self-harm
The motivations underlying DSH behaviours are not well understood; however, surveys of the FOV specifically report a variety of cited reasons. These can be roughly categorised into three general causative types: (1) social development; (2) personal gain; and (3) emotional release. The first involves thought processes such as wanting to know if one's friends are 'real friends' by gauging their reaction to self-defamatory content, wanting to prove 'toughness' by demonstrating one's ability to withstand such attacks, or wanting to start a fight with someone [1][2][3]. Posting or sending bullying content to oneself can produce personal gains such as sympathy and attention from others, or a reaction that the individual finds personally humorous [2,3,8].
The third category of FOV motivations encompasses those that appear to be emotion-based-individuals engage in this behaviour as a manifestation of depressive symptoms or self-hating thoughts, or to alleviate them [2,3]. Patchin and Hinduja [2] found that FOV was more common amongst teens who were bullied, depressed, or harmed themselves offline. FOV being related to poor mental health or self-esteem is further supported by findings revealing which individuals are more likely to participate in this activity. Young people who are part of a minority, particularly a stigmatised one, appear to be at higher risk of such behaviours. Specifically, students who are LGBTQ +, had a mental or physical disability, or who behaved deviantly were more likely to anonymously cyber-bully themselves [2,3]. Supporting the idea of peer ostracisation as a key component to this behaviour, both Patchin and Hinduja [2] and Erreygers et al. [8] found that DSH was more common amongst individuals who had been bullied (both offline and online). Additionally, Erreygers et al. [8] found negative correlations between self-reported FOV and self-esteem, subjective well-being, and life satisfaction.
It is likely, then, that young people who do not feel accepted and safe in their social circles are at greater risk of victimising themselves online. This suggests that, in some cases, FOV serves a similar purpose to 'traditional' forms of selfharm, that is, individuals experiencing emotional turbulence may engage in harmful behaviours (e.g., cutting or bullying themselves) as a form of affect regulation. Self-report and experimental research on physical self-harm in adolescents has consistently found that the primary reinforcers of self-harm are emotional, that is, individuals harm themselves either because the behaviour alleviates their unwanted emotions (a form of negative reinforcement), or because it promotes pleasant emotions (a form of positive reinforcement) [9][10][11].
Two recent studies support emotional regulation as a motivator for FOV. Meldrum et al. [7] used structural equation modelling to determine the relationships between bullying victimisation, negative emotions, and FOV in Florida teenagers. Being bullied was associated with higher rates of digital self-victimisation, and this effect occurred both directly, and via the mediation of negative emotions. That is, harmful online behaviours appear help relieve the negative affect caused by peer rejection and harassment. Critically, these results parallel those of a similar study investigating physical self-harm in UK adolescents [12]. Structural equation modelling indicated that being bullied prior to high school was directly related to adolescent self-harm. There was also an indirect association which was mediated by depressive symptoms. Thus, it is likely that both FOV and physical self-harm are used as a form of affective regulation in young people, particularly those who are experiencing higher rates of negative affect due to peer victimisation.
FOV sharing a common causal pathway with physical self-harm is further supported by the work of Semenza et al. [13], who also analysed data from Florida highschoolers. Using logistic regression, it was found that FOV was negatively associated with sleep duration, which is also an established predictor of physical self-harm. Importantly, individual's depressive symptoms partially attenuated this relationship, indicating that while poor sleep is a correlate of FOV independently, depression also plays a significant role. The authors suggest that reduced sleep duration negatively affects emotional regulation, leading to an increased likelihood of depression and thus a reliance on problematic online behaviour as a means of alleviating these negative emotions. The study is not conclusive, but it provides further evidence for FOV being a form of affect regulation in the same way that physical self-harm can be.
Unfortunately, literature on the motivations and risk factors associated with non-fictitious forms of DSH is scarce. In their mixed-methods case study of DSH in three recovering eating disorder patients, Pater et al. [14] found that online content was used to sustain or justify harmful activities. During their teenage years, at the height of their eating disorders, individuals would seek out media that glorified extremely thin bodies or unhealthy eating habits in order to make themselves feel bad, or to find a sense of community and support for their disordered thinking and behaviour. As of yet, it is unclear how demographic factors interact with these 'self-triggering' online behaviours.

Assessing digital self-harm
The inherently modern nature of DSH also means that assessment of the phenomenon has yet to be formalised. Early studies of FOV interrogated the behaviour using a variety of question formats and time frames. Englander [1] asked first year college students whether they had 'falsely posted a cruel remark 'against' themselves' or cyberbullied themselves in high school. The frequency of the behaviour, and motivation underlying its occurrence, was assessed with multiple choice questions. In their study, Patchin and Hinduja [2] asked teenage participants whether they had anonymously posted something online that was mean to themselves, or cyberbullied themselves, in their lifetime. Participants were able to choose from 'never' , 'once' , 'a few times' , or 'many times' . Motivation was assessed with a single open-ended question. Pacheco et al. [3] used similar questions to determine prevalence but asked teens about their behaviour within the past year. Participants also had to choose from the options given regarding the reasons for their behaviour. Only Englander [1] and Pacheco et al. [3] asked if subjects had achieved the desired goal using FOV.
Erreygers et al. [8] developed a standardised index for the assessment of FOV. The index avoids several of the limitations that characterise previous assessments-subjects answer multiple items referring to specific behaviours (rather than a single general item) and provide information on what the FOV was about (e.g., their appearance, etc.), and how it was enacted (the digital platform, the identity the subject used, how they disguised their own). Erreygers et al. [8] report that the instrument met the appropriate validation requirements; however, it is somewhat unclear why the authors chose to develop an FOV index instead of a scale. According to Diamantopoulos and Winklhofer [15], indices consist of formal indicators, which are 'observed variables that are assumed to cause a latent variable' . Erreygers et al. [8] do not specify a latent variable that is assumed to be caused by the behaviours assessed in the FOV index.
In terms of non-fictitious DSH-that is, engagement with online spaces or media that promotes emotional and physical damage-no standardised instruments have been developed. This is likely a consequence of non-fictitious DSH being recognised as a phenomenon only recently, and hence being poorly characterised. It remains unclear whether the behaviours included under the category of non-fictitious DSH are direct translations of physical self-harm, and can thus be assessed with similar instruments, or if they contain meaningful differences that require a different approach for assessment.

Why should digital self-harm be investigated?
DSH is an important issue to examine as it may have serious impacts on the physical and mental health of today's youth. The prevalence of traditional self-harm is estimated to be between 7.5 and 46.5% in teenagers, and 38.9% in university students [16]. Students who had depressive symptoms, or who had previously self-harmed physically, appear more likely to anonymously cyber-bully themselves [2,8], and this behaviour seems more common in individuals with an existing mental health problem [3]. This suggests that rates of self-harm in young people may be even higher than previously estimated when self-harm in cyberspace is accounted for, particularly when the ubiquity of social media and internet use in modern young people is factored in [17,18]. Such increased prevalence of DSH is a cause for concern, as adolescent self-harm is a risk factor for subsequent depression and anxiety disorders, substance abuse, and poorer educational and occupational prospects [19]. In cases of chronic self-harm, individuals are also at higher risk of suicide [20].
It is highly likely that negative physical and emotional effects are also consequences of self-harm that occurs through digital means, whether fictitiously or not. A recent longitudinal study of weight-loss behaviour in the members of a 'pro-eating disorder' online community found that users' current and desired body weights decreased over a period of 15 months, and that many were pursuing weight goals that were dangerously low and unattainable [21]. While the study did not examine DSH specifically, the findings indicate that involvement in online spaces that contain highly triggering content can have serious real-world effects. The possibility of lifelong negative consequences as a result of DSH makes the scarcity of peer-reviewed empirical research a pressing issue. More information is required on the prevalence, causes, risk factors of DSH, as well as how this data is affected by the type of DSH, which can help aid successful future interventions or treatments.

Directions for future research
Currently, there is little peer-reviewed research on how and why fictitious and non-fictitious DSH occurs. While we have a general idea of the prevalence of FOV (approximately 6% of US high schoolers; [2]), the corroboration of this statistic in different populations is required. Additionally, to our knowledge, no peer-reviewed studies have examined whether FOV occurs in older youth, such as those past high school. In terms of non-fictitious self-harm, some research has been conducted on the online communities that form around eating disorders and physical self-harm [5,22], but there is minimal data on how common it is for young people to visit these sites or social media pages.
Critically, studies have yet to compare forms of online and offline negative behaviours. Current research suggests that both FOV and non-fictitious DSH share certain features with physical self-harm, such as the population most affected (young people, especially those who abuse substances, have symptoms of psychopathology, and experience bullying) and key motivators (emotional regulation and social functioning) (see Table 1). However, these commonalities have yet to be assessed empirically side-by-side, and some aspects of DSH-aetiology, long-term impacts-remain unknown.
As such it is, currently unclear what drives individuals towards, for example, cutting themselves versus visiting an online forum that supports cutting versus anonymously posting a hateful message about themselves, and how these behaviours may interact. It is possible that young people who physically self-harm today are more likely to also use digital content to either justify or trigger their activity. Comparative research would provide greater clarity regarding how we define forms of DSH, and would also enable our current understanding of self-harm in youth more broadly to be updated. Finally, more research is required to validate existing assessments of DSH and to develop new instruments for measuring different behaviours within this category.  [16] Approximately 6% of teenagers Risk factors Psychopathology, substance abuse, disordered eating, peer victimisation, familial dysfunction [16,23] Psychopathology, substance abuse, disordered eating, peer victimisation, minority status (LGBTQ +, disability) physical self-harm Causes Childhood trauma, emotional dysregulation, stress, dissociation, depressive symptoms [16] ? Function Affect regulation, self-punishment, distraction, social functions [16] Affect regulation, social motivations, personal gain Effects Increased risk of depression, anxiety, substance abuse, and suicide; reduced educational and occupational outcomes