Research

Research Spotlight

Anonymity, Reduced Accountability, and Adolescent Moral Reasoning Online

What We Looked At

This review examined how anonymity and reduced accountability alter moral reasoning in adolescent online interactions, drawing on Suler’s (2004) foundational account of the online disinhibition effect and subsequent research in cyberpsychology. Suler identifies six factors that reduce the behavioural inhibitions typically present in face-to-face communication: dissociative anonymity (concealment of offline identity), invisibility (not being physically seen or heard), asynchronicity (no requirement to cope with others’ reactions in real time), solipsistic introjection (constructing an imagined representation of others), dissociative imagination (treating online interactions as separate from real-world responsibility), and minimization of status and authority (reduced reliance on conventional markers of power and hierarchy). Suler distinguishes between benign disinhibition, for example seeking support about a socially embarrassing issue, and toxic disinhibition, where reduced accountability enables harmful behaviour.

The review considers how these factors intersect with adolescent development and moral reasoning, the blurring between victim and perpetrator in online contexts, and the degree to which each of Suler’s six factors has held up against more recent empirical and theoretical scrutiny. It also addresses deindividuation and subsequent frameworks that build on or revise Suler’s original model.

What the Research Found

Adolescents are considered particularly vulnerable to online harm because they combine greater online access and reduced parental oversight with developmental characteristics including sensation-seeking, social comparison, risk-taking, and identity exploration, alongside still-maturing cognitive control (Siegel, 2014; Spear, 2000; Johnson et al., 2009; Steinberg, 2010). Hernandez et al. (2021) found that narcissism, lack of empathy, disinhibition, and extraversion increase sexual-erotic online behaviour in adolescents, while Wang et al. (2021) found that high levels of online disinhibition were associated with high levels of cyberbullying perpetration among male adolescents with low empathy. Pluta et al. (2023) found that exposure to hate speech can impair empathy and perspective-taking at the neural level. Adolescents’ arousal-seeking nature also means that content warnings may sometimes provoke curiosity rather than avoidance (UCL Uncovering Politics, 2026).

Accounts of adolescent moral development have shifted away from Kohlberg’s (1958, as cited by McLeod, 2013) stage model towards a view of moral reasoning as a gradually developing process shaped more by emotional and social motivations than by cognitive deficits alone (Bajovic and Rizzo, 2020; Malti et al., 2021). McKee et al. (2024) found that adolescents’ drive to belong overrides moral judgment, and Zimbardo (1969) argued that group affiliation supplies individual role expectations. Johnson et al. (2009) caution that neuroscientific evidence about the adolescent brain is insufficient to support policy-level determinations about when young people become morally responsible, and Wachs et al. (2022) stress the need to understand what differentiates behaviour among adolescents rather than treating them as a uniform group.

The relationship between victim and perpetrator online is frequently blurred. Wachs et al. (2022) found a positive correlation between online hate speech victimisation and perpetration, suggesting that adolescents who are victimised may come to treat perpetration as an acceptable behaviour, with revenge and lack of coping strategies providing additional motivation. Strohmeier and Gradinger (2022) identified moral disengagement and risky online behaviour as shared risk factors for both cyberbullying and cyber victimisation. Turner et al. (2024) noted the prevalence of peer and intimate partner perpetrators in image-based sexual exploitation and abuse of children, and warned that prevention strategies focused on stranger danger are therefore inadequate. The motivation-based online disinhibition model (Wen and Miura, 2023) proposes that toxic behaviour occurs only when high levels of online disinhibition combine with strong motivation, with victimisation itself a potential source of motivation for subsequent perpetration.

On Suler’s individual factors, the research offers a more complex picture than the original model anticipated. Regarding dissociative anonymity, while anonymity is associated with negative in-game conduct (Kordyaka et al., 2020, as cited by Merilanen and Ruotsalainen, 2024) and with gaming environments that embed racist, misogynistic, and homophobic content (Miller-Idriss, 2025), anonymity does not straightforwardly predict toxic behaviour. Cyberbullying frequently occurs without anonymity (Wen and Miura, 2023), and a meta-analysis found only a weak positive correlation between anonymity and online self-disclosure (Clark-Gordon et al., 2019, as cited by Wen and Miura, 2023). Adolescents may also use anonymous online spaces constructively, to explore identity and bridge the gap between actual and ideal selves (Kurek et al., 2019; Michikyan et al., 2015, as cited by Wen and Miura, 2023).

On invisibility, Lapidot-Lefler and Barak (2012) found that lack of eye contact was a more significant driver of toxic disinhibition than anonymity or invisibility per se, suggesting the original model may conflate distinct mechanisms. Hollenbaugh and Everett (2013) found, contrary to the disinhibition hypothesis, that bloggers who shared their photograph or real name disclosed more, not less. Wang et al. (2024) found that online invisibility had no direct effect on cyberbullying, though Merilanen and Ruotsalainen (2024) note that the absence of visible emotional responses makes it easier for perpetrators to persist.

Regarding asynchronicity, Wen and Miura (2023) observe that contemporary digital culture, with read receipts and always-on connectivity, can create pressure to respond immediately rather than psychological distance. Wang and Ngai (2020), in a study of 1,103 adolescents, found that asynchrony directly fostered cyberbullying and also operated indirectly through moral disengagement. Wang et al. (2024) report that studies on anonymity and cyberbullying have produced inconsistent results overall, and that online disinhibition functioned through moral disengagement and low self-control to influence cyberbullying rather than producing it directly.

On solipsistic introjection, Pandita et al. (2024) emphasise that the removal of facial expressions and contextual nuance in online communication requires people to fill in the gaps using memory, stereotypes, and imagination. Young people are more vulnerable to predators partly because they are less likely to mentalize malicious intent and more likely to project their own desires onto others. Algorithms can further restrict exposure to diverse viewpoints, compounding these tendencies.

On dissociative imagination, the Proteus effect (Yee and Bailenson, 2007) complicates Suler’s assumption that behaviour in online or gaming environments remains neatly contained there. Szolin et al. (2022) found that digital avatars can produce lasting changes in socio-political views, gender-conforming behaviours, and self-perception. Merilanen and Ruotsalainen (2024) found that while some young gamers explicitly distinguished between their online and offline selves, banal toxicity was an accepted norm in gaming culture, suggesting that behaviour can become normalised even if it is not straightforwardly carried over.

On minimisation of status and authority, adolescents may have greater online agency than offline, but this can leave them exposed: many age verification measures have proven ineffective (Forland et al., 2024; Jarvie and Renaud, 2024; Marsden, 2023), and predatory adults can more easily pose as peers. In group situations where power dynamics are obscured, Christie and Dill (2016) found that only individuals with high self-esteem and high autonomy or low social anxiety were willing to challenge inappropriate behaviour, with most members staying silent for fear of rejection.

More recent theoretical developments supplement and in some respects challenge Suler’s framework. The Social Identity model of Deindividuation Effects (SIDE) describes how group norms become more powerful than general social norms online as social identity is heightened, meaning whether a group tends towards prosocial or antisocial behaviour determines outcomes (Spears and Postmes, 2015). Given adolescents’ particular susceptibility to peer influence, this model may be especially salient for this age group; as one participant in Regehr et al.’s (2024) research described, being immersed in an extreme niche online gradually erodes the original self. Pandita et al. (2024) developed the Disembodiment, Accountability and Disinhibition (DAD) framework, which builds on Suler to explain how the abstract and disembodied nature of online communication, platform design features, and emotionally arousing content each impede moral reasoning and rational engagement.

What This Means in Practice

The review’s overall finding is that while anonymity and reduced accountability can facilitate moral disengagement, they are neither necessary nor sufficient conditions for toxic online behaviour. Additional factors are implicated, including identity formation processes, peer influence, access to age-inappropriate content, empathy, self-control, arousal-seeking, algorithmic design, desensitisation, and escalation. Some adolescents have the coping skills to navigate these risks; for others, these factors combine to create significant vulnerability.

Suler’s six-factor model retains descriptive value, but more recent research reveals that each factor is more complex in practice than the original formulation suggested. Anonymity is not a reliable predictor of toxicity, and the absence of eye contact may be more important than physical invisibility. Contemporary asynchronicity can increase rather than reduce pressure. The boundary between online and offline identity is more porous than dissociative imagination assumes. And the minimisation of authority online may obscure power dynamics rather than simply flatten hierarchies.

For practice, this evidence suggests that interventions focused narrowly on reducing anonymity, or on encouraging young people to behave online as they would offline, are unlikely to be sufficient. The social motivations that shape online behaviour, including the powerful pull of group membership and peer acceptance during adolescence, need to be addressed directly. Media literacy approaches that help young people understand how online environments are designed, how algorithms shape what they encounter and how they engage with it, and how collective norms develop and normalise harmful behaviour may offer more traction than individual-focused approaches alone. Pandita et al.’s (2024) observation that research has yet to adequately characterise adolescent-specific vulnerabilities to moral outrage and affective polarisation online points to a significant gap in the current evidence base.

References