Understanding precisely how online harassment translates into lasting psychological harm matters enormously for schools and clinicians designing interventions. Most programs treat cyberbullying as a single-pathway problem, but new longitudinal evidence suggests the damage travels through distinct emotional mechanisms depending on whether the outcome is internalizing distress or outward behavioral disruption.
A two-wave study of 501 Chinese adolescents aged 15–18, assessed six months apart, used structural equation modeling to disentangle three candidate mediators: loneliness, emotion dysregulation, and self-esteem. Cyberbullying victimization at baseline predicted elevated emotional and conduct problems six months later. Loneliness and emotion dysregulation emerged as significant mediators for both outcome types, while self-esteem did not reach significance for either. Critically, subscale-level analysis of emotion dysregulation revealed a mechanistic split: limited access to adaptive regulation strategies primarily explained the pathway to emotional problems such as anxiety and depression, whereas impulse-control difficulties — the inability to manage behavior when distressed — most strongly accounted for the link to conduct problems including aggression and rule-breaking.
This bifurcation finding is the study's most clinically meaningful contribution. Prior cyberbullying research has largely treated emotion dysregulation as a monolithic construct; collapsing its facets obscures meaningfully different intervention targets. The impulse-control pathway to conduct problems aligns with established work on frustration-aggression responses and executive function deficits in victimized youth. Loneliness emerging as a shared mediator reinforces the social-disconnection hypothesis, consistent with rejection-sensitivity models of peer victimization. The self-esteem null finding is notable given how prominently self-esteem features in anti-bullying programming — it suggests cognitive self-appraisal may be downstream of, rather than a conduit for, behavioral outcomes. Key limitations include a single cultural context limiting generalizability, a six-month window that may miss longer-latency effects, and reliance on self-report measures. This is an incremental but methodologically careful advance that strengthens the case for emotion-regulation-targeted, multi-component cyberbullying interventions.
