evidence & Research

Civility Bureau’s framework is anchored in research consensus and prevalence data. Every attestation sub‑category is supported by evidence showing both its real‑world impact and its relative frequency online. This ensures that scoring is proportional, transparent, and regulator‑ready.

Why Evidence Matters

  • Legitimacy: Categories reflect documented harms, not arbitrary labels.

  • Proportionality: Severity tiers are grounded in prevalence and impact.

  • Auditability: Evidence anchors make scoring rules verifiable and defensible.

Moderation reviews data

Evidence by Category

Category / Sub‑category Prevalence Insight Why it matters for scoring
Insults / Harassment 1 in 3 young women report online harassment (EU Fundamental Rights Agency, 2020). High frequency, often lower severity → proportional scoring avoids over‑penalisation.
Unwanted sexual exposure 1 in 5 children in Europe exposed to sexual images online (EU Kids Online, 2020). Lower frequency but higher harm → higher severity tier.
Coercion / Extortion Europol reports thousands of sextortion cases annually, with rising prevalence. Severe impact, repeat risk → top severity tier.
Hate speech Eurobarometer surveys show 60%+ of Europeans encounter hate speech online. High prevalence, societal impact → mid‑to‑high severity.
Misinformation / Disinformation EU Commission reports 70% of citizens exposed to disinformation during COVID‑19. Systemic risk, but less direct individual harm → proportional scoring needed.
Bullying / Repeated targeting OECD studies show 15–20% of adolescents experience repeated online bullying. Frequency and persistence elevate severity over one‑off insults.
Impersonation / Identity abuse ENISA highlights identity misuse as a growing online threat. Lower prevalence but high trust impact → mid‑tier severity.
Exposure of private information (doxxing) Research shows doxxing incidents are increasing, with severe offline consequences. Rare but high‑impact → escalated severity.

Research Foundations

  • Criminology: The online disinhibition effect explains why ordinary users engage in harmful behaviour online.

  • Behavioural science: Evidence shows that graduated deterrence is more effective than blanket bans.

  • Policy context: Our framework aligns with EU Digital Services Act enforcement priorities and Horizon Europe’s digital safety agenda.

Together, these insights explain the high prevalence of harassment and exposure, and underline the need for proportional deterrence rather than blunt enforcement.

Our Position

By grounding every sub‑category in prevalence data and research consensus, Civility Bureau ensures that its scoring is not only technically auditable but also socially legitimate. Regulators can trust that proportionality is evidence‑based, transparent, and defensible, never arbitrary.