Working definitions for the terms used across the atlas. These are operational, not final. They describe how the platform uses each word, not how every field defines it. Where scholars disagree, we note that disagreement rather than hiding it.
A measure of how much people dislike or distrust members of the opposing political party or faction, regardless of specific policy disagreements. High affective polarization often precedes political violence and democratic backsliding.
Hostility toward, prejudice against, or discrimination targeting Jewish people, ranging from conspiracy narratives and stereotypes to violence and institutional exclusion. The atlas tracks both incident counts and the social conditions that sustain them.
A sourced example of reconciliation, de-escalation, reform, or successful inter-group repair. Bridge stories are paired with harm patterns so the platform never presents violence without what has reduced it.
A focused narrative entry that follows a specific episode or pattern of hostility across time, sources, and outcomes. Case studies link back to the countries, groups, and themes they belong to, and include scholar notes for deeper reading.
A hereditary system of social hierarchy, most formally documented in South Asia but present wherever inherited status determines access to work, marriage, housing, and justice. Caste hostility is recorded through atrocity data, exclusion studies, and community reporting rather than a single source.
A rating attached to each country or claim that reflects source diversity, coverage, recency, and methodological rigor. Confidence is not a measure of hostility; a high-confidence country may have high or low scores, and a low-confidence country may be under-recorded rather than safe.
A country, region, or topic where reporting infrastructure is so weak or deliberately suppressed that the absence of data is itself a finding. Data voids are rendered visibly on the map rather than hidden behind a neutral color.
A political ideology that defines a nation primarily by shared ethnicity, language, or ancestry rather than by civic membership. Ethnonationalist framings correlate strongly with hostility toward migrants, minorities, and interfaith or interethnic households.
The intentional killing of women or girls because of their gender. The atlas uses UNODC and national femicide monitoring data where available, with caveats about the wide variation in how countries classify and report these deaths.
A criminal offense that is motivated, in whole or part, by the offender's bias against a protected characteristic such as race, religion, sexual orientation, or disability. Definitions and thresholds vary widely across jurisdictions, which is why the atlas does not treat hate-crime counts as globally comparable.
One of the 17 categories of inter-group tension tracked by the atlas, such as antisemitism, anti-LGBTQ+ hostility, or caste-based violence. Vectors are lenses for reading the evidence, not rankings.
An analytical split between hostility directed at a group inside its own country (inbound) and hostility the same group faces when present as a diaspora or minority elsewhere (outbound). The two readings often diverge sharply for the same group.
A sourced event pinned to a country or region: a protest, an act of violence, a law change, or a legal proceeding. Markers are illustrative rather than exhaustive and always link to the citation behind them.
The sociological theory and evidence base showing that sustained, equal-status contact between groups under cooperative conditions can reduce prejudice. Intergroup contact is a common ingredient in the bridge stories tracked by the atlas.
A framework for understanding how multiple identities (race, gender, class, caste, disability, sexuality) combine to shape specific experiences of hostility that would be missed by looking at any one axis alone.
Hostility toward Muslims or Islam expressed through discrimination, harassment, violence, or state policy. The atlas draws on official hate-crime data, survey work, and specialized NGO reporting to track Islamophobia across very different contexts.
The set of rules and source choices used to produce a score or narrative. Every metric on the atlas has a methodology note explaining what counts, what does not, and why. See the Methodology page for the full account.
Negative attitudes, beliefs, or actions directed at people perceived as belonging to a different group. Out-group hostility is measured through surveys, recorded incidents, and content analysis, each with different blind spots.
An editorial convention on the atlas: entries are titled in a way that foregrounds the people affected and the question being asked rather than sensational framings. For example, “Why did anti-Asian hostility spike during COVID-19?” rather than “America’s Asian Hate Wave.”
A severe form of polarization in which partisan identity becomes the dominant social identity, such that opposing groups are seen as existential threats. Political sectarianism is distinct from ordinary disagreement and tends to track closely with democratic erosion.
The documented origin of a claim or data point, including source, period, geography, and method. Every surfaced number on the atlas should be drillable into its provenance chip so readers can verify or push back.
Work that repairs relationships and institutional trust after a period of intergroup harm. Reconciliation is distinct from reform: reform changes laws and systems, reconciliation changes relationships between people and communities.
Structural change to laws, institutions, or policies intended to reduce discrimination or hostility. Reform can outpace reconciliation (a law changes, attitudes do not) or lag it (communities de-escalate before the state catches up).
An optional reading mode that reveals extra methodological notes, footnotes, and caveats written for researchers and reviewers. Scholar mode is off by default so that the surface experience stays approachable for general readers.
A category of evidence: official statistics, survey data, human-rights reporting, archives, conflict databases, academic work, and so on. The atlas deliberately mixes families so that no single source family dominates a claim.
The degree to which a government restricts independent reporting on discrimination, violence, or protest. State opacity is a measurable condition, not an absence of information, and is visualized on the map for countries that refuse to cooperate with international monitors.
Patterns of exclusion that are built into institutions, laws, or everyday practice rather than attributable to a single person’s bias. Structural discrimination is often invisible in hate-crime counts but shows up in housing, hiring, education, and criminal-justice data.
A topic-based entry on the atlas that pulls together everything tracked about a particular vector, such as antisemitism or caste discrimination, across countries, groups, and case studies.
A design approach that avoids sensationalizing harm, uses content warnings where appropriate, limits aggressive color and motion, and treats the reader’s wellbeing as a first-class constraint alongside clarity and accuracy.
The gap between recorded incidents and lived harm. Underreporting is caused by weak reporting systems, mistrust of police, legal thresholds, institutional filtering, or fear of retaliation. It is one of the main reasons raw hate-crime counts are not comparable across countries.
Hostility or prejudice toward people perceived as foreign or from outside one’s own community. Xenophobia is tracked through survey data, hate-crime records, and legislative monitoring, since it can show up in all three.
The glossary is expanded as new topics are added to the atlas. If a term is missing or a definition feels off, flag it through the contact page and it will be reviewed against the sources we rely on.