How CAWT Sources Conflict Data

Reviewed 18 September 2026. This is an evergreen explainer; case-specific legal findings should be checked against the newest authoritative source.

In short: The sourcing hierarchy behind CAWT conflict profiles, figures, updates and explainers—and how we handle claims that cannot yet be independently verified.

A conflict website can look authoritative while quietly mixing incompatible numbers, unattributed claims and stale information. CAWT is designed to do the opposite: keep the source visible, preserve the source’s own level of certainty, and separate durable reference material from fast-moving updates.

Our default source hierarchy

Where possible, CAWT starts with first-order or specialist sources: UN monitoring missions and agencies, the ICRC and other humanitarian institutions, treaty bodies, international courts, official statistical releases and well-documented conflict datasets. High-quality journalism is important for developments that institutional sources have not yet covered and for reporting from places where international access is limited.

Parties to a conflict are sources—but not neutral validators

Governments, militaries, armed groups and de facto authorities often possess important information first. Their statements can be newsworthy and sometimes the only immediate source. CAWT attributes such claims explicitly. A military statement that it destroyed a target, or an authority’s casualty total, does not become independently verified merely because it appears in a later article.

Preserve chains of attribution

Institutional reports often reproduce figures from other bodies. OCHA, for example, may report a casualty figure supplied by a health ministry while explaining that the figure has not been independently verified by the UN. CAWT should preserve that distinction rather than rewriting the number as “the UN says.”

Prefer methods you can inspect

A useful dataset explains definitions, coverage, dates, evidence thresholds, revision rules and limitations. Where those are missing, apparent precision can be misleading. For this reason, CAWT may prefer a smaller verified count with transparent methodology to a larger number whose construction is opaque—while still reporting other major estimates with clear attribution.

Do not silently reconcile incompatible numbers

If one source counts violent civilian deaths and another estimates excess mortality, the figures answer different questions. If two displacement databases use different reference dates, they cannot simply be added. CAWT explains the mismatch rather than manufacturing a synthetic total.

Corrections are part of the method

Conflict facts change and sources revise data. Correcting a number is not a failure if the earlier figure was accurately attributed at the time. The important question is whether the site records the change transparently and keeps links to the source material that supports the new version.

Frequently asked questions

Does CAWT use Wikipedia as a primary source?

No. It can be useful for orientation, but publishable factual claims should be traceable to stronger underlying sources.

Can a government source ever be reliable?

Yes, but reliability is claim-specific. Official sources may be authoritative for their own decisions while still requiring attribution for contested battlefield claims.

Why not combine every available dataset?

Because different datasets often measure different populations, time periods and event types. Combining them can create a number no source actually supports.