Last checked: 16 October 2026. Part of our Explainers series.
In brief: Ask who published the report, how they gathered the information, what exactly they claim, and what would prove them wrong. Good reports show their method, date every fact, separate what is verified from what is alleged, and give the accused party a chance to respond.
Human rights information reaches us from many directions: UN investigations, reports by groups such as Amnesty International and Human Rights Watch, local monitors, journalists, governments and social media posts from people on the ground. All of them can be valuable. None of them is automatically right. This checklist is the one we use ourselves, set out in our Methodology.
The 10-point checklist
1. Who is speaking, and what is their role?
Identify the author and the type of organisation. A UN fact-finding mission, a court, an international rights group, a local activist network, a news agency and a government ministry all have different mandates and pressures. A clear “About” page, named staff or a stated mission is a good sign. Advocacy is not a disqualifier, since most rights groups are open about campaigning, but you should know which role a source is playing when it speaks.
2. Is the method explained?
Serious reports explain how they were researched. Human Rights Watch, for example, publishes a page on its research methodology, describing an approach “based on information gathering from a broad range of sources, and with field-based research at its core”, including interviews with victims and witnesses. Look for a methodology section that says how many people were interviewed, how documents or videos were checked, and what limits the researchers faced, such as being denied access to a country.
3. Is it primary or secondary?
A primary source is the original: the court judgment, the UN report, the organisation’s own release, the video itself. A secondary source reports on it. Secondary sources are useful for context, but details often change in retelling. Before you share a striking claim, follow the links back to the original. If there are no links, treat the claim with caution.
4. Is every fact dated?
Situations change fast, and a figure that was right in March may be badly out of date by October. Check the publication date and the period the report covers. Be wary of old reports recirculated as new, a common problem with images and casualty figures.
5. What do the numbers actually count?
Monitoring groups usually report what they have been able to verify, so their totals are minimums, not estimates of the real number. Two groups can publish different numbers for the same events because they apply different verification rules. Note the definition used, whether a figure is a verified count, a reported count or a statistical estimate, and who produced it. Our Iran page shows how we present differing counts side by side. Statistical groups such as the Human Rights Data Analysis Group work on estimating violations that go unrecorded; their core concepts page explains why raw counts can mislead.
6. Is the legal language used precisely?
“Violation”, “war crime”, “crime against humanity” and “genocide” have specific legal meanings and different thresholds. Check who is applying the label and at what stage: an allegation by a witness, a finding by a UN commission, or a judgment by a court. Our guide War crimes, crimes against humanity and genocide explains each term and the standards of proof, and ICC vs ICJ explains which court decides what.
7. Is there corroboration?
A single account can be true, but a strong report brings several independent strands together: testimony from people who do not know each other, documents, medical records, satellite imagery, verified video and official statements. Ask whether the sources are genuinely independent or all trace back to one origin.
8. Did the accused party get a chance to respond?
Fair reporting asks the government, armed group or company concerned for its side, and publishes the reply or notes that none came. A report that never mentions the other side’s account is not necessarily wrong, but it is incomplete. We follow the same practice, and note on each country page when we have not found a government response.
9. Can the images be verified?
Photos and videos are among the most powerful and most misused evidence. Before trusting or sharing one:
- run a reverse image search, for example with Google Lens, to see whether it appeared earlier in another context;
- check for an original uploader, date and location, and whether the weather, signs, language and landmarks match the claim;
- look for verification by an established open-source team; the Bellingcat resources page explains common techniques;
- be alert to AI-generated or edited images, which are now easy to produce.
Professional investigators follow the Berkeley Protocol on Digital Open Source Investigations, published by the UC Berkeley Human Rights Center with the UN human rights office. It sets standards for collecting, preserving and verifying online material so that it can be used in legal proceedings.
10. Does sharing it put anyone at risk?
Human rights material often concerns people in danger. Identifying a detainee, a witness or a family member can expose them or their relatives to reprisals. Do not share faces, names, addresses or other identifying details of private individuals unless they or their representatives have chosen to make them public. Blur or crop where needed, and never share material that appears to show sexual violence or children in distress.
A note on government sources
Government statements are primary sources for what a government says, and sometimes the only source for official figures such as court sentences or detention statistics. Treat them like any other source: check the date, the definition behind any number, and whether independent monitors or courts confirm it. Official reports about other countries, such as the annual human rights reports of the US State Department, can be useful for comparison, but they are also produced by a government with its own foreign-policy interests. The same caution applies to state media. Quote them accurately, attribute them clearly, and do not treat repetition across outlets as confirmation.
Red flags at a glance
- No author, date or link to an original source.
- Round, unchanging numbers repeated across many accounts with no named origin.
- Legal labels used as headlines with no indication of who reached that conclusion.
- Screenshots of text instead of links to the original post or document.
- Strong emotional framing with few verifiable details.
- Claims that “the media are hiding” something, with no evidence offered.
Green flags at a glance
- A methodology section and a named, accountable publisher.
- Every figure tied to a source and a date.
- Clear wording that separates verified facts, allegations and estimates.
- A response from the accused party, or a note that one was requested.
- A published corrections policy and visible corrections when errors are found.
How to use this checklist with our reporting
Apply it to us, too. Each of our pages links to its sources, gives dates, and says what we do not know. Our Editorial policy sets out how we work, including how we use AI tools, and our Corrections policy explains how we fix mistakes. If something on our site fails one of these tests, please tell us through our Contact page.
What you can do
Before sharing a report about Iran, the United States, Africa or anywhere else, take two minutes to run through the first four questions. Share the original source with its date, not a screenshot. For more ways to help, visit Take Action.
Additional sources
- Human Rights Center, UC Berkeley, home page, with research and training on open-source investigation.
- Committee to Protect Journalists, cpj.org, for context on the risks facing reporters who document abuses.
- Amnesty International, About us, describing its research and campaigning roles.
- US Department of State, Country Reports on Human Rights Practices, an example of a government source to compare with independent reports.