How to check facts

Fact-checking is the process of verifying the accuracy of reporting or statements. It often involves tracing the origin of a statement back to its original source (for example, a piece of research) and considering its original context, or comparing reporting in one source with the reporting in another (for example, multiple witness accounts of the same event).

This article looks at how fact-checking works. It is part of the Common’s Library’s good information toolkit, which reviews the strategies and criteria for spotting misinformation and working out what information can be relied upon.

Fact-checking organisations

For much of the 20th century, fact-checking occurred in dedicated teams of fact-checkers situated within news publications and publishers, scrutinising the accuracy of their writers’ own material. While this work continues, nowadays fact-checking is often used to describe dedicated journalists checking, critiquing and debunking public claims made by other people. While it is not specifically a fact-checking organisation, the House of Commons Library can check facts for MPs and their staff through our confidential requests service.

In recent years, fact-checking organisations have established an identity of their own and numbers have significantly increased: in 2023, researchers at Duke University identified over 400 fact-checking websites worldwide, double that in 2016.

While fact-checking organisations are perhaps best known for their work on political claims, many also work on social media misinformation across a range of topics. A major reason for this was the direct funding available from Meta to fact-check content on its platforms (primarily Facebook and Instagram), although as of 2025 Meta is ending its third-party fact-checking programme in the US. The International Fact-checking Network estimated that almost half of the fact-checking industry’s income in 2022 came from Meta’s third-party fact-checking programme.

Standards for fact-checkers

Fact-checkers are usually signatories of independent standards schemes like the International Factchecking Network (IFCN) code of principles or the European Fact-Checking Standards Code. Similar to general journalistic codes such as the Independent Press Standards Organisation’s Editors’ Code of Practice these include a commitment to accuracy, and they also require fact-checkers to be politically impartial and cite publicly identifiable sources wherever possible.

While adhering to these codes provides some reassurance about the quality of fact-checks, most organisations still urge individuals to read their information critically.

For example, the UK’s largest independent fact-checking organisation Full Fact says, in response to the question “Why should I take your word for anything?”:

We’d prefer you not to. Anything we publish is attributed and linked to its source, so you can make your own mind up about it.

The IFCN Code of Principles says that fact-checkers “want their readers to be able to verify findings themselves” rather than taking their word for it.

The fact-checking process

The process of fact-checking a statement involves working out whether pieces of information (or facts) in a statement are verifiable and then examining whether other sources confirm or contradict them. The following steps are sourced from several fact-checking organisations, including:

• Full Fact, a UK fact-checking and campaigning charity
• FactCheck.org, a US non-profit organisation focused on fact-checking US politics
• AFP Fact Check, a department of the French international news agency Agence France-Presse
• Craig Newmark Graduate School of Journalism, part of the City University of New York
• Reuters Factcheck, part of the international news agency Reuters

Choosing what statements to check

Fact-checking organisations say that they usually choose to check statements that are in the public interest and could cause harm if false. Full Fact says that it “prioritise[s] claims that have the most potential to cause harm to people’s lives”.

The US organisation FactCheck.org, which focuses on statements made by senior US political figures, says that it “seek[s] to devote an equal amount of time reviewing claims by Republicans and Democrats”.

Reuters factcheck includes the following checklist in how it decides what to fact-check:

• Editorial value: is the topic timely and of public interest?
• Could the material potentially cause real-world harm, if it is inaccurate?
• Reach: how far has the claim travelled? We examine the level of interactions on individual posts, as well as the visibility of the claim across different platforms
• Potential reach: is the information likely to be shared further? We examine how quickly the post is generating interaction and consider whether it may be repeated by others.
• Balance of fact vs opinion: is it possible to isolate certain claims from the material?

Working out what it’s possible to check

The next step involves evaluating a claim and what parts of it are verifiable ‘facts’. Statistics, definitions, quotations and dates can often be verified, whereas conjecture and opinion cannot.

Fact-checkers also need to consider what assumptions are implied by a claim and look for evidence to assess whether those assumptions are justified.

Full Fact highlights that, in some cases, factually correct information can be misused to draw a misleading or incorrect conclusion; for this reason it says it “doesn’t only fact check the evidence used in a claim, but also the underlying assumption”.

For more on how factually correct information can be misused, see the Library guide on How to spot spin and inappropriate of statistics.

What can be verified more easily?

Claims are easier to verify when they refer to:

• something that is recorded or attributed to another source
  • For example, if someone claimed an expert said a policy would damage the economy, a fact-checker could contact that expert and see if that is what they said.
• something that is measurable and which someone reputable has measured
  • For example, if someone claims unemployment is high, a fact-checker could refer to a reputable source of labour market data, such as the Office for National Statistics. To check whether unemployment is “high”, the fact-checker might compare the unemployment rate to historical data or comparable data in other countries.

What can’t always be verified easily?

Claims aren’t always easy to verify when there is no definitive source that could confirm or refute it. This might be because:

• It is difficult or impossible to collect data that could confirm or refute it; for example, data on the effect of a policy before it has come into force.
• It would be unethical to collect the data; for example, testing the efficacy of a life-saving drug by denying it to one group and seeing if they live as long as the group taking it.
• Different sources with different but equally valid methods of measuring something come to opposite conclusions; for example, there are two main ways of measuring crime (surveys of households or police records) and these two measures show different trends because they capture slightly different types of crime.
• The claim involves one thing directly causing another, which is more difficult to prove than a correlation between two variables; this is particularly true for social and economic trends (such as crime or productivity) which depend on innumerable factors that can’t be controlled in experiments.

In these cases, a fact-checker might survey the evidence and explain what interpretations of it are or aren’t valid and under what assumptions.

What can never be verified

Matters of opinion or principle can’t be verified by fact-checking.

Opinions are sometimes obvious, such as “this is wrong”, but they can also be hidden. For example, if someone claimed “wages are too low”, a fact-checker could look at wage data to see whether average wages were low relative to historical rates. But they could not check whether current wages are “too low” – lower than they should be – because that is an opinion. That opinion might be based on what the effects of low wages are and how they value those effects compared with the effects of higher wages.

However, a fact-checker who wanted to evaluate a statement of opinion might refer to a range sources with different opinions and their reasons. They could then summarise this range of opinion, making the assumptions behind them explicit, for readers to make up their own minds.

Corroborating the information

Finally, fact-checkers compare the verifiable information with other sources.

If it’s clear what the original source of the information is, fact-checkers can compare what was in a statement to what was in this original source to see if the statement accurately represented it. If the statement was made by an identifiable person, fact-checking organisations commit under the International Factchecking Network Code of Principles to contacting them and asking for supporting evidence.

They will also use other sources to verify or counter the claims made in a statement. Full Fact says that it only uses publicly available sources:

We draw on publicly available information, including statistics and primary research.

We provide links to primary sources for all factual assertions wherever possible and aim to provide at least two sources to verify the central claim of a check unless there is only one relevant source. We prefer to link to data tables, legal documents, or other primary sources rather than press releases or executive summaries of statistical releases.

In publishing the results of fact-checking, organisations explain the sources they have used and any information about the sources that might affect their credibility (such as conflicts of interest).

Verifying information in conflicts and other developing situations

This section looks at the specific situation of verifying information in conflicts or other fast-moving and uncertain situations.

It focuses on the methodology of verifying individual reports and images, rather than entire narratives of conflicts or questions.

Verification poses unique challenges during conflicts

While best practice for verifying information in armed conflicts is similar to the general process above, conflicts can present some unique challenges to verification. These include:

• difficulties of journalists, researchers and civil society organisations accessing a conflict zone to independently verify reports
• the fast-moving nature of a situation, the speed at which information is generated, and the time-sensitive nature in responding
• allowance under international law for “ruses of war” (including camouflage, mock operations and misinformation) “intended to mislead an adversary”, so long as the ruse does not infringe on any other rule applicable in armed conflict
• deployment of ‘information operations’ and censorship by one or more parties to a conflict

Duke University's reporters lab website provibes a map of fact-checkers in different parts of the world.

Methods to assess text, video and image sources

Journalists use a range of methods to assess the authenticity of reports in conflict. The information here is summarised from several sources:

• Press Gazette, How to debunk misinformation coming out of war in Ukraine, 24 March 2022
• Full Fact, Ukraine and beyond: How to fact check misleading videos, 25 February 2022
• Stanford Report, Seven tips for spotting disinformation related to the Russia-Ukraine conflict, 3 March 2022
• Reuters Institute for the Study of Journalism, BBC expert on debunking Israel-Hamas war visuals, 13 October 2023

Plausibility of images and videos

Images and videos from previous conflicts or other situations may be re-circulated or re-labelled so they seem to be related to a different conflict.

One way of checking images is to look for discrepancies between what they purport to show and what they actually show. For example, do the language (in speech or text), dress, and landscape features match what you would expect for the region?

For more, see the Commons Library article on verifying images and videos.

Headlines, quotes and people or organisations named

Text-based stories may combine several elements of new and old reporting. Several methods can be applied to deconstruct an article:

• Check the headline. Is the headline supported and evidenced by the main story and its content?
• Is the story sourced? Reputable reports should cite the sources of their information.
• Check the quotes. A web search will help determine whether the quotes have been taken out of context and the original source of the quote.
• Check named individuals and organisations. A search will determine whether attribution is correct and allow an assessment of the reliability of the source.
• If in another language, search for a translation by a reputable organisation (which will likely provide its own report of an event).

Recognition of multiple perspectives and uncertainty

Most fact-checking organisations and established media will state competing positions and interpretations of an event.

Most will also flag any uncertainties in data or reporting.

Range of organisations reporting the same story

A major event is likely to produce varied reporting and varied images across a range of international and domestic media platforms (this would not include word-for-word repetition of the same report on other platforms and in different media).

Be aware of false fact-checkers or false attribution to established media

If uncertain, go to the original organisation’s website, where any attributed videos or reports should be available.