Fact-Checking at Scale: A UK Journalist's Guide

Misinformation has become one of the defining editorial challenges of the 2020s. The speed of social media means that false or misleading claims can reach millions of people before a single journalist has had time to check them. For UK newsrooms operating under IPSO's Clause 1 accuracy obligation — and for broadcasters bound by Ofcom's due-accuracy requirement — developing robust, repeatable fact-checking processes is both an ethical imperative and a regulatory necessity. This guide draws on the published methodologies of the UK's leading fact-checking organisations to give every journalist a practical framework for working at scale.

The UK Fact-Checking Landscape

The United Kingdom has one of the most developed fact-checking ecosystems in the world. Several organisations now dedicate significant editorial resources to verifying public claims, and their work has shaped both journalism practice and public understanding of misinformation.

• Full Fact is the UK's leading independent fact-checking charity, founded in 2010 and a signatory to the International Fact-Checking Network (IFCN) Code of Principles. It covers political claims, health misinformation, and viral content circulating on social media.
• BBC Verify is the BBC's dedicated specialist team, launched in 2023, which uses open-source intelligence, geolocation, and data analysis to verify claims for live and investigative broadcasting.
• Channel 4 FactCheck operates within Channel 4 News and has been publishing fact-checks of political claims since 2008, with a particular focus on parliamentary debates and party manifestos.
• The Reuters Institute for the Study of Journalism at the University of Oxford produces authoritative research on how fact-checking organisations operate, their impact on public knowledge, and the spread of misinformation in the UK media environment.

These organisations publish their methodologies and correction records openly, providing models that any newsroom or individual journalist can learn from and adapt. The Reuters Institute's annual Digital News Report is an essential reference for understanding how UK audiences encounter and process fact-checked content.

Beyond these specialist bodies, many UK national and regional newsrooms have embedded fact-checking responsibilities into their general editorial workflow. The Guardian's data team, the Times's investigations desk, and regional operations including the Manchester Evening News and Yorkshire Post all employ journalists who apply structured verification to claims before publication — often without labelling the output as a “fact-check.” Understanding the principles behind the specialist organisations helps any journalist bring the same discipline to their own work.

Workflow for Live Events: Elections and Breaking News

For individual journalists covering elections, party conferences, or major public inquiries, the live events workflow requires the same discipline as a specialist unit — just at a smaller scale. The key discipline is always completing verification before broadcast or publication, even under time pressure; a wrong fact published at speed and corrected hours later is still a wrong fact that an audience will have acted on.

The tension between speed and accuracy is most acute during live events. Elections, major political announcements, and breaking news create high demand for rapid claim verification at precisely the moment when the normal pace of editorial checking is hardest to maintain.

The Reuters Institute's research into the 2024 UK general election found that misinformation spread fastest in the hours immediately after a major campaign event, before established fact-checkers had had time to publish their verdicts. This window — sometimes called the “virality gap” — is the period when preparation and pre-researched briefs have the greatest value.

BBC Verify developed a real-time model for its 2024 general election coverage that illustrates how to manage this tension. Rather than waiting until the end of a broadcast to correct claims, the Verify team monitored live output simultaneously, flagged checkable claims in real time, and published brief online verdicts within minutes of broadcast. This required a clear internal protocol: a designated monitor, a rapid-response researcher, and a senior editor empowered to approve short-form verdicts under time pressure.

• Pre-event preparation: Anticipate likely claims. For an election night, prepare fact-check briefs on the claims most frequently made by each party during the campaign so that researchers can reach for pre-researched evidence rather than starting from scratch.
• Triage: Not every claim can be checked in real time. Prioritise claims that are specific, verifiable, and likely to mislead a significant audience. Vague or unmeasurable assertions can be flagged for post-broadcast analysis.
• Version control: Live fact-checks are updated as new evidence emerges. Use a version-controlled document (a shared Google Doc or a newsroom CMS with revision history) rather than a static email thread so that every team member has access to the most current evidence.
• Signposting: Make clear to your audience when a fact-check is preliminary and subject to revision. Transparency about uncertainty is not weakness — it is part of responsible live journalism.

Full Fact's Methodology

Full Fact publishes its methodology in full, making it one of the most transparent fact-checking operations in the world. Its core workflow has four stages:

1. Claims monitoring: Full Fact uses a combination of manual monitoring and automated tools to identify claims in the public sphere that are checkable, significant, and relevant to its audience. The organisation has invested in automated claim detection software that can process large volumes of text from parliamentary debates, media appearances, and social media.
2. Evidence gathering: Researchers identify primary sources for the figures or assertions in question — official statistics, peer-reviewed research, government publications — and assess whether the claim accurately represents those sources.
3. Rating and publication: Full Fact uses a nuanced verdicts system (detailed further in the ratings section below). Every published fact-check includes a link to the primary evidence and a clear explanation of the reasoning behind the verdict.
4. Follow-up and corrections: Full Fact actively pursues corrections from public figures and media organisations when claims are found to be inaccurate. It publishes the outcomes of those requests regardless of whether the correction is made.

Full Fact's IFCN signatory status requires it to meet the Code of Principles, including commitments to non-partisanship, transparency of funding and methodology, and an open corrections policy. The IFCN conducts periodic assessments to verify compliance.

Full Fact has also been a pioneer in automated fact-checking research. Its automated system monitors thousands of claims daily across parliamentary proceedings, broadcast media, and social platforms, flagging those that match known checkable assertions for human review. This human-in-the-loop model — where automation handles monitoring and triage while trained journalists make the editorial judgements — is increasingly seen as the most practical approach to fact-checking at genuine scale. No fully automated system yet achieves the accuracy and contextual nuance required for publication-ready verdicts.

BBC Verify: The Broadcaster's Approach

BBC Verify, launched in May 2023, brought together the BBC's existing disinformation, data journalism, and visual verification teams under a single brand. Its work falls into three broad categories:

• Geolocation and OSINT: Verifying the location of images and videos using satellite imagery, street-level mapping (Google Street View, Mapillary), and visual analysis of landmarks, street signs, and infrastructure. Bellingcat's published guides on geolocation are a widely used reference within the verification community.
• Open-source verification: Using publicly available data — company registers, court documents, shipping records, flight-tracking databases — to verify factual claims about people, organisations, and events.
• Data analysis: Checking statistical claims against official datasets, assessing whether figures have been cherry-picked or presented in misleading ways.

BBC Verify operates within the BBC's existing editorial guidelines, which require accuracy and impartiality, but adds specialist verification capacity that general-purpose reporters may lack. Its published output includes clear sourcing for every claim and explicit notes when the team was unable to verify or disprove an assertion.

One of BBC Verify's most significant contributions has been normalising the explicit acknowledgement of uncertainty in broadcast journalism. Rather than presenting unconfirmed information as established fact, the team's output regularly uses language such as “we have not been able to independently verify this claim” or “the evidence is consistent with but does not conclusively establish.” This transparency about the limits of verification is a model worth adopting across all journalism, not only dedicated fact-checking units.

Building a Distributed Fact-Check Team

Most newsrooms outside the BBC and major national titles lack the resources to maintain a dedicated fact-checking unit. A distributed model — in which fact-checking responsibilities are shared across an editorial team rather than concentrated in a specialist unit — can deliver comparable rigour at lower cost, provided roles and processes are clearly defined.

• Claim monitor: During live coverage, one journalist is responsible solely for listening, reading, or watching the event in question and flagging specific verifiable claims for checking. This role requires discipline to focus on factual assertions rather than opinion.
• Research lead: A second journalist or researcher takes flagged claims and locates the primary evidence. This person should have quick access to official statistics databases, government publications, and academic sources.
• Senior editor (approver): A senior editor reviews the research and approves the verdict before it is published. This is especially important for claims that touch on contested or politically sensitive topics.
• Version control discipline: All fact-checks should be stored in a shared, version-controlled environment. If a verdict changes as new evidence emerges, every version should be logged with a timestamp and the reason for the update. A simple convention in a shared Google Doc is sufficient for small teams.

Larger operations may add dedicated OSINT researchers, data analysts, and legal reviewers. The Reuters Institute has published research on the internal structures of leading fact-checking organisations globally, which provides useful benchmarks for teams building out their own capability.

Ratings Systems and Transparency

One of the defining features of professional fact-checking is the use of structured ratings systems. Rather than simply labelling claims as “true” or “false,” leading organisations use a more granular set of verdicts that acknowledge complexity and partial accuracy.

• True / Mostly True: The claim is accurate and supported by primary evidence.
• Misleading: The claim contains accurate elements but is framed in a way that creates a false overall impression.
• Missing Context: The claim is technically accurate but omits information that would materially change how it is understood.
• False / Mostly False: The claim is not supported by primary evidence.
• Unverified: The claim cannot be confirmed or disproved with available evidence.

Channel 4 FactCheck typically applies a similar graduated scale. Regardless of the specific labels used, every ratings system should be applied consistently regardless of which political party or ideological position is being evaluated. Inconsistency in applying ratings is one of the most common criticisms levelled at fact-checking operations and the fastest way to undermine credibility.

The IFCN Code of Principlesrequires signatories to publish their ratings criteria openly so that readers can understand how verdicts are reached. Full Fact publishes its ratings framework on its website and applies it consistently across political parties and ideological positions — a requirement of the non-partisanship principle.

Transparency about methodology also means publishing corrections promptly and prominently. When a verdict changes, the original should be updated with a clear correction notice, not silently deleted. This reflects both good editorial practice and the requirements of IPSO's Clause 1.

Tools for Scale

Several tools have been built specifically to support fact-checking at volume:

• ClaimBuster: A machine-learning tool developed by the University of Texas at Arlington that analyses text and scores sentences by their “check-worthiness.” It can process transcripts of political speeches, debates, or press conferences and prioritise the claims most worth investigating.
• Google Fact Check Tools: The Fact Check Tools Explorer allows journalists to search a global database of published fact-checks using the ClaimReview schema. If a claim has already been investigated by another IFCN signatory, you may find an existing verdict before committing research time.
• CrowdTangle alternatives: Following Meta's deprecation of CrowdTangle, journalists tracking the spread of claims on social media have moved to tools including Brandwatch, Meltwater, and the academic CrowdTangle Archive. For smaller teams, the Meta Content Library (available to approved researchers) provides partial replacement functionality.
• Bellingcat OSINT guides: The Bellingcat resource library covers geolocation, satellite imagery analysis, social media investigation, and verification of documents and video. Many BBC Verify and Full Fact researchers cite Bellingcat guides as foundational references.
• Wayback Machine: Essential for capturing screenshots of pages and claims before they are deleted, and for establishing when a claim first appeared online.
• ONS and government statistics portals: The Office for National Statistics, HMRC, NHS England, and other UK government bodies publish datasets that underpin the majority of political fact-checks. Bookmark the relevant portals for your beat.

The ClaimReview schema, developed collaboratively by Google, Bing, and fact-checking organisations, allows publishers to mark up their fact-checks in structured data that search engines can surface directly in results. Implementing ClaimReview on your fact-check pages means your verdicts can appear in Google's Fact Check Explorer and in rich results on the SERP, increasing the reach of your work. UK publishers using Next.js or similar frameworks can add ClaimReview as a JSON-LD script block in the same way as Article and BreadcrumbList markup.

For a broader toolkit covering digital security and investigation, see our guide to open-source intelligence for UK journalists.

IPSO Accuracy Obligations and Corrections Policy

For the majority of UK print and online publishers, fact-checking is not optional — it is a regulatory obligation under IPSO's Editors' Code of Practice. Clause 1 (Accuracy) is the most frequently cited clause in IPSO adjudications and complaints.

• Clause 1(i): The press must take care not to publish inaccurate, misleading, or distorted information or images, including headlines not supported by the text.
• Clause 1(ii): A significant inaccuracy, misleading statement, or distortion must be corrected promptly and with due prominence.
• Clause 1(iii): The press, while free to be partisan, must distinguish clearly between comment, conjecture, and fact.

“Due prominence” is a key phrase. IPSO has found against publications that buried corrections in footnotes or published them significantly less prominently than the original inaccurate article. Best practice is to publish a correction that is at least as prominent as the article it corrects, ideally linked from the original page.

Where an inaccuracy arises from a source providing incorrect information, the publisher is not automatically absolved of responsibility. Clause 1(i) requires the press to “take care” — a standard that has been interpreted by IPSO to mean that reliance on a source does not discharge the obligation to verify. For statistics and factual claims, primary source verification is expected. This applies equally to information received from press officers, official spokespersons, and government departments.

Building a structured fact-checking workflow before publication is the most effective way to avoid IPSO complaints under Clause 1. The standard of “care” required by Clause 1(i) means that a post-publication correction, however prompt, does not necessarily satisfy IPSO if the original publication failed to take reasonable precautions. Documented pre-publication verification steps provide evidence that the required care was taken.

The BBC is regulated by Ofcom rather than IPSO, but its editorial guidelines impose comparable accuracy obligations. Broadcasters are additionally subject to Ofcom's Broadcasting Code, which requires due accuracy in news programmes.

The Scottish regulation picture differs slightly. IPSO covers Scottish publishers who have signed up to the scheme, but the Scots law of delict (rather than English defamation law) governs civil claims for inaccurate statements. Scottish publishers should be aware that the threshold tests and defences differ from those under the Defamation Act 2013, which does not apply in Scotland.

Online-only publishers that are not IPSO or IMPRESS signatories have no formal regulator mandating accuracy standards, but they remain subject to defamation law under the Defamation Act 2013. A false statement of fact that damages a person's reputation and is published to a third party is potentially actionable. Building a documented fact-checking workflow is therefore risk management as much as it is good journalism practice, regardless of regulatory status.

Practical Checklist

Before publishing any fact-check or claim-based story, work through all of the following steps:

• Identify the specific, verifiable claim being checked — strip away rhetoric and isolate the factual assertion.
• Search the Google Fact Check Tools Explorer to check whether an IFCN signatory has already investigated this claim.
• Locate the primary source for any statistic or assertion (official dataset, peer-reviewed study, government publication).
• Check that the figure cited in the claim matches the primary source and has not been taken out of context.
• Use ClaimBuster or manual triage to prioritise claims by check-worthiness before investing research time.
• For image or video claims, run a reverse image search and apply OSINT geolocation techniques.
• Apply the appropriate ratings category (True / Misleading / Missing Context / False / Unverified) and document your reasoning.
• Contact the person or organisation that made the claim before publication to give them an opportunity to respond.
• Have the verdict reviewed by a senior editor before publication.
• Publish all primary source links and explain your methodology in the fact-check itself.
• If the verdict changes after publication, update the article prominently with a dated correction note and log the change in your version history.
• Add ClaimReview JSON-LD structured data to the published fact-check so that search engines can surface your verdict in Fact Check Explorer and rich results.
• Retain all research notes and source documents in case the verdict is challenged by the subject, IPSO, or a legal representative.

Common Mistakes

• Checking the wrong claim. Politicians and public figures frequently embed a checkable statistic within a broader rhetorical framing. Ensure you identify and check the specific factual assertion, not just the general impression the speaker is trying to create.
• Relying on secondary sources. Fact-checking requires primary evidence. If your evidence is a newspaper article citing a report, find the original report. Secondary sources may themselves contain errors.
• Treating “unverified” as “true.” An inability to disprove a claim is not confirmation that it is accurate. When evidence is insufficient, say so clearly.
• Ignoring context and methodology behind statistics. Official figures are often reported without the caveats that statisticians consider essential. Check the ONS or relevant body's notes on data quality and comparability before citing figures.
• Publishing without giving the subject a chance to respond. Not only is this poor journalism practice, it may expose your publication to a defamation or accuracy complaint that could have been avoided.
• Burying corrections. An IPSO Clause 1 finding can result from a correction that is technically published but not sufficiently prominent. Match the prominence of the correction to the prominence of the original error.
• Assuming a previous fact-check is still current. Statistics change. A fact-check published six months ago may be based on superseded data. Always re-check the primary source before relying on an older verdict.
• Neglecting to implement ClaimReview structured data. If you are publishing fact-checks and not marking them up with ClaimReview JSON-LD, your verdicts are invisible to search engines' fact-check features. This is a missed opportunity for reach and impact.
• Applying ratings inconsistently across political parties or ideological positions. Inconsistency is the fastest way to attract accusations of bias. Apply your ratings criteria identically regardless of who made the claim.

Primary Sources

• Full Fact: Our Work and Methodology — Full Fact's published approach to claim selection, evidence gathering, and ratings
• BBC Verify Explained — the BBC's overview of its specialist verification unit's remit and methods
• IFCN Code of Principles — the five commitments required of all IFCN signatories, including Full Fact
• Reuters Institute for the Study of Journalism — research on fact-checking effectiveness, newsroom structures, and the Digital News Report
• IPSO Editors' Code of Practice — Clause 1 (Accuracy) and the corrections requirements for regulated UK publishers
• Google Fact Check Tools Explorer — searchable database of published fact-checks using the ClaimReview schema
• Bellingcat OSINT and Verification Guides — practical how-tos covering geolocation, satellite imagery, social media investigation, and document verification