L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Module 6 · Lesson 1

Writing a Verdict That People Trust

How PolitiFact, Snopes, and AP Fact Check frame their rulings — and why the language of a verdict matters as much as the evidence.
What separates a verdict readers believe from one they dismiss?

In March 2020, a viral claim spread on Facebook: "The US has tested more people for COVID-19 than any other country." PolitiFact assigned it False. But the verdict only landed because the team paired the ruling with a precise, citable sentence: "South Korea had tested more than 270,000 people by that point; the US had tested fewer than 75,000." The number was the verdict. The label alone would have been ignored.

Why Verdict Framing Is a Skill

Fact-checkers at organisations like PolitiFact, AFP Fact Check, and Full Fact all follow a common discipline: the verdict statement must stand alone. A reader who never reads the full article should be able to understand the ruling from a single sentence. This discipline protects against the most common failure in public fact-checking — verdicts that are technically correct but functionally useless because they are vague.

In 2019, Reuters Fact Check established an internal rule that every published verdict must contain: (1) the original claim restated accurately, (2) what evidence was found, and (3) the ruling with a specific justification. That three-part structure is now standard across most major fact-checking organisations affiliated with the International Fact-Checking Network (IFCN).

The Anatomy of a Strong Verdict Statement

Examine how the best-performing fact-check articles (by reader trust metrics) structure their verdicts. Three elements appear consistently:

Element 1 — Restate the Claim Precisely
Do not paraphrase in a way that weakens or strengthens the original. When Snopes investigated the 2016 claim that "George Soros owns voting machine companies," the verdict began by quoting the claim verbatim, not a summary of it. Paraphrasing introduced bias in early drafts that readers immediately challenged in comments.
Element 2 — Anchor the Verdict to a Specific Finding
Vague verdicts ("There is no strong evidence for this") score lower on reader trust than specific ones ("The company's SEC filings show no Soros ownership stake"). The difference was documented in a 2021 Reuters Institute study on fact-check credibility across 14 countries.
Element 3 — Use a Consistent Rating Label
PolitiFact's six-point scale (True, Mostly True, Half True, Mostly False, False, Pants on Fire) was designed in 2007 specifically because binary True/False ratings frustrated readers when claims were partially correct. The Washington Post Fact Checker uses 1–4 Pinocchios for similar reasons. Consistent scales build reader literacy over time.

How AI Tools Assist Verdict Writing

In 2022, the Duke Reporters' Lab surveyed 92 fact-checking organisations globally. Forty-one percent reported using AI tools — primarily to draft initial verdict summaries and flag logical inconsistencies in researcher write-ups. The AI did not determine the verdict; human editors retained that authority. The AI's role was editorial scaffolding: catching passive constructions that obscured responsibility ("mistakes were made") and suggesting more precise claim restatements.

Tools like the ClaimBuster system, developed at the University of Texas Arlington, automatically detect "check-worthy" claims in text and can draft preliminary verdict summaries that researchers then refine. As of 2023, ClaimBuster had analysed more than 20 million sentences from public discourse.

Key Principle

A verdict is not an opinion. It is a structured finding backed by citable evidence. The moment your verdict uses words like "seems," "appears," or "might be," you have written an inference, not a fact-check result. Readers detect this, and trust falls.

Key Terms

Verdict statementA standalone sentence that conveys the ruling, the evidence anchor, and the claim — readable without the full article.
Rating scaleA standardised set of labels (e.g., PolitiFact's six-point scale) that allows readers to compare verdicts across time and topics.
Claim restatementRepeating the original claim in the verdict with precision, neither softening nor amplifying it.
IFCNInternational Fact-Checking Network — the global body that sets editorial standards for accredited fact-checkers.

Lesson 1 Quiz

Writing a Verdict That People Trust · 4 questions
1. According to Reuters Fact Check's internal rule, how many elements must every published verdict contain?
Correct. Reuters established the three-part structure: restated claim, evidence found, and ruling with specific justification.
Not quite. Reuters' rule specifically requires three parts: the original claim restated accurately, what evidence was found, and the ruling with a specific justification.
2. Why did PolitiFact create a six-point rating scale rather than a binary True/False system?
Correct. PolitiFact's six-point scale was designed in 2007 specifically to address reader frustration with binary ratings on partially correct claims.
Not quite. The six-point scale was created because binary True/False ratings frustrated readers when claims were partially correct — it is about nuance, not compliance.
3. What does the 2021 Reuters Institute study suggest about specific vs. vague verdict anchors?
Correct. The Reuters Institute study across 14 countries found that verdicts anchored to specific evidence (like citing exact SEC filings) earned higher trust than vague formulations.
Not quite. The Reuters Institute study across 14 countries specifically found that specific, evidence-anchored verdicts scored higher on reader trust than vague ones.
4. What role did AI tools play in the 41% of fact-checking organisations that used them, according to the 2022 Duke Reporters' Lab survey?
Correct. The survey found AI was used for editorial scaffolding — drafting summaries and catching logical issues — while human editors retained verdict authority.
Not quite. The survey found AI served as editorial scaffolding: drafting initial summaries and flagging inconsistencies. Human editors always retained final verdict authority.

Lab 1 — Verdict Drafter

Practice structuring a fact-check verdict with AI feedback · 3 exchanges to complete

Your Task

You have investigated a claim. Now you need to write a verdict statement that follows the Reuters three-part structure: restate the claim precisely, anchor it to specific evidence, and state the ruling with justification. Submit a draft verdict below and the AI will critique it using IFCN editorial standards.

Try drafting a verdict for this claim: "A photo circulating in 2022 showed a Ukrainian soldier holding a Nazi flag." — Use evidence you can cite (reverse image searches, date metadata, news archives). Write your three-part verdict statement and submit it for critique.
Verdict Critique Assistant
IFCN Standards
Welcome to the Verdict Drafter. Write a three-part verdict statement for the claim about the 2022 photo. Include: (1) the claim restated precisely, (2) specific evidence you found, and (3) your ruling with justification. I will critique your draft using IFCN editorial standards and suggest improvements. Go ahead — paste or type your verdict draft.
Module 6 · Lesson 2

Choosing the Right Platform for Your Fact-Check

From Twitter threads to long-form articles — how fact-checkers at AFP, BBC Reality Check, and Bellingcat match format to audience.
Does a true finding spread as far as a false one — and can platform choice change that?

When Bellingcat published its initial findings linking a Russian Buk missile launcher to the MH17 crash in July 2014, it chose a medium-length blog post with embedded high-resolution images. The format was deliberate. Investigative lead Eliot Higgins later explained in interviews that Twitter would have lost the evidentiary chain, and a PDF report would have reached only specialists. The blog post format — image-rich, linkable, accessible — reached both mass audiences and intelligence analysts simultaneously.

By 2016, that single post had been cited in the Dutch Safety Board report, shared more than 140,000 times, and translated into six languages. Format was not incidental to impact. It was the mechanism of impact.

Platform Characteristics That Shape Reach

A 2023 study by First Draft News analysed 4,200 fact-check articles published across eight major organisations. It found that platform format — not just content quality — was the second strongest predictor of reach, after headline clarity. The organisations that consistently reached the largest audiences adapted their findings to multiple formats simultaneously.

Long-Form Article
Best for complex, multi-part claims. Snopes and PolitiFact use this for claims with significant context. Average reading time 6–12 minutes. High trust, lower initial reach.
Twitter / X Thread
Best for time-sensitive breaking claims. AFP Fact Check uses numbered threads of 5–8 tweets with a final verdict tweet pinned at top. High immediate reach, low context depth.
Video (YouTube / TikTok)
BBC Reality Check and Full Fact use 60–180 second videos for visual claims (photos, videos). Highest engagement with under-35 audiences. Requires visual evidence to be effective.
WhatsApp Broadcast
Used in India (Boom Live), Nigeria (Dubawa), and Brazil (Agência Lupa) where misinformation spreads via messaging apps. Push model — audience must opt in.
Instagram Carousel
Lead slide shows verdict label; subsequent slides show evidence steps. Used by AFP, Reuters, and Rappler for photo and video misinformation. High engagement with visual claims.
Podcast / Audio
FactCheck.org and BBC More or Less use audio for statistical and economic claims. Effective for audiences that commute. Evidence must be explained verbally — no visuals.

The Cross-Platform Rule

In 2022, Full Fact (UK) published an internal review of its distribution strategy. It found that fact-checks published only in long-form article format reached an average of 12,000 readers. The same fact-checks adapted into a Twitter thread plus an Instagram carousel reached an average of 340,000 people. The evidence did not change. The format tripled the potential audience eleven times over.

The practical implication for fact-checkers: write your long-form article first (it is your evidence record), then strip it into platform-specific formats. The article is the source of truth. The thread, carousel, and video are distribution vehicles. Never reverse this order — distribution-first formats tend to lose evidentiary precision.

Documented Case — AFP Fact Check 2021

When a viral photo claimed to show Brazilian election fraud in October 2022, AFP Fact Check published a long-form article, then immediately adapted it into a five-tweet thread (in Portuguese), an Instagram carousel with reverse-image-search screenshots, and a WhatsApp broadcast to 80,000 subscribers in Brazil. The fact-check reached 2.1 million people within 48 hours — a reach AFP attributed directly to the multi-platform adaptation strategy.

AI's Role in Multi-Platform Adaptation

Tools like Logically AI and Google's Fact Check Explorer now offer automated format adaptation: paste a long-form fact-check article and receive a Twitter-optimised summary, an Instagram caption, and a WhatsApp message draft. These tools were used by 23 IFCN-accredited organisations as of the 2023 Duke Reporters' Lab survey. Editors report the AI drafts require significant revision but reduce adaptation time by approximately 60%.

Key Terms

Distribution vehicleA platform-specific format (thread, carousel, video) used to spread findings from a primary evidence record.
Evidence recordThe long-form article that contains full citation chains, methodology, and ruling — the authoritative source for all adapted formats.
Cross-platform adaptationReformatting the same fact-check findings for different platforms while preserving evidentiary accuracy.

Lesson 2 Quiz

Choosing the Right Platform · 4 questions
1. What did Bellingcat's Eliot Higgins identify as the reason for choosing a blog post format over Twitter for the MH17 investigation?
Correct. Higgins explained that Twitter would lose the evidentiary chain while a PDF would only reach specialists — the blog post format reached both mass audiences and intelligence analysts.
Not quite. Higgins' reasoning was about evidentiary integrity and audience reach: Twitter loses the evidence chain, PDFs reach only specialists, but a blog post serves both.
2. According to the First Draft News 2023 study, what was the second strongest predictor of a fact-check's reach after headline clarity?
Correct. The First Draft News study of 4,200 fact-check articles found platform format was the second strongest predictor of reach, after headline clarity.
Not quite. Platform format — not follower count, celebrity subject, or article length — was the second strongest predictor of reach according to the First Draft News 2023 study.
3. What did Full Fact's 2022 internal review find about long-form-only vs. multi-platform distribution?
Correct. Full Fact's review found long-form-only fact-checks averaged 12,000 readers; the same content adapted into thread plus carousel averaged 340,000.
Not quite. Full Fact found the multi-platform approach (article + Twitter thread + Instagram carousel) increased average reach from approximately 12,000 to 340,000 readers.
4. What is the correct order of operations when creating multi-platform fact-check content, according to the lesson?
Correct. The article is the source of truth and evidence record. Platform formats are distribution vehicles. Reversing this order loses evidentiary precision.
Not quite. The long-form article must come first — it is the evidence record. Threads, carousels, and videos are distribution vehicles derived from it. Distribution-first formats lose evidentiary precision.

Lab 2 — Platform Adaptation Planner

Design a multi-platform distribution strategy for a real fact-check · 3 exchanges to complete

Your Task

You have just completed a fact-check. Now you need to plan how to distribute it across platforms. Using what you learned from the AFP Brazil election case and Full Fact's strategy, design a multi-platform distribution plan. Tell the AI which platforms you would use, in what order, and why — for the scenario below.

Scenario: You have just published a 1,200-word fact-check article debunking a viral TikTok video that falsely claims a 2023 wildfire in Canada was deliberately started. Your audience is primarily English-speaking, ages 18–45. Design your multi-platform distribution plan and explain your choices to the AI.
Platform Strategy Advisor
Distribution Planning
Welcome to the Platform Adaptation Lab. Describe your multi-platform distribution plan for the Canadian wildfire fact-check. Tell me: which platforms you would use, in which order, what format for each, and your reasoning. I'll give you feedback on whether your strategy matches what organisations like AFP and Full Fact have found most effective.
Module 6 · Lesson 3

Handling Pushback and Bad-Faith Responses

What happens when subjects of fact-checks fight back — lessons from PolitiFact's legal challenges, Snopes lawsuits, and coordinated harassment campaigns.
When a fact-check goes public, who pushes back — and how should you respond?

In 2016, a content company called Proper Media sued Snopes co-founder David Mikkelson, alleging in part that Snopes had unfairly profited from licensed content. Simultaneously, a coordinated social media campaign began labelling Snopes as "left-wing" and "unreliable." The campaign cited the lawsuit as evidence of institutional corruption — even though the lawsuit was a business dispute between co-owners, entirely unrelated to editorial integrity.

Snopes documented the conflation publicly, publishing a detailed explanation separating the business dispute from editorial processes. Independent media analysts including those at PolitiFact and the Poynter Institute confirmed no editorial compromise had occurred. The episode became a textbook case of source-attack misinformation — attempting to discredit a fact-checking organisation by associating it with an unrelated controversy.

Categories of Pushback

Research by the Shorenstein Center at Harvard (2020) identified four distinct categories of responses that fact-checkers receive after publishing verdicts, ranked by frequency:

  • Legitimate corrections (8%). The subject provides new evidence or identifies a factual error in the fact-check itself. IFCN standards require a published correction with a timestamp.
  • Good-faith disagreement (19%). The subject contests the interpretation without new evidence. Appropriate response: publish the subject's statement with the fact-check's maintained verdict, clearly labelled.
  • Source-attack (38%). The subject or their supporters attack the fact-checker's funding, political affiliation, or personal history rather than the evidence. Appropriate response: do not engage with the ad hominem; restate the evidence chain.
  • Coordinated harassment (35%). Organised campaigns targeting the fact-checker personally via social media, email, or phone. This is documented to have affected at least 64% of IFCN-accredited organisations as of 2022 (Committee to Protect Journalists data).

The PolitiFact Response Protocol

PolitiFact publishes its correction policy publicly. If a subject provides new evidence after a verdict is published, editors review the evidence within 48 hours and either update the verdict (with a documented edit history) or publish a statement explaining why the new evidence does not change the finding. This transparency is central to IFCN accreditation.

In 2020, when the Trump campaign formally disputed a PolitiFact ruling on COVID test numbers, PolitiFact published the campaign's full written response alongside the original fact-check, with a paragraph-by-paragraph editorial response. No verdict was changed because no new evidence was provided — only rhetorical challenge. The exchange was itself cited in journalism schools as an example of transparent pushback management.

The Harassment Problem

A 2022 report by the Coalition for Women in Journalism found that female fact-checkers receive harassment at 3.4 times the rate of male colleagues after publishing politically sensitive verdicts. The International Center for Journalists (ICFJ) launched the OnTheLine helpline in 2021 specifically to support journalists facing online harassment, and by 2023 had assisted more than 800 cases, many involving fact-checkers.

AI Tools for Monitoring Pushback

Several fact-checking organisations now use AI-assisted monitoring to detect when their published fact-checks are being misrepresented online. The tool Meltwater, used by AFP and Reuters, monitors social media for mentions of a published fact-check and flags instances where the verdict has been inverted or stripped of context — for example, a headline screenshot shared without the verdict, or a partial quote from the article used to suggest the opposite conclusion.

In 2021, AFP Fact Check used Meltwater to identify that a Brazilian political network was sharing screenshots of an AFP fact-check with the verdict cropped out, making it appear AFP had confirmed rather than debunked the false claim. AFP published a follow-up article documenting the cropping tactic within 24 hours.

Key Terms

Source-attack misinformationAttempting to discredit a fact-checker by associating them with unrelated controversies rather than challenging the evidence.
Legitimate correctionNew evidence or a factual error in the fact-check itself, requiring a published correction with timestamp per IFCN standards.
Coordinated harassmentOrganised campaigns targeting fact-checkers personally, documented in 64% of IFCN-accredited organisations by 2022.
Verdict inversionSharing screenshots or excerpts of a fact-check in a way that reverses its actual conclusion.

Lesson 3 Quiz

Handling Pushback and Bad-Faith Responses · 4 questions
1. According to the Shorenstein Center research, what percentage of post-verdict responses are legitimate corrections with new evidence?
Correct. Only 8% of post-verdict responses are legitimate corrections with new evidence. The majority (38%) are source-attacks and 35% are coordinated harassment.
Not quite. The Shorenstein Center found only 8% of responses are legitimate corrections. Source-attacks (38%) and coordinated harassment (35%) are far more common.
2. What tactic did a Brazilian political network use against an AFP fact-check in 2021, as detected by Meltwater?
Correct. The network shared AFP fact-check screenshots with the verdict cropped out, making it appear AFP had confirmed rather than debunked the false claim — a classic verdict inversion tactic.
Not quite. The tactic was verdict inversion: sharing screenshots with the verdict cropped out so the article appeared to confirm the false claim rather than debunk it.
3. What does IFCN standards require when a fact-checker publishes a correction?
Correct. IFCN standards require a published correction with a timestamp — maintaining an auditable edit history is fundamental to accreditation.
Not quite. IFCN standards specifically require a published correction with a timestamp. Deleting or rewriting without documentation violates transparency standards.
4. How did PolitiFact respond when the Trump campaign disputed a 2020 COVID verdict without providing new evidence?
Correct. PolitiFact maintained the verdict (no new evidence was provided) but published the campaign's full written response with an editorial reply — a model of transparent pushback management.
Not quite. PolitiFact published the campaign's full response with a paragraph-by-paragraph editorial reply, but maintained the verdict because no new evidence was provided — only rhetorical challenge.

Lab 3 — Pushback Response Trainer

Practice responding to hostile post-verdict challenges using IFCN protocol · 3 exchanges to complete

Your Task

You have published a fact-check. Now you are receiving pushback. The AI will play the role of a subject disputing your verdict. Identify the type of pushback (legitimate correction, good-faith disagreement, source-attack, or coordinated harassment) and draft an appropriate response following IFCN protocol.

Setup: You published a verdict rating "False" a politician's claim that crime rates doubled in their city last year. The politician's communications team has just sent this response: "Your organisation is funded by left-wing foundations. This fact-check is politically motivated. We demand a retraction." — Identify the pushback type and write your IFCN-compliant response to the AI.
Pushback Response Coach
IFCN Protocol
You have received a pushback message from a politician's team. Your job is to: (1) identify which of the four Shorenstein categories this falls into, (2) draft an IFCN-compliant response, and (3) state whether you would change the verdict. Write your analysis and response draft — I'll assess whether it follows best practice from PolitiFact and Full Fact's documented approaches.
Module 6 · Lesson 4

Building a Fact-Check That Travels

How CheckDesk, CrossCheck, and collaborative networks amplify individual verdicts into durable public knowledge records.
What makes a fact-check findable, citable, and useful a year after it is published?

In the weeks before France's 2017 presidential election, First Draft News coordinated a coalition of 37 newsrooms — including Le Monde, AFP, BuzzFeed News France, and regional outlets — under the CrossCheck project. Every participating organisation agreed to use a shared taxonomy of claim types and verdict labels, so that a claim debunked by one newsroom would be instantly findable by readers of any other participating outlet.

The coalition processed 300 individual fact-check submissions in seven weeks. Post-election analysis by the Reuters Institute found that claims debunked within the CrossCheck network spread at significantly lower rates on French social media than similar debunked claims in the 2016 US election, where no coordinated network existed. Coordination multiplied the impact of each individual fact-check.

Why Isolated Fact-Checks Fade

A 2022 study by the Empirical Studies of Conflict project at Princeton analysed 18,000 fact-check articles published between 2015 and 2021. It found that 73% of fact-check articles received fewer than 500 organic page views after their first week of publication, regardless of their accuracy or importance. The problem was discoverability: most fact-checks were not indexed in ways that allowed readers encountering a false claim months later to find the existing debunking.

Google's Fact Check Tools were introduced in 2017 specifically to address this. When news organisations use the ClaimReview structured data markup — a free schema that embeds verdict metadata into article HTML — their fact-checks become indexable by Google's Fact Check Explorer. As of 2023, more than 11,000 fact-checking pages globally use ClaimReview markup.

The ClaimReview Standard

ClaimReview is a technical schema developed by schema.org that allows fact-checkers to embed structured metadata in their articles. The metadata includes: the claim text, the claimant, the claim date, the verdict label, and the rating value. When embedded correctly, Google Search displays the verdict directly in search results — a fact-check label appears next to the original false claim in Google Search, giving readers the debunking at the moment they encounter the claim.

The Duke Reporters' Lab maintains a real-time database of ClaimReview-marked fact-checks globally. Their 2023 annual census found 68 countries with at least one IFCN-accredited fact-checking organisation using ClaimReview, and estimated that ClaimReview-enabled fact-checks received on average 340% more organic search traffic than equivalent fact-checks without the markup.

Case Study — Poynter International Fact-Checking Network

The IFCN's CoronaVirusFacts Alliance, formed in March 2020, united 100 fact-checking organisations across 45 countries to collectively debunk COVID-19 misinformation. Every partner used ClaimReview markup and shared verdicts via a central database. By December 2020, the alliance had published more than 9,000 individual fact-checks, all cross-searchable. The WHO cited the alliance database in its own communications as a reference source — making individual journalists' fact-checks part of a global public health evidence record.

AI and Claim Matching

One of the most significant practical benefits of AI in fact-checking distribution is claim matching: automatically detecting when a claim circulating online has already been fact-checked and surfacing the existing verdict. Google's Fact Check Explorer API, ClaimBuster, and the Full Fact automated fact-checking system all perform some version of this function.

In 2021, Full Fact demonstrated that their automated system could match a new instance of a circulating claim to an existing fact-check in under 3 seconds, compared to an average of 47 minutes for a human researcher to perform the same task. The system was used to respond to claims in UK parliamentary debates in near-real time during the 2021 pandemic response.

Building Your Own Durable Record

Even without institutional infrastructure, individual fact-checkers can build durable records. The steps followed by student journalism programmes at Northwestern, Columbia, and City University London include:

  • Write the full article with all sources hyperlinked and archived (use archive.org or archive.ph for source preservation).
  • Add ClaimReview markup to your article's HTML metadata using Google's Structured Data Markup Helper (free tool).
  • Submit your fact-check to Google's Fact Check Tools index via the Search Console.
  • Archive the fact-check itself at archive.org so it persists even if your publication changes.
  • Tag the claim in an open database (e.g., the Duke Reporters' Lab database or MediaBias/FactCheck) so it is findable by other researchers.
  • Review the fact-check at six-month intervals and update if new evidence emerges, with a documented edit log.

Key Terms

ClaimReviewA schema.org structured data standard that embeds fact-check verdict metadata into article HTML, making verdicts indexable by Google Search.
Claim matchingAI-assisted detection of when a circulating claim matches an already fact-checked claim, surfacing existing verdicts automatically.
CrossCheck modelA collaborative fact-checking network where multiple newsrooms use shared taxonomy and verdict labels to multiply the reach of individual fact-checks.
CoronaVirusFacts Alliance100-organisation IFCN coalition formed in 2020 that produced 9,000+ ClaimReview-marked COVID fact-checks searchable by the WHO.

Lesson 4 Quiz

Building a Fact-Check That Travels · 4 questions
1. What did the 2022 Princeton Empirical Studies of Conflict analysis find about fact-check discoverability?
Correct. The Princeton study found 73% of fact-check articles received fewer than 500 organic page views after their first week, regardless of accuracy or importance — a systemic discoverability failure.
Not quite. The Princeton analysis found that 73% of fact-check articles — regardless of accuracy — received fewer than 500 organic page views after their first week. Discoverability was the core problem.
2. According to the Duke Reporters' Lab 2023 census, how much more organic search traffic do ClaimReview-marked fact-checks receive compared to unmarked equivalents?
Correct. The Duke Reporters' Lab found ClaimReview-enabled fact-checks received on average 340% more organic search traffic than equivalent fact-checks without the markup.
Not quite. The Duke Reporters' Lab 2023 census found the figure was approximately 340% more organic traffic — a very significant discoverability advantage.
3. What was the CrossCheck coalition's documented impact on the 2017 French election compared to the 2016 US election?
Correct. The Reuters Institute found CrossCheck-debunked claims spread at significantly lower rates on French social media than similar debunked claims in the 2016 US election — where no coordinated network existed.
Not quite. The Reuters Institute's post-election analysis found CrossCheck-debunked claims spread at significantly lower rates than comparable claims debunked without coordination — demonstrating the network multiplier effect.
4. In Full Fact's 2021 demonstration, how quickly could their automated claim-matching system find an existing fact-check for a newly circulating claim?
Correct. Full Fact's automated system matched claims to existing fact-checks in under 3 seconds, versus an average of 47 minutes for a human researcher performing the same task.
Not quite. Full Fact's demonstration showed the system performed claim matching in under 3 seconds — compared to an average of 47 minutes for a human researcher. This enabled near-real-time responses during UK parliamentary debates.

Lab 4 — Durable Record Builder

Plan a ClaimReview strategy and long-term preservation plan for a fact-check · 3 exchanges to complete

Your Task

You have written a complete fact-check article. Now you need to make it durable and discoverable for the long term. Using the six-step framework from student journalism programmes at Northwestern, Columbia, and City University London, build a preservation and discoverability plan for the scenario below. Discuss your choices with the AI.

Scenario: You have just published a fact-check debunking the claim that "5G towers caused the first wave of COVID-19." The article is on a small independent journalism website with no institutional support. Outline your complete ClaimReview strategy, source archiving plan, and six-month review protocol — and explain why each step matters.
Durable Record Advisor
ClaimReview · Preservation
Welcome to the Durable Record Builder. Describe your full preservation and discoverability plan for the 5G/COVID fact-check. Walk me through each step you would take — ClaimReview markup, source archiving, index submission, long-term review protocol — and explain your reasoning for each. I'll assess it against the standards used by IFCN-accredited organisations and suggest improvements.

Module 6 Test

Broadcast Your Truth Verdict · 15 questions · Pass mark 80%
1. What are the three parts of Reuters Fact Check's mandatory verdict structure?
Correct. Reuters requires: the original claim restated accurately, what evidence was found, and the ruling with a specific justification.
Reuters' three-part structure is: restated claim, evidence found, and ruling with justification — the foundation of a standalone verdict statement.
2. Which organisation's 2021 study across 14 countries found that specific evidence-anchored verdicts earned higher reader trust than vague ones?
Correct. The Reuters Institute 2021 study across 14 countries documented the trust advantage of specific, evidence-anchored verdicts.
The Reuters Institute conducted the 2021 study across 14 countries that found specific evidence anchors drove higher reader trust.
3. How many fact-checking organisations were surveyed by the Duke Reporters' Lab in 2022, and what percentage used AI tools?
Correct. 92 organisations were surveyed and 41% reported using AI tools for tasks like drafting verdict summaries and flagging inconsistencies.
The Duke Reporters' Lab 2022 survey covered 92 organisations; 41% used AI tools primarily for editorial scaffolding.
4. What format did Bellingcat choose for its MH17 findings, and why was it strategically superior to alternatives?
Correct. The image-rich blog post format preserved the evidentiary chain (unlike Twitter) while remaining accessible to non-specialists (unlike a PDF).
Bellingcat chose a blog post because it preserved the evidentiary chain while reaching both mass audiences and intelligence analysts — Twitter loses evidence chains; PDFs only reach specialists.
5. According to Full Fact's 2022 internal review, what average readership did long-form-only fact-checks achieve?
Correct. Long-form-only fact-checks averaged approximately 12,000 readers — compared to 340,000 for the same content distributed across multiple platforms.
Full Fact found long-form-only fact-checks averaged about 12,000 readers. The same content adapted to multiple platforms averaged 340,000.
6. What is "verdict inversion" as detected in the 2021 AFP Brazil case?
Correct. Verdict inversion involves sharing screenshots with the verdict portion cropped out, making it appear the fact-checking organisation confirmed the false claim.
Verdict inversion specifically means sharing screenshots with the verdict cropped out — so the article appears to confirm rather than debunk the claim. AFP detected this via Meltwater monitoring.
7. According to Shorenstein Center research, which category of post-verdict pushback is most frequent?
Correct. Source-attacks — attacking the fact-checker's funding or affiliation rather than the evidence — were the most common response at 38%.
Source-attacks (38%) are the most frequent pushback category — more common than coordinated harassment (35%), good-faith disagreement (19%), or legitimate corrections (8%).
8. What specific tactic characterised the pushback against Snopes during the 2016–2017 Proper Media lawsuit?
Correct. The campaign associated the business dispute with editorial bias — a source-attack that had nothing to do with the actual legal case's subject matter.
The pushback was source-attack misinformation: conflating an unrelated business dispute between co-owners with claims of editorial bias to label Snopes unreliable. PolitiFact and Poynter confirmed no editorial compromise.
9. ClaimReview is a standard developed by which organisation to embed verdict metadata in article HTML?
Correct. ClaimReview is a schema.org standard — though Google Fact Check Tools adopted it and made it searchable via Google Search and Fact Check Explorer.
ClaimReview is a schema.org standard. Google adopted it for Fact Check Explorer, but schema.org developed the specification.
10. How many newsrooms participated in the CrossCheck coalition during the 2017 French election?
Correct. CrossCheck united 37 newsrooms, including Le Monde and AFP, under a shared taxonomy and verdict label system. They processed 300 submissions in seven weeks.
CrossCheck coordinated 37 newsrooms for the 2017 French election — including Le Monde, AFP, and BuzzFeed News France — under shared claim taxonomy and verdict labels.
11. How many COVID-19 fact-checks did the IFCN CoronaVirusFacts Alliance produce by December 2020?
Correct. The CoronaVirusFacts Alliance — 100 organisations in 45 countries — published more than 9,000 ClaimReview-marked fact-checks by December 2020. The WHO cited the database as a reference source.
The CoronaVirusFacts Alliance produced more than 9,000 fact-checks by December 2020, all cross-searchable via ClaimReview markup. The WHO cited the alliance database in its own communications.
12. What is the correct order for multi-platform fact-check production, per the lesson?
Correct. The article is the evidence record and source of truth. All distribution formats derive from it. Distribution-first approaches lose evidentiary precision.
The long-form article must come first — it is the evidence record. Threads, carousels, and videos are distribution vehicles derived from it, not the other way around.
13. What did PolitiFact do when the Trump campaign disputed its 2020 COVID verdict without providing new evidence?
Correct. PolitiFact maintained the verdict (no new evidence provided) but published the campaign's full response with an editorial reply — a transparency model cited in journalism schools.
PolitiFact published the campaign's full response with a paragraph-by-paragraph reply, but maintained the verdict because only rhetorical challenge — not new evidence — was offered. This became a transparency model.
14. According to the 2022 Coalition for Women in Journalism report, at what rate do female fact-checkers receive harassment compared to male colleagues after politically sensitive verdicts?
Correct. The 2022 Coalition for Women in Journalism report found female fact-checkers receive harassment at 3.4 times the rate of male colleagues after politically sensitive verdicts.
The Coalition for Women in Journalism found the rate was 3.4 times higher for female fact-checkers — prompting the ICFJ to launch the OnTheLine helpline in 2021.
15. How quickly did Full Fact's 2021 automated claim-matching system perform compared to human researchers, and in what setting was this demonstrated?
Correct. Full Fact's system matched claims in under 3 seconds (vs. 47 minutes for a human) and was used to respond to claims in UK parliamentary debates in near-real time.
Full Fact's 2021 demonstration showed under 3 seconds vs. 47 minutes for human researchers — demonstrated during UK parliamentary debates during the 2021 pandemic response.