Fact-Checking Guide
Verify claims and sources systematically for accuracy.
Usage
- Identify the specific claim to verify (who said what, when, with what evidence)
- Trace the claim to its original source — not the article that cites another article
- Evaluate the source's credibility and potential biases
- Check if the statistics or data are presented accurately and in context
- Look for corroboration from independent, reputable sources
Examples
- Statistical claim verification: Claim: "Crime increased 50% last year." Verify: 50% of what? (If from 2 incidents to 3, that's 50% but meaningless.) What type of crime? What geographic area? What time comparison? Is this from official government data or a survey? Is the baseline year unusually low (making the increase look larger)? Always ask: "50% of what, measured how, by whom?"
- Source evaluation (CRAAP test): Currency (when was it published?), Relevance (does it address your question?), Authority (who is the author, what are their credentials?), Accuracy (is it supported by evidence, peer-reviewed?), Purpose (inform, persuade, sell, entertain?). Score each 1-5. Below 15/25: unreliable
- Lateral reading technique: Don't just evaluate a source by reading it — open new tabs and search for what others say ABOUT that source. A professional-looking website could be a known misinformation outlet. Check: Wikipedia entry for the organization, Media Bias/Fact Check rating, Snopes or PolitiFact for specific claims
- Image/video verification: Reverse image search (Google Images, TinEye) to find the original context. Check EXIF data for date and location. Look for visual inconsistencies (shadows, edges, proportions) suggesting manipulation. InVID browser extension analyzes video metadata and keyframes
Guidelines
- Extraordinary claims require extraordinary evidence — the more surprising a claim, the more scrutiny it deserves
- Primary sources (original research, official data, court documents) are more reliable than secondary sources (news articles, blog posts, social media)
- Beware of cherry-picked statistics: a single data point can be accurate but deeply misleading without context
- "A study found..." means nothing without: who conducted it, sample size, methodology, peer review status, and whether it's been replicated
- Your own biases are the biggest fact-checking obstacle — we instinctively accept claims that confirm our beliefs and scrutinize those that don't
- When in doubt, say "I don't know yet" rather than sharing unverified information — uncertainty is more honest than false confidence