What distinguishes a verifiable claim from a headline
Before opening any link, it is useful to distinguish between a verifiable fact and an opinion or a sensational headline. A verifiable claim usually contains concrete data: dates, figures, organization names, or measurable results. A headline, by contrast, may appeal to emotion or urgency without offering elements that allow for cross-checking. This initial distinction determines the depth of verification you need: if you only want a general explanation, confirming that the publication is reputable is enough; if you need to cite the data, you must trace its origin.
The SIFT method, developed by Mike Caulfield since 2017, proposes four moves to tackle this task. It is not a list of visual characteristics of a page, but a sequence of active actions. The key is that each move relies on a specific search technique, allowing you to move from intuition to a documented evaluation. It is not a one-time test, but a habit that adapts to the context: a quick read to share a news item or a deep investigation for academic work.
Why intent matters
The intent with which you consume information defines the level of verification. Caulfield notes that both approaches are equally useful: quick and superficial investigations constitute most of our web activity, allowing us to spend more time on what truly matters. However, in both cases, periodically stopping to re-evaluate the search strategy is fundamental to avoid falling into a distraction loop.
The four SIFT moves and lateral reading
The acronym SIFT summarizes the four steps: Stop, Investigate the source, Find better coverage, and Trace claims, quotes, and media back to the original context. The first step is the simplest and the most ignored: before reading, ask yourself if you know the website or source. If you do not, do not read or share until you use the other moves to understand what you are seeing.
The second move, investigating the source, does not require exhaustive journalistic research, but it does require about sixty seconds to identify who is publishing. If you are reading an economics article written by an award-winning economist, you should know this before reading. Similarly, if you see a video about the benefits of a product published by the industry that manufactures it, that contextual information is crucial for interpreting its message. Knowing the expertise and agenda of the source helps decide if it is worth your time and better understand its reliability.
Lateral reading as a central technique
The third move, finding better coverage, sometimes involves ignoring the source that reached you to seek more reliable reports or analyses on the same claim. This is where lateral reading comes in, a strategy identified in the research by Wineburg and McGrew. Unlike vertical reading, which involves staying within a website to evaluate its reliability, lateral reading involves leaving the site after a quick scan and opening new tabs to judge the credibility of the original source. According to the study, professional fact-checkers reached better-grounded conclusions in a fraction of the time that historians and students did, who often fell for traps of logos and domain names that look official.
The fourth move, tracing the original context, is essential when content has been stripped of its context. A video of a fight may show a person as an aggressor, but what happened before? An image may look real, but the caption may be misleading. A claim about a medical treatment may be based on a research finding, but does the cited article really say that? Tracing the claim to its source allows you to see it in its original context and evaluate whether the version you saw was presented accurately.
The four moves of SIFT
Use these four actions to check a digital claim and restore context. The sequence is a practical framework, not a guarantee that a claim is true.
-
1. Pause
Check your purpose and what you know about the source.
-
2. Investigate
Identify the publisher, expertise, and interests.
-
3. Find coverage
Seek stronger reporting and assess whether it is independent.
-
4. Trace context
Return to the original claim, quotation, or media.
SIFT supports evaluation but cannot guarantee accuracy; conclusions depend on the claim and evidence examined.
Sources and references1
How to distinguish independent coverage from repetition
A common mistake is assuming that if multiple media outlets publish the same information, it is more likely to be true. This logic is incorrect. Repetition is not corroboration. Many websites copy the same press release or are syndicated with each other, meaning they are sharing the same original source without independence. To distinguish independent coverage, you must identify if the media have separate newsrooms, if they investigated on their own, or if they simply retransmitted an agency wire.
The reputation of a domain or the aesthetics of a site do not guarantee truth. A site with a professional design and a .org domain may be funded by corporate interests. The Stanford study on the Civic Online Reasoning (COR) curriculum highlights that students learned to critically ignore low-quality sites and not be fooled by misleading cues like search result position or domain type. A site appearing on the first page of search results may be optimized for search engines (SEO) and still spread false information.
Criteria for evaluating independence
To evaluate if coverage is independent, consider these criteria:
- Origin of information: Does the media cite a primary source or retransmit a press release?
- Own reporting: Does the article contain analysis, interviews, or exclusive data from the media?
- Editorial context: Does the media have a clear editorial line and a history of corrections?
- Funding: Who funds the media or the organization publishing the information?
Hypothetical example of tech verification
Imagine a tech blog claims that a new company has developed a chip that reduces server energy use by 90% without sacrificing performance. The figure and comparison make this a specific, checkable claim. Applying SIFT, pause before sharing and investigate who published it, what expertise or interests they may have, and what context might help assess the claim. An unfamiliar source is a reason to look further, not proof that its claim is false.
To find better coverage, search for reporting or analysis of the same claim and check whether the sources are independent or rely on the same original material. Then trace the claim to its original context, such as the research or evidence cited. If you cannot corroborate it, treat it as unverified, rather than presenting it as established fact. Lack of corroboration alone does not show that a claim is false.
Process log
A log can record the steps you took, what you found, and the conclusion you reached. This makes the evaluation process easier to review and share.
Reusable log sheet
To make the process reproducible, you can use a log sheet that captures the key steps. This sheet is not a rigid form, but a guide to document your evaluation. You can use it in a text document, a spreadsheet, or a note in your knowledge management system.
Sheet structure
- Claim: Describe the specific claim you are evaluating.
- Original source: URL of the site or publication where you found the claim.
- Step 1: Stop: Did you know the source? What was your intent in reading it?
- Step 2: Investigate the source: Who is publishing? What is their expertise and agenda?
- Step 3: Find better coverage: Did you find other sources? Are they independent?
- Step 4: Trace to original context: Did you trace the claim to its origin? What did you find?
- Conclusion: Is the claim verifiable, unverifiable, or false? What practical decision do you make?
- Date: Date of the evaluation.
This sheet lets you audit your own process and share it with others. If someone asks why you did not share a news item, you can use the sheet to show the steps you followed.
Commented readings and limits
Mike Caulfield's article, "SIFT (The Four Moves)", published in 2019, is the primary source of the method. Caulfield, a professor at Washington State University Vancouver, presents SIFT as an alternative to the CRAAP test, which he considers insufficient for the modern web. The article is licensed under Creative Commons Attribution 4.0 International, allowing its use and adaptation with attribution. It is not an academic document, but a practical guide based on Caulfield's teaching experience since 2017.
The study by Wineburg and McGrew, "Lateral Reading and the Nature of Expertise", published in 2019 in Teachers College Record, is academic research that identifies lateral reading as the key strategy of professional fact-checkers. The study sampled 45 experienced Internet users: 10 PhD historians, 10 professional fact-checkers, and 25 Stanford students. Participants thought aloud while evaluating live websites. Historians and students often fell for the traps of logos and domain names, while professional fact-checkers reached better-grounded conclusions in less time. This study is not a universal validation of lateral reading, but an observation of the strategies of a specific group of experts.
The Stanford study on the Civic Online Reasoning (COR) curriculum was a randomized experiment involving about 500 high school students in an urban Midwestern school district. Students who received the curriculum took six 50-minute lessons and were twice as likely to identify questionable websites as they had been before instruction. Even afterward, they earned only half the possible points on the test. The study has limits: it examined one district and does not establish that lateral reading is the only effective strategy. The study was funded by Google.org.
Limits of the evidence
It is important to distinguish between lack of corroboration and falsity. A claim may be true but lack multiple coverage, or it may be false but have many sources repeating it. The reputation of a source, the look of a site, or the number of mentions do not guarantee truth. Lateral reading and the SIFT method are tools to reconstruct the context that the web often removes, not to guarantee absolute truth. In most cases, these techniques allow for more productive engagement with digital information, but they are not infallible.