How to Check Something Online Before You Share It
The SIFT Method

Most of what reaches you online arrives without its context: a screenshot, a quote on a picture, a link from someone you half know. You do not need a research degree to check it. You need a short routine that you will actually use.

A well-known routine is SIFT. It comes from Mike Caulfield — later co-author, with the Stanford researcher Sam Wineburg, of the 2023 book Verified — who set it out in a blog post in June 2019, building on “Four Moves” he had been teaching students since 2017. In his words:

“The four moves: Stop, Investigate the source, find better coverage, trace the original context.”

Move 1 — Stop

Before you read or share, ask whether you know the website or source of the information, and the reputation of both the claim and the website. In Caulfield’s words: “If you don’t have that information, use the other moves”.

Caulfield separates two kinds of checking. When you are doing deep research of your own, you may want to chase down individual claims and verify them independently. But, as he puts it, “Quick and shallow investigations will form most of what we do on the web.”

Move 2 — Investigate the source

Spend about a minute finding out who is behind what you are reading. Caulfield puts it this way:

“Taking sixty seconds to figure out where media is from before reading will help you decide if it is worth your time”

The trick is not to judge a website by looking at the website. Official-looking logos and domain names are what the Stanford researchers below call “easily manipulated features”. Instead, open a new tab, search for the organisation or author’s name, and see what other sources say about them.

Researchers call this lateral reading: “leaving a site after a quick scan and opening up new browser tabs in order to judge the credibility of the original site.”

Why leaving the page works

The evidence comes from a study by Sam Wineburg and Sarah McGrew of Stanford University, published in Teachers College Record in 2019. They watched three groups evaluate real websites while thinking aloud: 10 PhD historians, 10 professional fact checkers and 25 Stanford undergraduates.

The groups behaved very differently.

• Historians and students tended to read vertically — in the study’s words, “staying within a website”. The researchers found they “often fell victim to easily manipulated features of websites, such as official-looking logos and domain names.”
• Fact checkers read laterally. Faced with an unfamiliar organisation’s site, they left its landing page in about half a minute on average and went to find out who ran it.

The authors summed it up in one line: “Fact checkers, in short, learned most about a site by leaving it.”

Two details are worth keeping, because they stop the story being told too neatly:

• Fact checkers were not faster at everything. On an open search task, the students found the answer first. The authors note the fact checkers “were the most selective when it came to the sites they visited and took the most time to verify their answers.”
• Not every historian read vertically. Two of the ten historians behaved more like the fact checkers than like their colleagues.

Move 3 — Find better coverage

If the claim matters more than the source — a statistic, a news event, a scientific finding — then the original post is often the worst place to learn about it. Caulfield’s advice:

“your best strategy may be to ignore the source that reached you, and look for trusted reporting or analysis on the claim”

Look for whether well-established news organisations, reference works, official bodies or specialist publications have covered the same claim, and what they say.

This does not mean you must accept whatever the majority says. Caulfield answers that question directly: “Do you have to agree with the consensus once you find it? Absolutely not!” The point is to know what the consensus is before you decide to disagree with it.

The trap: searching the headline itself

“Just search it” sounds like good advice, and it can backfire.

In a set of five experiments published in Nature (online in December 2023, in the January 2024 issue), Aslett and colleagues asked people to use online search to check news articles, including false ones. Searching increased the chance that people rated a false or misleading article as true. In their first study, the authors describe this as “a 19% increase in the probability that a respondent rated a false or misleading article as true.”

What people typed mattered. When the search was the false article’s own headline or web address, the results were far more likely to include unreliable sources: 77% of those searches returned at least one unreliable news link, against 21% of other searches.

So the SIFT move is not “search the claim”. It is find coverage you have reason to trust — search for the topic in your own words, go to sources you already know, and look at who is reporting it.

Move 4 — Trace claims, quotes and media to the original context

A great deal of what spreads online has been cut away from where it came from. As Caulfield puts it, “Much of what we find on the internet has been stripped of context.”

Tracing means going upstream until you reach the original:

• A quote — find where it was first said or written: the book, the speech, the interview. Our quote checks show how often a famous line turns out to belong to someone else.
• A photo — search for earlier copies of the image with a reverse image search tool, and see where and when it first appeared.
• A statistic or study — find the report or paper itself, and check the claim matches what it actually found, not what a headline said it found.

What the evidence supports — and what it does not

What it supports. A study of college students in a civics course found that, after instruction in lateral reading, they “were more likely to use lateral reading to fact-check and correctly evaluate the trustworthiness of information than controls.”

What it does not yet show.

• The foundational study was small. Ten historians and ten fact checkers, all in the United States. The authors themselves note that small samples exaggerate differences, and that more research is needed to know whether the strategies generalise across topics, sites and searches.
• The college study is limited. It came from one university and was not a randomised trial, and one of its co-authors is Caulfield himself. It says further research is needed to know whether the improvement is maintained over time.
• Whether it lasts has not been established. The studies that looked are few, and they tested quite different things.
  • A 2023 study of Canadian school students taught with SIFT-based lessons found that “Gains were maintained from posttest to delayed posttest” — a follow-up about a month later, but with no comparison group at that stage.
  • A 2026 randomised controlled trial with 2,666 adults in Germany, published in Psychological Science, tested a short lateral-reading video and a claim-search strategy. Both showed small improvements, not all of them statistically distinguishable from zero. At a two-week follow-up, discernment had improved in every group, including the control group, and “differences between the intervention and control groups were no longer statistically distinguishable.” In a report on the study the researchers note that the trained groups’ scores did not decay; the control group caught up.
  • A 2025 study of 459 upper-secondary students in Swedish classrooms found that a fact-checking workshop did not significantly improve their ability to evaluate credible and misleading news, and students “did not report increased use of digital verification tools three months after the intervention”.
  • These do not cancel each other out, because they tested different doses on different people: a course of classroom lessons, a five-minute video, a single workshop. What none of them shows is a lasting benefit measured against a comparison group.
• The sixty seconds is a rule of thumb, not a finding. It is Caulfield’s guidance for investigating a source, not a measured result.

That is a fair place to leave it: a sensible routine, backed by promising but modest evidence, that costs you a minute and can save you from repeating something false.

The routine, on one screen

1. Stop. Do I know this source? If not, don’t share yet.
2. Investigate. Open a new tab. Who runs this site or account, according to other sources?
3. Find better coverage. What do sources I already trust say about this claim?
4. Trace. Where did the quote, photo or number first appear, and does it say what the post claims?
5. Don’t search the headline word for word. Describe the claim in your own words instead.

Sources

• Mike Caulfield, “SIFT (The Four Moves)”, Hapgood, 19 June 2019.
• University of Chicago Press, catalogue page for Mike Caulfield and Sam Wineburg, Verified (2023).
• Sam Wineburg and Sarah McGrew, “Lateral Reading and the Nature of Expertise: Reading Less and Learning More When Evaluating Digital Information”, Teachers College Record 121(11), 2019 (authors’ accepted manuscript, Stanford Digital Repository).
• Brodsky, Brooks, Pavlounis and Johnston, “Instruction Increases Canadian Students’ Preference for and Use of Lateral Reading Strategies to Fact-Check Online Information”, AERA Open 9, 2023.
• J. E. Brodsky and colleagues, “Improving college students’ fact-checking strategies through lateral reading instruction in a general education civics course”, Cognitive Research: Principles and Implications 6, article 23, 2021.
• L. Oswald, A. Kozyreva, S. M. Herzog, P. L. Nickl and R. Hertwig, “Boosting Media Literacy Using Lateral Reading and Online Search Interventions”, Psychological Science (published online July 2026).
• T. Nygren and E. Efimova, “Investigating the long-term impact of misinformation interventions in upper secondary education”, PLOS ONE (2025).
• K. Aslett and colleagues, “Online searches to evaluate misinformation can increase its perceived veracity”, Nature 625 (published online December 2023).

Checked September 2026.

Related: Did They Really Say That? 22 famous quotes checked · Why critical thinking matters more than ever with AI · Overcoming survivorship bias