Suppose you land on a website with a clean design, a dot-org address, an About page full of high-minded language, and footnotes at the bottom of every article. Is it trustworthy? Most people, including very educated ones, try to answer by studying the site itself. Professional fact-checkers do something different, and the difference turns out to matter a great deal.

Historians, students and fact-checkers

Researchers at Stanford University led by Sam Wineburg and Sarah McGrew ran a revealing test, first reported in 2017. They asked three groups to evaluate websites and social media posts: 10 PhD historians, 10 professional fact-checkers and 25 Stanford undergraduates. All were skilled readers. The historians were experts at weighing documents, and the students attended one of the most selective universities in the world.

One task compared two organizations with almost identical names: the American Academy of Pediatrics, the main professional body for children's doctors in the United States, and the American College of Pediatricians, a much smaller group founded in 2002 with a distinct social and political agenda. After reading the pages closely, many historians and students judged the smaller group's site to be as reliable as the larger one's, or more so. The fact-checkers quickly worked out who each group was.

The reason was method. The historians and students read vertically: they stayed on the page, scrolling up and down, judging its look, its language and its links. The fact-checkers read laterally: they left the site quickly to see what others said about it, opening new tabs and searching for the organization's name. They reached better conclusions in a fraction of the time.

Why staying on the page fails

Vertical reading fails because everything on a page is controlled by whoever made it. A professional design, an official-sounding name, citations and a dot-org address can all be bought or copied cheaply. Asking a site whether it is reliable is like asking a stranger whether they are honest. Lateral reading asks the neighbors instead.

Fact-checkers also practiced what the researchers called click restraint: scanning the whole results page before choosing a link, reading snippets and web addresses rather than clicking the first result. The first result is often the one best at search optimization, not the most informative.

Wikipedia, perhaps surprisingly, was one of their most useful tools. Its articles on organizations often summarize funding, history and criticism, with references to follow. Fact-checkers used it not as a final authority but as a quick first stop to learn who is behind a source.

SIFT: four moves

The digital literacy educator Mike Caulfield turned these habits into a short routine called SIFT, published in 2019:

• Stop. Before reading or sharing, pause and ask what you know about the source and what you want from it. A surge of strong emotion is itself a cue to stop.
• Investigate the source. Spend a minute learning who is behind it, by reading laterally.
• Find better coverage. If the claim matters, look for whether trusted outlets have reported it. Often the fastest way to check a claim is to set aside the source you found and look for a better one.
• Trace claims to their original context. Follow quotations, statistics and images back to where they started, and see whether they say what the post claims.

SIFT does not ask anyone to become an expert. It asks for a minute or two of checking before belief and sharing, scaled to how much the claim matters. Studies of short lateral reading lessons in schools and colleges have found that even brief instruction can improve students' judgments, though the gains fade without practice.

Limits and fairness

Lateral reading has critics. Learning who funds a site tells you about its likely slant, not whether a particular claim is true, and dismissing an argument only because of who makes it is a classic fallacy. The fair use of these tools is triage: deciding how much trust a source has earned and how carefully to check it, while staying willing to weigh a good argument from an unexpected place. Lateral reading also depends on other sources being reliable, which is one reason a healthy press and well-run reference works are public goods.

Worked example

A post shares a striking chart: "Crime up 300 percent since the new policy," from a site called the Center for Public Safety Research. Apply SIFT. Stop: the claim is dramatic and aimed at a heated political issue, so slow down. Investigate: a search for the center's name in a new tab, not on its own site, shows it was founded last year by a political consultant and lists no researchers. Find better coverage: searching for the city's crime statistics leads to police data and a local newspaper report. Trace: the 300 percent figure turns out to be the rise in one category, bicycle thefts, from 4 to 16 in a single neighborhood. The chart rested on real numbers, but the post's claim about crime was false. The whole check took about four minutes, and none of it required judging the site's design.