What is lateral reading, and why do fact-checkers use it?

Summary: Lateral reading means leaving an unfamiliar page to find out who published it before you read what it says. Professional fact-checkers leave the page by reflex, most readers stay on it, and a 2021 Stanford study found that 4 hours of practice moved a class of 87 college students from 3 lateral readers to 67.

A page can't vouch for itself

Every signal on an unfamiliar website was placed there by the people who own it. The About page, the mission statement, the footnotes, the .org at the end of the address, and the clean layout each cost the publisher close to nothing and say only what the publisher wants said. Reading those signals closely feels like diligence. It's the same move as asking a stranger whether they're honest and writing down the answer as evidence.

Lateral reading is the alternative. You open a new tab, search the publisher's name, and read what newspapers, encyclopedias, and watchdog groups say about it before you return to the page. The question shifts from "does this look credible?" to "who is behind this, and what do people with no stake in it say about them?" The name comes from the direction your attention moves: A vertical reader scrolls down one page, and a lateral reader moves across several.

Researchers at Stanford borrowed the method from professional fact-checkers, who treat leaving a site as the first step of judging it rather than the last. It has since become the core habit taught under the name news literacy.

Why careful readers stay on the page

The habit of staying put is older than the web. On paper, a publication reached you through gatekeepers: An editor, a printer, a library, and a newsstand. Each had something to lose by passing along junk, so by the time a page was in your hands the vetting had happened upstream, and you could spend your attention on the text. The checklist most of us learned in school (an author, a date, citations, and a reputable-looking domain) was built for that world, where faking those marks was expensive.

Online, every mark on the checklist is free. A site can give itself an editorial board and a .org address in an afternoon. So the careful reader who checks the boxes is applying a paper-era test to a medium that passes it automatically, and the more closely they read the page, the more convinced they get.

A 2021 study in the Harvard Kennedy School Misinformation Review shows what the checklist habit looks like in numbers. Joel Breakstone, Sam Wineburg, and colleagues at Stanford tested 87 undergraduates on 13 source-evaluation tasks before and after a 4-module online course, with each module taking an hour. At the start, 3 of the 87 read laterally. Another 23 leaned on a .org domain as a sign of trust, 14 on how the site looked, and 14 on its About page, and the group averaged 3.95 of 13. After the course, 67 of 87 read laterally, reliance on the .org domain fell 69% (from 23 students to 7), reliance on the About page fell from 14 students to 2, and the average rose to 7.08 of 13, a gain of 3.13 points. The study had no control group, so part of that gain may be familiarity with the test, and the authors say as much. The shift in method, from 3 lateral readers to 67, is harder to explain away than the score gain.

Who vouches for the vouchers?

The strongest case against lateral reading is that it moves the trust problem one step over instead of solving it. When you leave a page to see what an encyclopedia or a newspaper says about its publisher, you're now trusting the encyclopedia and the newspaper, and both get things wrong. There's no floor to that regress. A second objection is practical: Most news reading happens on a phone, where juggling tabs is a chore, and nobody reads laterally for 40 articles a day.

Both objections are partly right. You are trusting a second source, and you'll never reach certainty about a publisher any more than about a claim. What lateral reading buys is independence rather than certainty: A publisher that 3 unrelated sources describe the same way is a better bet than one that describes itself, and when the 3 disagree, the disagreement is itself the finding. As for time, the fix is to read laterally per publisher, not per article. You check a source once, write down the verdict, and from then on the cost is zero. Most people read the same handful of publishers week after week, so the list fills within weeks.

The content of a page gives less help than it seems to. In 2018, Pew Research Center asked 5,035 US adults to sort 10 statements into fact and opinion. Among adults with high political awareness, 36% classified all 5 factual statements correctly, compared with 17% of those with low awareness; Among the very digitally savvy, 44% got all 5 opinion statements right, compared with 21% of everyone else. Educated, fluent readers often can't judge a sentence from the inside. Judging the source is the lever within reach, and it works whether the page is misinformation from a sincere publisher or disinformation from a deliberate one.

How to read laterally in 60 seconds

The method works on a newspaper, a search result, a forwarded link, or a post.

1. Find the publisher, not the author. The author's name tells you little; The organization paying for the page tells you a lot. On a printed paper that's the masthead; On a site it's the name in the address bar or the footer.

2. Open a new tab and search the publisher's name with a word like "owner," "funding," or "bias." Skip every result the publisher controls, including its own social accounts.

3. Read 2 results from sources you already know. An encyclopedia entry and a newspaper's description are usually enough; A watchdog or fact-checking group's write-up is a bonus.

4. Give the publisher 1 of 3 labels: Known and reliable, known and slanted (read with the slant in mind), or unknown (read every claim as a claim). Write the label down.

5. Go back to the page and read it with the label in front of you.

In 2025, the Reuters Institute's Digital News Report found that 58% of people across 48 markets worry about telling real from fake online, and 73% in the US do. When unsure about a piece of news, 38% said they'd turn first to a trusted news outlet, 35% to an official source, 33% to a search engine, and 25% to a fact-checker. Most of those checks are lateral moves already. The adjustment is to point the search at the publisher rather than the claim: Checking a claim is a chase you run once per headline, while checking a publisher is a test you run once and reuse.

That test runs on every item before it reaches the brief at informed.now. Each of the 6 news summaries we send on a weekday morning traces to a primary source we've already read laterally, and the list of sources we cite is public, so the same 60-second check works on us. The method is the point, and the brief is one place it's applied.

The page you're on is the one source that can't tell you who made it. The rest of the web can, and that answer is one tab away.