The Ninety-Second Check
Almost everything you were taught about judging a website is wrong. Look for a professional design, check the domain, read the About page, see whether it cites sources—every one of those is a test the people you should worry about pass easily, because every one of them is a test they can see coming.
We know this because somebody watched. Stanford researchers sat 45 experienced internet users in front of live websites and recorded their screens while they thought aloud: ten history PhDs, ten fact-checkers employed at well-regarded news and fact-checking organisations, and 25 Stanford undergraduates, all of whom reported spending at least four hours a day online. The historians and the students fell for official-looking logos and domain names. The fact-checkers did not, and they finished first.
The Task That Separated Them
Participants compared articles about bullying on two sites: the American Academy of Pediatrics, the field’s largest professional body, and the American College of Pediatricians, a much smaller group founded in 2002. The names are nearly the same. Both sites look institutional.
Every fact-checker judged the Academy the more reliable source. The historians frequently equivocated, deciding both looked fine. And 60% of the Stanford students picked the College—by the end of ten minutes, only three of the 25 had worked out that the two organisations differed at all. The students had spent those ten minutes on the page, reading it.
One fact-checker spent eight seconds on the College’s landing page. He typed the organisation’s name into a search engine, opened its Wikipedia entry, read that it was founded as a protest against the Academy’s position on adoption by same-sex couples, and followed Wikipedia’s citations out to a Boston Globe story and a Southern Poverty Law Center report. Then he had his answer. The people who read the page carefully did not get one.
The Three Moves
The researchers named what the fact-checkers were doing.
- Take bearings. Before you read anything unfamiliar, work out where you have landed. Not what the page argues—who is arguing it. This is a navigation habit: the hiker who checks the compass before walking, rather than after an hour.
- Read laterally. Leave the page. Open new tabs and find out what the rest of the internet says about the site you were just on. Reading vertically—staying put and studying the page’s own account of itself—means judging a source using only the evidence that source chose to show you.
- Practise click restraint. Do not click the first result. Read the whole results page first: the URLs, the snippets, who is publishing what. The fact-checkers were sometimes slower to their first click and much faster to a sound answer, because a first click can put you on a path you never get off.
That is the entire technique. It is less work than what most people do, which is why it is worth adopting even when you are tired and the stakes are low.
The Version to Actually Use
Mike Caulfield turned the same research into four moves he calls SIFT: stop, investigate the source, find better coverage, and trace claims, quotes and media back to the original context. Stop comes first for a reason—the strongest signal that you are about to be fooled is the feeling that you cannot wait to send this to someone.
The Digital Inquiry Group, which ran the original fact-checker study, boils it down to three questions: who is behind the information, what is the evidence, and what do other sources say.
In practice, on a page you have never seen before:
- Search the site or organisation’s name, plus a word like “funding” or “who owns”. Do not use the site’s own search box.
- Read the results page before clicking anything.
- Open the Wikipedia entry if there is one, and go straight to its references. Use it as a directory of sources, not as the answer.
- Search the specific claim, in quotation marks, and see who else is carrying it and who is not.
- If the claim traces back to one anonymous account and nowhere else, you are done.
What This Does Not Do
Lateral reading tells you who is talking and whether they are generally reliable. It does not tell you whether this particular sentence is true. For that, somebody has to go and look at the documents, and that is a different job, done by people who do it full time—which is the next piece.
It also does not work on a claim that has never been examined by anyone. New, fast-moving and local stories are exactly where the technique thins out, because there is no lateral evidence yet. When you find nothing, the honest conclusion is that you do not know, and the useful move is to wait.