Three headlines, one of them built like a fake. Which one — and what gave it away?
Setup: all three headlines below were written for this page. Two are built the way real reporting is built; one is built the way fakes are built. Read them out loud, then everyone votes at the same time — no talking over each other.
A) “Supermarket recalls 3 batches of Golden Crunch peanut butter over a labelling error” — The Straits Times, Tuesday, quoting a statement from the supermarket chain.
B) “Teens who park their phones an hour before bed score 12% better on memory tests” — a university sleep study of 240 students, published last month.
C) “Free MRT rides for all students next Monday to mark the network’s birthday” — officials said, adding that “details would be shared soon”.
Worksheet Q1 — the three checks: for each headline write down (1) WHO says it — a named person, agency or company? (2) WHEN — a real date? (3) THE NUMBER — is there one, and can you look it up? One headline fails all three.
Worksheet Q2 — the gut test: which one would you forward to a friend BEFORE checking? Write down the feeling that made you want to. That feeling is exactly what fakes are built to trigger.
Worksheet Q3 — fix it: rewrite the fake headline so it passes all three checks. Keep it boring. Boring is the price of true.
Headline C is the fake — it fails every check
Here's the reveal, and it has nothing to do with whether free MRT rides would be nice. C fails all three checks at once. Who? “Officials” — no name, no agency, no company. Trains in Singapore are run by named operators and announced by a named authority; a real free-ride story has an organisation standing behind it, because someone has to answer for it. When? “Next Monday” — a date that moves every time you read it, instead of a real one like 27 October. The number? None. And the giveaway phrase: “details would be shared soon”. Nobody announcing a real benefit says that. Real announcements arrive with a date, a website, and a form that nobody can find.
Look at what A and B do instead. A names the outlet, the day, and the count — 3 batches — so you can go and check three batches. B carries a number with a unit (12%), a sample size (240 students) and a place to look (last month's study). You might still find a mistake in either one — real reporting gets things wrong — but both of them left you a handle to grab. C left you a feeling and nothing to hold.
Which explains the Q2 answers. Almost everyone wants to forward C first, and not because it's more likely — because it's more fun: free stuff, students, a deadline. Fakes are engineered for the moment before thinking. Outrage, free money, fear, “share before they delete it” — all of it is accelerant, and the accelerator is the point. Real reporting is deliberately boring, because it has to survive a lawyer and a correction.
So the fix from Q3 is the whole skill: name the who, stamp the when, keep the number. A headline you can check is a claim. A headline you can only feel is advertising.
Today's 1%: before forwarding anything, say the source out loud. If you can't name who said it, you don't have news — you have a rumour with good lighting.