The hammer rang once. A spark jumped off the anvil and did not go out. It waited. They followed.

The January issue of the Green Gazette had the biggest headline Tara had ever printed. It ran right across the top of the front page in thick black letters:

ENERGY DRINKS DOUBLE HEART RISK, STUDY FINDS

Ollie had written it. Ollie was in Year 10. It was his first front page, and he had been so nervous about it that he had rewritten the opening paragraph eleven times. He had found the story himself, in a press release from a university health department, and he had read it three times and checked every word against it.

The story said that a large study of young people had found that drinking two energy drinks a day doubled the risk of a heart-rhythm problem serious enough to need a hospital. It quoted the press release. It had a photograph of a fridge full of bright cans.

By Wednesday the canteen had stopped selling them.

By Friday there was a thank-you card in the Gazette box from a parent. My son drinks three of these a day before football, it said. He read your front page and poured one down the sink. I have been trying to get him to do that for a year.

Tara pinned the card to the board above the back table at the Whetstone café, where the Gazette did its work on Saturdays. It was the first time anybody had sent them a card.

Then she found the other letter.

It was underneath the card, on a single sheet of paper folded once. It was from Joel Mensah, in Year 13. Tara knew him a little. He ran the school chess club and was known for saying very little and being right when he did.

The letter said two words.

Twice what?

Tara turned it over. There was nothing on the back.

She felt the pride of the week tip, just slightly, into something more defensive.

"I was going to say it's accurate," Tara said.

She had brought both letters to the back table. Danya was there, doing homework he had promised his mum he would finish before lunch. Across the table sat Widen, who was in the year above Tara and wrote the Gazette's money pages. Widen kept a small notebook of what things cost, from a bus ticket to a bridge, and could not hear a number without wanting to put another number beside it.

"The study found the risk doubled," Tara said. "That's what the press release says. Ollie checked it. The headline says exactly what the study found. We didn't make anything up."

She meant to write back to Joel and say so. She already had the sentence half written in her head: Thank you for your letter. Our headline accurately reports the findings of the study.

It felt good. It felt solid. Being accurate was the whole point of the Gazette.

Widen read Joel's letter and smiled.

"Compared with what?" Widen said. "He's asked the right question. Is twice a lot, or does it just sound like a lot? We can't tell which yet."

"What question? He's only asked two words."

"Twice what?" said Widen. "Doubled compared with what? If something goes from one in two to two in two, that's doubled. If something goes from one in a billion to two in a billion, that's doubled too. Same word. Completely different news."

Tara opened her mouth. She shut it again.

"Compared with what, though, in the story?" Widen asked. "Did it say what it doubled from?"

Tara picked up the paper and read Ollie's story again, all the way through, slowly. Then she put it down.

"No," she said.

"Did the press release?"

"I don't know," said Tara. "I don't think so. Ollie would have put it in."

"One number on its own tells you almost nothing," said Widen. "We don't know whether this is big news or small news. We need to widen the view."

Danya looked up from his worksheet. "How do we know it's even big?" said Danya. "Twice of nothing is still nothing."

"Not nothing," said Widen. "But maybe not much."

Deed came in at half past eleven with her hair still wet from swimming. Deed read council minutes for fun and kept PDFs of planning applications on her phone, in folders. She had once found a mistake in the town budget on page two hundred and nine.

Tara showed her the front page and Joel's letter.

"Where's the study?" Deed asked.

"Ollie used the press release."

"Not the press release," said Deed. "Go to the original document. The press release is somebody's summary of the study, and the story is our summary of the press release. A summary of a summary drifts. Let's go and read it where it was first written down."

It took her nine minutes to find it. It was on a journal website, and the summary at the top was free to read. Danya leaned over to watch her scroll.

"Here it is, the methods and the table," said Deed. "Page four, table two." She turned the phone round.

The study had followed tens of thousands of young people for a year. It had sorted them by how many energy drinks they said they drank. Then it had counted how many of them went to hospital that year with a heart-rhythm problem.

Among the young people who drank no energy drinks, it was about one in every ten thousand.

Among those who drank two a day, it was about two in every ten thousand.

Tara read the two numbers several times.

"Says who, though? Two in ten thousand is twice one in ten thousand," said Danya, who was good at this.

"It is," said Deed. "So the headline is right. It doubled."

"How do we know that's not tiny?" said Danya. "It's one more. One more person in every ten thousand."

Nobody said anything for a moment.

"The press release kept the word twice," said Deed, "and dropped both the numbers. You can see why. Twice is a better headline than one more in ten thousand."

"Who checked that before they wrote the press release?" said Danya. "That's not fair. It sounds like a scam."

"No," said Deed firmly. "Read the next bit."

The next paragraph of the study said that one extra case in ten thousand was small, but that it was real, and that it mattered because so many young people drank these drinks every day. It also said that the drinkers and the non-drinkers might have differed in other ways. Some might sleep less, or have other health problems, and the study could not rule all of that out.

"It's careful," said Deed. "It's a careful study, if you read it where it was first written down. The care just didn't make it into the press release."

Widen had been writing in the notebook while Deed read.

"Is ten thousand a lot, or does it just sound like a lot?" Widen said. "It's a hard number to picture. I want to put it next to something real. How many students at our school?"

"About fifteen hundred," said Tara.

"Put it next to the whole school, then," said Widen. "Imagine every one of them drank two a day, every day, for a year. Two in ten thousand of fifteen hundred is nought point three. Point three of a person a year."

"Says who? You can't have point three of a person," said Danya.

"Widen the view and it means about one person every three or four years," said Widen. "And if none of them drank any at all, it's one in ten thousand of fifteen hundred. Nought point one five. About one person every six or seven years."

Tara wrote it down.

One every six or seven years, with nobody drinking them.

One every three or four years, with everybody drinking two a day.

"Twice as big," said Widen. "And still rare. Both of those are true at once. That's what compared with what is for."

Tara looked up at the board, where the thank-you card was pinned. A boy had poured a can down the sink because of their front page. She did not feel sorry about that. She did not think she should.

"Can I say one thing?" Tara asked. "Was the headline wrong?"

"Not wrong," said Widen. "Half. It's half a number. It gave people the times and not the thing being timesed."

"Who checked that? That's not a word," said Danya.

"It is now," said Widen.

Attest arrived last. Attest subbed the Gazette's copy and drew a red circle round every number that did not have a name next to it. The front page had come back from Attest with only one circle on it, round the word double. Nobody had noticed the circle, because it had not been round a number.

"Where did that number come from? I wondered when someone would write in and ask," Attest said.

Tara felt her face go warm. "You could have said."

"I circled it," said Attest. "Where did that number come from? That's what a circle asks." Then, more kindly, "I wasn't sure either. That's why it was a circle and not a line through."

They sat together and worked out what a correction would have to say, if it was going to be a correction at all.

"Let me listen first," Tara said, before anyone could start. She had learned that from the flooded-underpass photo in the autumn, when she had nearly defended a picture before she knew where it came from. She took out a pad and wrote three headings across the top: what we printed, what the study says, what is missing. Under the first she wrote the headline. Under the second she wrote Deed's two numbers. The third column stayed empty for a long time, and then she wrote the number that the word "double" belongs to.

"It isn't really a correction," said Attest. "Nothing we printed was false. It's a clarification, and a clarification names its source too. Put a name next to it: the study, not the press release."

"I was going to say, does that mean we take the story back?" Tara asked.

"No," said Attest. "The warning can stay. The study found a real increase. We're just putting the number next to the word. The box has to carry both numbers, one in ten thousand and two in ten thousand. It has to say where they came from, the study itself, not the press release. And it should say the thing the study said about itself, that the drinkers and non-drinkers might have been different in other ways."

"Hm. That's a lot for one box," said Tara.

"It's two sentences," said Attest. "Maybe three. Every claim names its source, and a times needs the number it multiplies. That's all."

Tara thought about the parent's card again. She thought about the boy pouring the can away.

"I think we missed someone," Tara said slowly. "What if we print the numbers, and he reads one in ten thousand and thinks it's fine, and starts drinking three a day again?"

Nobody answered at once.

"Who says so is settled, and every claim names its source," said Attest. "What to say is the hard bit. That's your bit."

Mrs Hallam, who advised the Gazette and signed off every issue, came in at noon with Nurse Adeyemi from the school medical room. Ollie came with them. He had heard about Joel's letter and looked as if he had not slept.

Nurse Adeyemi spoke first. She was small and brisk and had worked at the school for twelve years.

"I want to say something before you decide," she said. "That doubling is real. These drinks have a lot of caffeine in them. In this country the cans already have to say they aren't recommended for children. I see students every week who've had three before second lesson and can't sit still and can't sleep. If you print 'one in ten thousand' in big letters, half of them will hear 'safe'. That's what I'm worried about."

It was the strongest version of the argument, and she meant every word of it.

Mrs Hallam spoke next. "The headline matched the press release," she said. "A university wrote that press release. Ollie did what we teach him to do. He read it carefully and quoted it exactly. If we print a box saying the headline was misleading, it'll feel to him like we're blaming him for doing his job."

Ollie looked at the table.

"I checked it," he said quietly. "I really did check it."

"You did," said Tara. "Against the press release. That's what we all would have done."

Then she unfolded Joel's letter again, because he was not there to speak for himself. She knew what he would have said, because she had thought about it all morning.

A frightening word without its number spends trust, she thought. And the next warning, the one that really matters, will need that trust. If people learn that our headlines say double about things that are rare, then one day we'll print something big and nobody will believe us.

She said it out loud, so it would be in the room with the others.

Nurse Adeyemi nodded. "That's fair," she said. "I don't want a scary headline. I want them to stop drinking three before maths."

"Then put a name next to it, and the warning stays," said Attest. "And it gets its number."

Mrs Hallam took the corrections box out of her bag. It was really just a small grey panel at the bottom of page two, with a border round it and space for a few lines. She had used it four times that year. She set it on the table in front of Tara, with the January front page beside it.

On the page in front of Tara were Deed's two numbers, one and two in ten thousand. There was Widen's school-sized version, one every six or seven years against one every three or four. There was the study's own careful line about other differences, and the boy pouring a can down the sink. And there was Joel's letter, with its two words.

"You're the editor," said Mrs Hallam. "What does it say?"

She handed Tara the pen.

Tara looked at the empty box.

 

 

(Write the line Tara prints in the corrections box.)

 

The spark dimmed to a warm dot and settled back on the anvil. The Green was just the Green again.

Back on the Green

That week an advert on Danya's phone said a new game was three times more fun than the old one.

He went to find Lan, who was seven and was building a tower out of every cushion in the café, and asked her, very seriously, three times more fun than what. Lan said than the old one. Danya said exactly, and how fun was the old one. Lan thought about it and said not very. Danya said then three times not very is still not very.

Lan threw a cushion at him, which was a fair answer.

Afterword. The ideas in this case are relative risk and absolute risk. A relative risk compares the chance of something in one group with the chance in another: "twice the risk" is a relative risk of two. An absolute risk is the chance itself, such as one in ten thousand a year. Both are true at once, and each is misleading alone: "twice" without its starting number can make a rare thing sound common, and "one in ten thousand" without the comparison can make a real increase sound like nothing. Good reporting gives both, says where they came from, and says what the study could not show, such as whether something else explained the difference. A clarification adds what was missing without taking a warning back. Questions about your own health belong with a doctor, nurse or pharmacist. You can practise comparing risks and spotting a missing baseline in InferForge, and putting a number next to something real in NewsForge.