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

The June issue of the Green Gazette had a headline in the biggest type the printer would allow, and Tara had been pleased with it, because it felt like the paper doing what a paper is for.

THE GREEN SAYS: CLOSE THE SKATE PARK AT 8

It had started with a letter in May. The skate park at the bottom of the Green, a concrete bowl and a few rails behind the bowling club, stayed open until ten every night in summer, and people who lived in the houses on Mill Row said the noise kept them awake: the clatter of boards, the shouting, the music from somebody's speaker. The council was going to decide in July whether to change the hours.

Nadia in Year 9 had written the story. It was her first front page. She had had a good idea, too. Instead of just quoting the letter, she had asked the Gazette's readers. The paper had a website, which mostly had the crossword on it and a list of lost cats, and Nadia had put a poll on it for two weeks.

Should the skate park close at 8pm in summer instead of 10pm? YES / NO.

A hundred and forty people had voted. A hundred and fifteen had said yes. That was eighty-two per cent.

The front page had a big pie chart, mostly red, and a photograph of the empty bowl at dusk. The council officer had already quoted it in a newsletter: A recent poll found that over eighty per cent of residents support earlier closing.

On Tuesday there was a letter in the Gazette box. It was handwritten, on lined paper torn from a notebook.

Dear Gazette, it said. I'm the youth worker at the skate park. I'm there four nights a week. I have about forty regulars, mostly fourteen and fifteen, and not one of them had heard of your poll until they saw the front page. Neither had I. I would like to know who you asked, because it wasn't us. From Bex (and the kids).

Tara pinned it to the board above the back table at the Whetstone café. She read it twice.

She had been proud of the front page. Now she was uneasy. She did not know why yet.

"I was going to say the poll was open to everyone," Tara said.

She had brought the letter to the back table. Danya was there, with a skateboard of his own leaning against his chair, which he had been given for his birthday and could not yet stand on for more than three seconds. Across from him sat Corah, who wrote for the school paper and read the fine print of everything, and who got quietly angry about anything that seemed unfair.

"I was going to say it like this," said Tara. "Nobody was stopped from voting. It was on the website for two weeks. Anyone could click. A hundred and forty people did. That's more than came to the council meeting. I can write back to Bex and say the poll was open, and the skaters could have voted too."

She had the sentence half written. Thank you for your letter. The poll was open to all readers.

It felt good. It felt like defending something fair: a vote, open to anyone.

Corah read the front page, and then the letter, and then the front page again, more slowly.

"Who did they ask?" said Corah.

"Everyone," said Tara. "That's the point. Nobody was chosen. People chose themselves."

"Who got left out?" said Corah. "That's my question. Who did they test it on?"

"Nobody got left out," said Tara. "Anyone could vote."

"Anyone could vote," said Corah. "Who knew there was a vote? That sample doesn't look like us. A hundred and forty people. Forty skaters, and none of them knew. So who were the hundred and forty?"

"Says who they were different?" said Danya. "Maybe they were just people."

"Nobody says it," said Corah. "That's what bothers me. We printed THE GREEN SAYS, and nobody's checked who was speaking."

Tara looked at the letter. Not one of them had heard of your poll.

"How do we know who voted?" Danya asked.

Nobody knew. That was the problem.

Deed found it that evening.

Deed read council minutes for fun and kept documents on her phone, and she had a way of finding the page nobody else had opened. This time it was not a council report. It was the Gazette's own website. The website kept a record, behind the scenes, of where visitors had come from: whether they had typed the address in, or come from a search, or clicked a link somebody had sent them.

"Go to the original document," said Deed. "Not the pie chart. This. Where did every vote come from?"

She turned the phone round. It was a short table.

Most of the votes had come from one place. A link, posted on the second day of the poll, in a group chat called Green Neighbours.

"Ninety-six of the hundred and forty," said Deed. "Through that one link."

"What's Green Neighbours?" said Tara.

"It's a group for people on Mill Row and the streets round it," said Deed. "It started last year. I looked at the minutes of the residents' meeting. It was set up to organise complaints about the skate park. That's what it's for. Somebody posted our poll in it on day two, with a message saying Vote YES!!"

Tara looked at the pie chart. Mostly red.

"So most of the people who voted," she said slowly, "were already in a group that wanted it closed."

"Ninety-six of them," said Deed. "Read it where it was first written down. It's all there."

She felt something shift, the way it had in May with the crossings, and in the spring with the bikes. It was the feeling of a true number turning out to be standing on the wrong ground.

Prei built spam filters for fun and liked counting things. Prei took the phone, and a pad, and drew two piles.

"Count the flags, then count the right ones," said Prei. "Or here: count the votes, then count where they came from. Let's put it on paper with real numbers."

In the first pile, Prei wrote from the Green Neighbours link: 96 votes. Then Prei looked at the table again. Each vote was marked yes or no.

"Of those ninety-six," said Prei, "ninety-two said yes. Four said no."

In the second pile, Prei wrote everyone else: 44 votes. "And of these forty-four," said Prei, "twenty-three said yes, and twenty-one said no."

Tara added it up in her head. Ninety-two and twenty-three. A hundred and fifteen. The same as the front page. Every vote was real. None of them had been changed.

"So among the people who didn't come from that group," said Tara, "it's twenty-three out of forty-four."

"About fifty-two per cent," said Prei. "About half and half. A flag is not a fact. The eighty-two per cent was mostly one group's view, added to a coin toss."

"Hm," said Tara.

"The votes are real," said Prei. "Every one. That's not the problem. The problem is the headline. It says The Green. It's really Green Neighbours, plus forty-four other people."

Sana had spent last summer as an intern at a lab, and still checked every sample sheet twice. Sana looked at Prei's two piles for a long time without saying anything.

"Does the sample look like the whole?" said Sana at last. "That's what I want to know. Who's in the whole?"

"Everyone who lives on the Green," said Tara.

"Show me the whole sheet, not the tidy one," said Sana. "Who on the Green never saw the poll?"

Tara thought about it.

"People who don't read the Gazette's website," she said. "Which is most people, to be honest. It's mostly the crossword."

"Who else?"

"People who aren't in that group."

"Who else?" said Sana.

Tara looked at the letter on the board. About forty regulars, mostly fourteen and fifteen.

"The skaters," she said. "Nobody sent it to them. They're not in the Green Neighbours group. Most of them don't read the paper. Some of them are too young to have the right kind of phone." She stopped. "They're the people the decision is about, and almost none of them knew it was happening."

"The ones you threw away are data too," said Sana. "Even if you didn't mean to throw them away. A poll anyone can answer isn't a poll of everyone. It's a poll of the people who heard about it and cared enough to click. That's a real thing to know. It's just not the same as what the Green thinks."

"So a poll is useless," said Danya.

"No," said Sana. "A poll is a measure of who answered it. You have to say who that was."

Attest came last, with the Gazette's red pen. Attest subbed every story before it went to print and drew a red circle round any number without a name next to it. The June front page had come back with one circle, round the words THE GREEN SAYS. Next to it Attest had written: which green?

"Who says so? I wrote that," said Attest. "And then I thought I was being fussy, because the poll was open to all."

"I saw it," said Tara. "I thought you meant the headline was too long."

"I meant: who was speaking," said Attest. "But I let it go. Nobody followed it up, until Corah asked who got left out."

"Let me listen first," Tara said, before anyone started telling her what to write.

She took her pad and wrote three headings across the top: what we printed, where the votes came from, who never had the chance.

Under the first she wrote the Green says close at 8: 82%. Under the second she wrote 96 from one group, 92 of them yes; the other 44, about half. The third column filled quickly: people who don't read the website, then people not in the group, then the skaters.

"Hm. So what does a clarification like this have to carry?" Tara asked.

"Every claim names its source," said Attest. "Our own poll, and how we ran it: online, two weeks, anyone could vote. Where most of the votes came from, named: one group, set up about the skate park. That the votes are real but don't speak for the whole Green. And something about next time."

"Next time?"

"A correction names its source too," said Attest. "If we run another poll, we say who answered it, in the same size type as the answer. Put a name next to it."

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

"Four sentences," said Attest. "One of them is a promise, which is the hard one."

Tara looked at the council officer's newsletter, which somebody had clipped and pinned beside Bex's letter. Over eighty per cent of residents support earlier closing.

"I think we missed someone," Tara said slowly. "The council. They've already used it. If our box only says sorry, they'll still have the eighty-two per cent in their newsletter. We have to say it clearly enough that they can't."

Nobody answered at once.

"Who you say it to is your call," said Attest. "That's your bit."

Mrs Hallam, who advised the Gazette and signed off every issue, came in at four. Nadia came with her, very pale. Mr Okafor from Mill Row came at quarter past, in his work clothes. Kai came last, with Bex, the youth worker. Kai was fifteen and carried his board under his arm and did not sit down.

Nadia spoke first.

"The poll was open to anyone," she said. "I didn't post it in that group. I didn't even know it existed. I put it on our website and told people in the paper. Every vote is real. Nobody voted twice, I checked. If people who wanted it open didn't vote, that isn't my fault, is it?"

It was the strongest version of her argument, and every part of it was true.

Mr Okafor spoke next.

"I live at number four Mill Row," he said. "My bedroom window is forty metres from that bowl. Until ten o'clock every night from May to September there is a clatter, and a crash, and somebody's music. I work early. I have not slept properly for a year. I joined that group because I wanted somebody to listen, and nobody did, until your paper." He looked at the pie chart. "I'm sorry if our group voted a lot. We care a lot. That's not the same as cheating."

That was true too.

Then Kai put his board down on the floor, wheels up.

"It's the only place," he said. "On the whole Green. For people my age. There's the library, but you can't talk. There's the café, but you have to buy something. The skate park's free and nobody tells you to go home. If it shuts at eight in summer, it's light till ten, and we'll just be on the street." He shrugged. "Nobody asked us. That's all. Nobody asked."

Nobody said anything for a while.

"Who hasn't spoken yet?" said Danya, very quietly. "That's what you always ask."

"Nobody asked the bowling club either," said Bex. "Or the people in the flats on the other side, who never complain. Or the parents who'd rather their kids were in the bowl than on the street."

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

On the table in front of Tara was the front page, with its mostly red pie chart and its empty bowl at dusk. There was Deed's phone, open at the table of where the votes had come from. There was Prei's pad with its two piles: 96, 92 yes and 44, about half. There was Sana's list of the people who never saw it, and Attest's red circle: which green? There was the council's newsletter, with over eighty per cent. And there was one letter, on lined paper torn from a notebook, signed Bex (and the kids).

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

She handed Tara the pen.

Tara looked at the empty box.

 

 

(Write the clarification Tara prints under the June front page.)

 

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 weekend Lan, who was seven, asked everyone in her class whether the class pet should be a hamster or a fish.

She asked at break, by the climbing frame. Fourteen children said hamster. Two said fish.

Danya asked her who had been at the climbing frame.

Lan thought. She said it was mostly the children who liked running about. The children who liked quiet things had been in the library corner, reading.

She went to the library corner on Monday and asked them too. It was much closer after that.

Afterword. The idea in this case is self-selection: when people decide for themselves whether to answer a poll, the answers show who heard about it and cared enough to reply, which is often the people with the strongest feelings, not what everyone thinks. Every vote in this case was real, and none was cheated. But most came from one group that had been formed around the question, and the people most affected were never told. A famous example is a magazine poll in 1936 that got more than two million replies and still predicted the wrong American president. Before you trust an open poll, ask who saw it, who answered, and who never had the chance; and when you run one, say who answered it. You can practise checking who is in a sample in InquiryForge, and reading where a number came from in NewsForge.