A political statistic moves faster than any correction. You see a number on your timeline, it fits something you already suspected, and four seconds later it is on your timeline with your name underneath it. That is not a failure of intelligence; it is how attention works. But sharing is a small act of authorship. Once your name is attached, the claim borrows your credibility whether you meant it to or not.

The ten questions below take about two minutes. They are not a licence you have to earn before you are allowed an opinion. They are the checks a decent sub-editor would make before letting a number into print, and most of them you can run on a phone while the kettle boils.

Where did the number actually come from?

Start with the paperwork, not the post.

Can I find the original source — not a screenshot of it?

A cropped image of a table tells you nothing about who compiled it, when, or why. Search a distinctive phrase from the graphic in quotation marks and see what comes up. If every version you can find is somebody else's screenshot, you have found a rumour with good typography. Treat it as unverified and move on.

Who produced it, and what do they want?

A national statistics agency, a company press release, a campaign group and an anonymous account are four different kinds of object. All four can be right. Only some of them are checked by someone whose job is to check. Ask whether the organisation exists to measure things or to win arguments. The answer does not disqualify the figure — it tells you how much weight to put on it and how to describe it when you pass it on.

Is the source making the claim, or is the sharer making it for them?

This is where most distortion happens. A report might say a trend "suggests" or "is consistent with" something; the post above it says the report "proves" it. Compare the wording of the claim with the wording of the source. The gap between the two is where the spin lives, and it is usually visible in under a minute.

Check the date, and the units

When was it published?

Old figures get recycled endlessly. A number about NHS waiting times, wages or recorded crime from five years ago can still be circulating as though it were this morning's news. Look for a publication date on the original page — not the date of the post you happened to see.

Is it a rate, a total, or a projection?

"One in five" and "20 per cent" are the same thing. "20 per cent more than last year" is not "20 per cent of". Watch the difference between per cent and percentage points: a move from 2 per cent to 3 per cent is one percentage point, and it is also a 50 per cent rise. Both statements are true, and one is chosen to sound enormous. Forecasts need the same care — a projection for 2035 is not a measurement of 2024, however confidently it is drawn.

Look at the framing, not just the figure

The number can be accurate and the message can still be misleading. Three things to check:

• The comparison. "Up 40 per cent" compared with what, and why that baseline? A rise from a very small number is still a very small number.
• The chart axis. A bar chart of values between 96 and 98, with the axis starting at 95, turns a rounding error into a cliff edge. Sometimes that is legitimate; often it is decoration.
• The verb. "Linked to", "associated with" and "caused by" are not interchangeable. Two things rising together is a fact about the numbers. One causing the other is a claim that needs evidence of a different kind.

If the graphic falls apart once you read the small print, that is your answer. You do not need to find the rebuttal; you just need to not be the one who spread it.

Test the claim against yourself

Would I share this if it argued the opposite?

This is the hardest question and the most useful. Most of us apply a stricter standard to numbers we dislike. Try the swap: same source, same method, opposite conclusion. If that version would have you reaching for the "misleading" reply, slow down.

Could I explain it in my own words, without the post?

Say out loud what the figure measures, over what period, for whom, and compared with what. If you cannot, you do not yet know what you would be sharing. Paraphrasing is the test, because it forces the vague bits into the open.

If you still can't verify it, say so

You will not always be able to settle a claim, and you do not have to. "I can't find the source for this" is a perfectly good contribution. So is waiting a day. If something is important, someone with more time and better tools will have checked it by tomorrow, and the post will still be there.

What you can do is refuse to launder the uncertainty. Sharing with "if true, this is huge" does not protect you: the headline travels, the caveat does not.

The two-minute habit

Nobody runs ten checks on every post. Pick two that catch most of the nonsense — find the original source, and check the date — and make them automatic. Add the "would I share the opposite?" test when a claim flatters you, because that is exactly when your judgement is weakest.

The aim is not to become the person who fact-checks the group chat. It is to keep your own name attached to things that are true, and to make your timeline slightly less useful to people who rely on you not checking. Two minutes is a cheap price for that.

Photo: DayronV / Pixabay