For about fifteen years the standard advice was to look at how a post looked. Bad spelling, a strange web address, an image that had clearly been through five screenshots. That advice was never great and it is now close to useless, because making something look professional costs nothing. The checks below are about the world around the claim rather than the claim itself, and none of them take longer than a minute.
Check One: Who Else Is Reporting It?
Start with the size of the claim. If something enormous just happened — a leader resigned, a country closed its borders, a famous person died — then the story is not going to exist in one place. Enormous events leave marks everywhere.
So search the plainest version of the claim, in three or four words, and look at what comes back. If the answer is silence everywhere except the post that told you, that silence is the finding. It is not proof of a lie, but a real event of that size with no coverage is a contradiction that needs explaining.
The trap is counting copies. Twenty accounts posting the same paragraph is not twenty sources. Look for somebody who would have their own way of knowing rather than somebody who reshared.
Check Two: Where Did the Picture Come From?
A huge share of misleading posts are honest photographs with a dishonest caption. The picture is real; it is from a different country, or from six years ago, or from a film set.
The check is a reverse image search, which most browsers and phones can do from the image itself. What you are looking for is the earliest appearance. If the same photo shows up attached to a completely different story two years earlier, the caption in front of you is the invention, not the image.
Check Three: What Is It Asking You to Feel?
Notice your own reaction before you decide anything. Posts that make people furious or vindicated get forwarded more, so more of what reaches you was shaped to do that. That is a fact about distribution, not about truth — a true story can be enraging.
Treat the strength of the reaction as a speed limit rather than a verdict. The specific danger is a claim that is exactly what you already believed. That is the one you will check least carefully, and it is the one someone would build if they wanted you to spread it.
Check Four: Who Is Actually Speaking?
Open the account. Not to judge the person, but to see whether the account existed last month. Accounts created recently, posting constantly about one subject, with a name one character away from a real organisation, are a pattern worth recognising.
The same goes for the site behind a link. A page can carry a masthead, a logo and an About section and still have been made last week. The question is not whether it looks like a news organisation. It is whether anyone has ever heard of it.
Check Five: What Would Change Your Mind?
This one is uncomfortable and it is the most useful. Ask yourself what evidence would make you drop the claim. If the honest answer is "nothing would", you are not evaluating it any more, you are defending it.
It works in the other direction too. If someone tells you a thing you did not want to hear, the same question applies: what would it take to convince you? Having an answer is the difference between being sceptical and being stubborn.
The Part People Skip: Forwarding Is Publishing
Sending something to a group chat with "is this real?" attached does not stay attached. It gets screenshotted, it gets forwarded without your question, and it arrives at the next person as something somebody they trust sent them.
If you genuinely do not know, the honest move is to check first or to say nothing. "I could not find anyone else reporting this" is a perfectly good thing to send.
The Digital Ethics & Citizenship test puts eight of its twenty-four situations here, and several of them are deliberately built so that no check gives a clean answer — because the real skill is knowing what to do when you still cannot tell.