The Headline Was True. It Was Still Misleading.

The annoying thing about misleading headlines is that they don’t actually have to lie. Lying would almost make things easier. If a headline said something completely false, someone could check it, prove it wrong, and we’d all move on. The more interesting version is when every word is technically true.

Take something like “Risk increases by 50%.” That sounds serious. Your brain immediately gives that number some weight because, well, 50% is a lot. Except maybe the risk went from 2 people in 10,000 to 3 people in 10,000. That’s still a 50% increase. Completely true. It’s also a rather different story from whatever your brain probably imagined when it first saw the headline.

This happens everywhere. A company “loses $20 billion in value.” A city sees a “record increase” in something. A new study finds that people who do X are “twice as likely” to develop Y. Support for a politician “surges.” Crime “soars.” None of those statements necessarily need to be false. They just need to leave something out, and humans are extremely good at filling in the missing part ourselves.

Numbers are particularly good at this

Imagine I tell you that “Complaints about Briflio increased 100% this month.” That’s not exactly the headline I’d want floating around. But suppose last month one person complained and this month two people did. Congratulations, complaints have increased 100%. Technically, we have a crisis.

This is why percentages without the original numbers can tell you surprisingly little. The same thing works in reverse. Imagine a disease affects 20 out of every 10,000 people, and a treatment reduces that to 10. You could say “New treatment cuts risk by 50%.” Or you could say “New treatment prevents illness in 10 out of every 10,000 people.” Same data. Very different feeling. Neither statement is necessarily wrong.

That’s the weird part. We tend to think facts and opinions live in completely separate boxes. A fact is objective. An opinion is subjective. But there’s something sitting between those boxes that’s arguably just as powerful: which fact did you choose to show me?

A headline is a tiny window

Suppose unemployment falls. Good news. But perhaps the number of people who stopped looking for work increased. Suppose house prices rise. That’s probably good news if you own one and considerably less exciting if you’re trying to buy one. Suppose a company’s revenue hits a record high. Fantastic — unless its costs grew even faster.

You can describe the exact same reality in several completely truthful ways. That’s why reading news isn’t only about asking “Is this true?” Sometimes the better question is: “What would I think differently if I saw the number next to this one?”

And that’s a much harder question to answer, because the missing context rarely fits nicely into a headline.

Then there’s the timeline trick

This one is almost beautiful. You can make the same graph tell dramatically different stories simply by deciding where it begins.

Imagine something normally sits around 100. Last month it suddenly falls to 60. This month it climbs to 80. One headline could say “Value jumps 33% in one month.” Another could say “Value remains 20% below previous levels.” Both are true.

If you only see the first, you’ll probably think things are booming. If you only see the second, you’ll probably think things are struggling. The disagreement isn’t necessarily about the data at all. It’s about where somebody decided the story should begin.

Once you notice this, you’ll start seeing it everywhere: markets, elections, immigration, crime, salaries, company performance and climate statistics. “Highest since…” “Biggest increase since…” “Down sharply from…” “Up dramatically compared with…” There’s one tiny question that’s worth asking whenever you see language like that: compared with what?

This doesn’t mean the media is secretly lying to you

This is where conversations about media literacy sometimes become a little silly. Someone notices that framing exists and immediately concludes that everything is propaganda and nothing can be trusted. That’s not particularly useful either.

Journalists have an unavoidable problem: reality is enormous and an article isn’t. Something has to be selected. Something has to become the headline. A graph needs a starting date. A journalist has to decide which statistic matters. An editor has to decide which part of a two-hour interview belongs in an 800-word article. There is no version where everything makes it in.

But there’s another problem sitting on top of that: you have to click.

“New Study Produces Interesting but Highly Conditional Result That Requires Considerable Context” is going to have a difficult day competing with “Scientists Discover SHOCKING Link Between…”

The internet rewards certainty, surprise, fear, conflict and enormous numbers. Reality, unfortunately, has terrible marketing. Reality says things like it depends, the sample was small, this is correlation rather than causation, we don’t know yet, or the increase looks large in percentage terms but remains small in absolute terms. Try fitting all of that into a push notification.

Your brain isn’t helping either

We like big numbers. We like simple explanations. And we really like information that confirms something we already suspected.

If I already think AI is destroying employment, I’ll probably notice every headline about AI-related layoffs. If I think AI is creating incredible opportunities, I’ll notice every headline about new AI jobs. Eventually, two people can consume news about exactly the same world and come away with completely different versions of reality.

Neither person necessarily encountered fake news. They may simply have encountered different selections of true news.

I think that’s much more interesting — and slightly more uncomfortable — than the usual conversation about misinformation. Spotting a completely fabricated story can sometimes be easy. Spotting a carefully framed truth requires you to think.

Give the headline five seconds

You don’t need to fact-check every article like you’re preparing evidence for court. But when something sounds particularly dramatic, give it five seconds.

If something increased by 80%, what was the original number? If something is at a “record high,” how far back does that record actually go? If a study says something “doubles the risk,” what’s the absolute risk? If a poll says most people believe something, who exactly was asked? If a company announces “record revenue,” what happened to its profit? If something affected “millions of people,” millions out of how many?

You don’t need a statistics degree to do this. Sometimes you just need to be slightly annoying and ask one more question.

Because the headline’s job is to get an enormous, complicated world through a very small door. Things are inevitably going to get left outside.

And perhaps the biggest mistake we make isn’t believing fake headlines. It’s assuming that because a headline is true, we’ve been given the whole truth.

Those are two very different things.