This guy read a headline. You'll never believe what happens next.

This article is an ongoing work in progress, not an article with a publishing date. The listed date reports when the article was started but content here may be newer.

The Web is full of misleading information. This is not new: old media was full of misleading information as well, but old media didn’t have the rapid dissemination that we do now so the competition for attention has ramped up. Headline writing on websites has turned into an obscene craft whereby the contents of the articles often bear no resemblance to the headline.

This is compounded by the fact that new headlines and taglines are often added after the fact, often by third parties resharing an original piece but also by other departments within an organization.

These are some guidelines I find valuable when looking at headlines. Honestly, I get fooled by these all the time, but now at least I can collect them here for future reference.

If the headline comes in the form of a question, the answer is «no».

This idea, stated less strongly, is known as Betteridge’s law due to Ian Betteridge’s 2009 article TechCrunch: Irresponsible journalism, which is worth the read. What a relief it was to finally read someone put forth the proposition:

And please, please, I hope no one brings up that old chestnut of “it’s only a blog, we don’t have to adopt proper standards for reporting”. The moment you can have a serious effect on a company or individual, you owe it to the world to be sure of what you say.

I’ll add that it’s a «serious effect» if a single individual believes you. Which doesn’t mean anyone has to adopt proper standards for reporting. Just be honest.

One study is news. Plural «studies» is an advertisement.

Multiple studies on the same topic are rarely published on the same day. So when a headline uses the plural to describe research, it’s telling you right up front that it doesn’t have anything new to say, but is trying to promote an idea.

This doesn’t mean it doesn’t have anything valuable to say. It’s perfectly legitimate to reexamine data in light of other data on the topic. But when writing about that one should not lead with the fact that the studies exist, but with the proposition itself. That is, unless you are trying to mislead people into thinking that your article contains new information.

«Followed by» means the reporter couldn’t establish any causal link, but still wants you to think there is one.

The fallacy here is that correlation is causation. Most people do understand that, but it’s easy to ignore when skimming headlines. Certainly, effects do follow causes (it’s actually the definition of past and future) but generally when someone wants to say that one event caused another, he or she will say so outright using phrases like «from», «because», «as a result of», or «caused by».

If someone dies of exsanguination as the result of being stabbed with a knife, the headline will likely read, «Area Man Dies From Knife Wounds.» Perhaps for the sake of brevity it will be shortened to the less explicit, «Stabbing Victim Dies». But when the less-explicit headline is longer than the explicit version, that is highly suspect. «Area Man Dies Following Knife Attack» could mean that a witness to the attack had a heart attack from the traumatic sight. Or there could be no relation whatsoever. The headline is still literally true but misleading.