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.

Leave a Reply