Shark Attacks, Availability Bias, and a Possible Conspiracy by the Shark-Industrial Complex

Everybody collects something. We start as kids: dolls, action-figures, cards, coins. As adults, we either give up those collections or intensify our efforts. Like sharks, our interests can only live by moving forward. But if they die, other obsessions take their place. One of my friends collects images of unexpected groups of famous people. (Lou Gehrig and the Marx Brothers, Muhammed Ali and the Beatles, Nancy Reagan sitting on Mr. T’s lap.)

But who am I to judge? I collect shark-attack stories. Rather than just watching from the sideline (shoreline?), I have contributed to shark literature. Two of my favorite blogs have concerned the prevalence of shark-attack stories out of proportion with their actual threat. Two summers ago, I examined the question, “which is more dangerous: sharks or horses?” (Spoiler alert: horses, by a ratio of more than 20-to-1.) Last year, I wrote about the summer’s “fad” danger, which was death by selfie. My attention was piqued by an article pointing out selfie-related deaths now outnumber those caused by sharks. Both of those blogs discussed availability bias, a common behavior pattern in which we misjudge the frequency of certain events due to their relative availability in our memory. A real-world consequence of this is that we spend 25 times as much money fighting terrorism as cancer, even though cancer kills 2,000 times as many people.

I hope my efforts to put particularly cinematic forms of mayhem in perspective are having an effect. You can judge — and potentially reach either conclusion — from a leading Memorial-Day shark-danger article in the Washington Post. The 19-word headline suggests that the Postunderstands and does not want to contribute to availability bias:“Expert warns 2016 could see a record number of shark attacks, but remember, accidental drownings are way more common.” On the other hand, for readers saying, “you had me at ‘record number of shark attacks’,” it took five paragraphs for the Post to mention “statisticians remind beachgoers that the chances of being attacked by a shark are infinitesimal….”

The article is an interesting balance of facts culled to justify shark attacks as news and reminders that such attacks are freakishly unlikely. For example, the article says 98 people were bitten by sharks worldwide last year, “setting a record for attacks.” The director of the International Shark Attack File said 2016 could be a record year for shark attacks. From this, the article deduced, “if so, it would follow a trend that began last year”. How is one data point a trend?

Although being attacked by a shark seems super-gory, none of the shark experts will own up to how few people die from shark attacks. The closest attempt to quantify this is a quote from an Atlantic Journal-Constitution article saying the 2014 “fatality rate for shark attacks in the United States was 1.7%”. If 98 shark attacks is a worldwide total, is it possible that the number that die in U.S. attacks annually is ONE? In contrast “plain old accidental drowning” — that’s a quote from the article — kills 3,500 per year.

Author of Thinking in Bets and How to Decide. Co-founder of The Alliance for Decision Education

Author of Thinking in Bets and How to Decide. Co-founder of The Alliance for Decision Education