Sharks at the Beach

Saturday Morning Breakfast Cereal posted a comic today that does a good job of illustrating a metaphor I like to use. “Sharks at the Beach.” The shark bit comes in at the little red vote button. The comic is about how bad people tend to be at evaluating risk, mixed with a dash of political commentary about the “War on Terror.” Educating a person about statistics or even simple probability can be a big help in getting them to understand the real risks in life.

For those who haven’t read my use of the “Sharks at the Beach” metaphor before, it’s an illustration I like to use about bias in evaluating risk. A lot of people tend to fear dramatic, spectacular, or otherwise “showy” risks, like being attacked by a shark at the beach. At the same time, they generally ignore the mundane risks of everyday life, like getting into a fatal car accident while driving to the beach. An encounter with a shark would be quite dangerous to life and limb, but such encounters are unlikely to happen, especially for people who don’t live near the coast. Going on an occasional beach vacation isn’t likely going to lead to a shark attack.

Opposite that, for many people, driving is a frequent activity. A single car trip isn’t likely to end badly. Because we have frequent safe trips, many people end up dismissing the danger and file driving as a “safe” activity, while risks with sensational results are deemed “dangerous,” even if the hazard is unlikely. A lot of people end up extending this way of thinking to black-and-white extremes and can end up insulating themselves due to overhyped mass media scares or neglecting basic safety measures in everyday activity because they don’t perceive the risk. One example of the former I recall hearing on NPR was a woman’s mother who, a few years after 9/11, was terrified of her daughter traveling by airplane because her long blonde hair would make her stand out as a target for being taken hostage by terrorists. Despite terrorist incidents, travel by airplane is still pretty safe compared with other methods of long distance travel.

The political aspect in this is the cost/benefit ratio of many anti-terrorist measures. As I see it, every new anti-terrorist countermeasure involves some cost to innocent people, and there’s going to be a point where that cost outweighs the benefit of preventing attacks. The law of diminishing returns applies. Spending a little more effort at airports looking for suspicious activity and potential bombs is perfectly understandable. On the opposite extreme, there’s panicky civilians and racially-profiling security who cause trouble every time someone brown-skinned or dressed up in exotic attire does something vaguely odd but innocent (or for merely existing) and act on it. Too easily this ends up as theater where innocent people are dragged off the plane by security, presumed guilty by the public, and the troublemakers are effectively rewarded for their paranoia. More sensitive detection means more risk of false positives and associated costs. With all the disproportionate fear being spread around by demagogues, who needs terrorists?

Advertisements

3 responses to “Sharks at the Beach

  1. One thing to remember about probability is that it isn’t identical for every single individual. One can give a “general” probability for being attacked by sharks vs being attacked by terrorists vs being killed in a car accident, but there are extenuating circumstances that will change those odds. For example, I live in the very center of America. The odds that I will be attacked by a terrorist are vanishingly small compared to someone living in a major city or working at a famous landmark of a place. Also, I am landlocked. The odds I’ll be attacked by a shark are effectively zero, so long as some freak accident involving a shark delivery and a plane crash doesn’t occur. SKY SHARKS! Protect your loved ones! Should we allow these dangerous animals to be carted in the unassuming clouds above? It’s a ticking time bomb! Wolf Blitzer reports.

    By the same token, car accidents are higher and lower for individuals based on where they are driving, how often they drive, and HOW they drive (also how the people around them are driving). Here’s another one. We’ve all heard about the odds that any glass of water you drink has at least one molecule of water that has passed through the bladder of Oliver Cromwell. The naked statistics work out, but only if the assumption that all water had an equal chance of passing through Oliver Cromwell’s bladder holds. Well, turns out it doesn’t. If your cool drink of water came from that recently discovered underground water in a Canadian mine, the odds of any of it having passed through anyone’s bladder are zero, because that water was locked away for a billion years, before bladders were invented.

    Odds are contingent, and the more “generalized” a statistic the more likely it’ll turn out to be wrong at the individual level, even if right at the scale it was intended for.

    Sorry, I’ve been reading “the Signal and the Noise” lately.

    • It’s okay. Probability’s a complex topic, so it’s hard to cover all the aspects. As you’ve probably guessed, I skipped the nuance for the sake of talking about common flaws in human perception of risk.

      On your water example, it reminds me of one quirk of carbon dating: There are some ecosystems that are isolated enough that they don’t get their carbon-14 replenished like most other ecosystems do, leaving the inhabitants with mostly carbon-12 to consume. This makes the organisms look older than they really are if you treat the results like those of an open ecosystem. Of course, Creationists love to treat that narrow case as if it invalidates all radioisotope dating via the Perfect Solution Fallacy. Scientists, on the other hand, know that all tools have flaws and limitations to take into account when using them.

  2. I know I’m late, but here goes…
    My favorite “probability trope” WRT driving, is the one where the driver refuses to wear a seatbelt because of the probability that he’ll be crashed into a river and get drowned in the car cuz he can’t undo the seatbelt fast enough. Totally disregarding the probability he’ll land in a river, nevermind that the car will sink faster than he can get the seatbelt off, and completely ignoring all the other possible events (more probable events) that the seatbelt protects one during.
    Just my small contribution, so sick of hearing that one.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s