Ten years ago, during a routine checkup, my doctor asked me if I had a gun in the house. He thought that having a gun in the house was a serious hazard to my young children. What a dufus. Since my children were not safe crackers, I didn't worry about it. Besides, according to economist Steven D. Levitt, yearly in the U.S. "there is 1 child killed by a gun for every 1 million-plus guns." Levitt also observes that yearly there is one drowning of a child for every 11,000 residential swimming pools.
So why didn't the doctor ask me if I had a swimming pool? Why do people people grossly overestimate the danger of trivial risks, while ignoring the danger of greater ones? Why do I see obese people smoking cigarettes slopping on sunscreen to avoid skin cancer? Why do people who spend hours driving to work each day on busy freeways worry about dying in an airplane crash? Cripes, auto accidents and falls kill 25 times more people in the U.S. each year than airplane crashes and firearms accidents combined. So what's wrong with these people? I want to hit them on the head with a bat.
The answer is The Availability Heuristic. Airplane crashes are far more dramatic than auto accidents, and remain more accessible in a person's memory. Similarly, an accidental gun death, especially if a child is involved, is far more dramatic than a falling, choking, or drowning death -- and so also remains more accessible in a person's memory. A mass homicide causes widespread consternation, even if the risk of dying in one is vanishingly small.
This problem is compounded by the news media, which (1) perseverates over dramatic
events, regardless of how inconsequential they may be to the reader or viewer, and
(2) never puts anything in perspective.
At 6:05 p.m. CDT on August 1, 2007 the main spans of the I-35W bridge crossing the Mississippi river in downtown Minneapolis, Minnesota collapsed. Thirteen people died as a result. A media feeding frenzy ensued with calls for reconstruction of practically every bridge in the United States, regardless of cost. These demands were supported by authoritative statements from such unbiased authorities as the CEO of the Minnesota chapter of the Associated General Contractors, who "said everything MnDOT does is based on cost-benefit analysis." (See Minneapolis Star Tribune, August 3, 2007). The death toll amounted to 1/3,000 th of annual U.S. highway fatalities (42,642 in 2006, according to the NHTSA), but of course nobody in the news media mentioned this fact. Widespread discussion of gephyrophobia ensued, with post-traumatic stress counselors offering advice to Californians traumatized by media coverage of the event. The story soon faded as attention turned to a Utah mine collapse in which three (3) miners and as many rescuers were lost. Mooing and lowing about workplace safety quickly replaced babbling about bridge safety.
The net result of all this is that public policy is distracted in bizarre and useless directions. Instead of cautioning Americans to watch their weight, quit smoking, wear seatbelts and life jackets, and take care when using ladders, we have endless grandstanding over safety measures like trigger locks on handguns (a gun safe would make ten times more sense), radon amelioration, and photocells on garage door openers. These are nice ideas, but require much effort and expense for trivial results. Better to spend the effort where it will have a detectable effect.
Solutions:
Journalists should:
- Understand risks, costs, and benefits. News organization should offer a mini course (a few days) in probability theory and statistics, and require that all employees take it and pass.
- Temper dramatic stories with a few facts. When speaking of airplane crashes, bridge collapses, homicides, and firearms accidents, it might help the reader/listener/viewer to note occasionally that his odds of being killed in such a way are extremely remote.
- Insist that anyone presenting statistical data identifies, specifically, the source thereof. If he cannot do so, either disregard the data, or report it "as is", noting that it is probably unreliable.
- Insist that anyone suggesting a safety measure gives a reliable estimate of the total cost of the measure (including indirect costs), and a reliable estimate of value of the benefits to be achieved thereby. Spending a billion dollars to save one life per year is not economically sustainable.
- Scrutinize closely information provided by government officials (who often just want more money for their own departments), consumer groups, and industry groups, all of whom have axes to grind.
- Disregard completely any information offered by product liability attorneys, including information from shill organizations employed by them.
- Understand risks, costs, and benefits. One would think that they do, but often they do not.
- Understand risks, costs, and benefits.
- Insist that probability theory and statistics are taught in our high schools and colleges.
- Not permit elected representatives to distract the public by concentrating attention on bizarre and trivial risks, while disregarding truly important matters.
References
- Deaths: Preliminary data for 2011 (National Vital Statistics Reports, Volume 61, Number 6, October 10, 2012). Also available here if the previous link changes. Table 2 is available in spreadsheet form thanks to me.
- For data on the percentage of homicide victims with criminal records, consult the many sources available. This varies widely by geography, exceeding 90% in some cities.
- For entertaining information on irrational thinking and how to avoid it, consult the works of Steven D. Levitt (economist), John Paulos (mathematician), Garrett Hardin (physical scientist), etc.