Statistical thinking is an unnatural act for humans. We evolved to generalize from single examples. After all, the penalty for caution was lost time, while the penalty for impetuousness could be death. So when I point out to others that statistics show how violence promotes more violence, I don't get much traction. Everyone has a favorite anecdote about the person who behaves counter to the average. The example of one trumps the example of many.
We're better at probability, since we like to gamble. So, when I point out that three successive screens meant to prevent deranged individuals from buying guns — even if individually only 50 percent effective — will result in nearly 90 percent of potential mad shooters being stopped, I might actually get someone's attention. Unless, of course, they are driven by dogma instead of evidence. In that case, I'll be treated to an anecdote about the one who got through, and the arms-bearing citizen who stopped them.
We're excellent at counting. After all, money is involved! So I expect that it's not a stretch for anyone to determine that, on average, a clip that holds 10 bullets kills one-third as many people as one that holds 30. And I suspect that I can even argue past the person who claims to be able to change clips so fast that bullet count doesn't matter. After all, the alleged Arizona shooter appears to be skilled with guns, but was stopped when he paused to reload.
I guess I should stick with counting. Yet some part of me wants to press all three arguments. In most alternative universes where political discourse is more civil, prospective gun buyers are screened multiple times for mental instability, ammunition clips whose only proper use is on a field of battle are banned from the streets, and a nine year-old girl is still alive.
Isn't that worth a bit of compromise?
With a Perspective, I'm Paul Wolber.
Paul Wolber is technical manager for a major Silicon Valley company specializing in measurement products and services.