Which is more valuable: statistics or experience?
Don't answer right away; take all the time you need. I'll wait. This question isn't as obvious as you might think, and the first answer you're tempted to give might not be the one you decide to stick with.
Humans are not born statisticians. Let's face it, most of us are lousy at statistics. My brilliant friend Julia Galef, who publishes articles, hosts podcasts, and gives TED talks on rationality, holds a degree in statistics from Columbia, and can quote you statistical formulas like nobody's business -- but even she decided not to make her living as a statistician.
We tend to trust personal experience more than abstract numbers. This works very well for children. Once you've burned your hand on a hot stove, you've learned to be careful around stoves. Once you've been stung by a bee, you'll always be wary of bees -- and probably a whole lot of harmless insects that happen to look like bees. Telling a child that 34,000 children under the age of eight are seriously burned by stoves in the United States every year probably won't have the same effect.
But what happens as we get older? We're exposed to more complex situations, and we begin to face what's known as relative risk: is it safer to ride the train home from the city late at night through some sketchy neighborhoods, or to drive when you're already tired and had two drinks with dinner? And humans are notoriously bad at assessing relative risk. For example, after 9/11/2001, thousands of Americans who had been planning airplane trips changed their plans and decided to drive instead, fearing more hijackings. But hijackings remain an exceedingly rare event, whereas motor vehicle accidents are quite common. So it's safe (if you'll pardon the use of the word here) to say than many more Americans died on the road who would have lived to a ripe old age had they simply flown to their destinations.
Printing statistics of lung cancer deaths from smoking in the newspaper doesn't get people to quit smoking. Airing a public service TV spot showing someone with a laryngectomy and multiple amputations from smoking does; all those nicotine fiends see that and think, "Holy snot, I sure don't wanna end up like that!"
In my own field of medicine, the way should be clear. Academicians deride those of us "in the trenches" by suggesting that we depend excessively on the imaginary Journal of Anecdotal Medicine. We are often brought up short by well-designed studies that find that some treatment that we've been using for years is actually largely ineffective, or frankly worse than no treatment at all. Statistics save lives. Yet, there are countless instances I can think of from my own practice when I intuitively knew the best course of treatment for a patient based not on the dry statistical analysis at the end of the latest study, but on my years of clinical experience. I could have read the same studies as a younger, more inexperienced clinician, but the patient would still not have fared as well.
So what am I saying here? We skeptics need to find the best ways to integrate statistical findings with personal experience. Throwing a snowball onto the floor of the Senate does not allow one to proclaim that one has disproved human-induced climate change. Personal experience should not be discarded out of hand, but should be viewed...skeptically. Conversely, statistics don't always give us the whole story, either. Reading the numbers usually isn't a powerful enough tool to sway the emotions. And we humans are, at base, emotion-driven beings.
Don't answer right away; take all the time you need. I'll wait. This question isn't as obvious as you might think, and the first answer you're tempted to give might not be the one you decide to stick with.
Humans are not born statisticians. Let's face it, most of us are lousy at statistics. My brilliant friend Julia Galef, who publishes articles, hosts podcasts, and gives TED talks on rationality, holds a degree in statistics from Columbia, and can quote you statistical formulas like nobody's business -- but even she decided not to make her living as a statistician.
We tend to trust personal experience more than abstract numbers. This works very well for children. Once you've burned your hand on a hot stove, you've learned to be careful around stoves. Once you've been stung by a bee, you'll always be wary of bees -- and probably a whole lot of harmless insects that happen to look like bees. Telling a child that 34,000 children under the age of eight are seriously burned by stoves in the United States every year probably won't have the same effect.
But what happens as we get older? We're exposed to more complex situations, and we begin to face what's known as relative risk: is it safer to ride the train home from the city late at night through some sketchy neighborhoods, or to drive when you're already tired and had two drinks with dinner? And humans are notoriously bad at assessing relative risk. For example, after 9/11/2001, thousands of Americans who had been planning airplane trips changed their plans and decided to drive instead, fearing more hijackings. But hijackings remain an exceedingly rare event, whereas motor vehicle accidents are quite common. So it's safe (if you'll pardon the use of the word here) to say than many more Americans died on the road who would have lived to a ripe old age had they simply flown to their destinations.
Printing statistics of lung cancer deaths from smoking in the newspaper doesn't get people to quit smoking. Airing a public service TV spot showing someone with a laryngectomy and multiple amputations from smoking does; all those nicotine fiends see that and think, "Holy snot, I sure don't wanna end up like that!"
In my own field of medicine, the way should be clear. Academicians deride those of us "in the trenches" by suggesting that we depend excessively on the imaginary Journal of Anecdotal Medicine. We are often brought up short by well-designed studies that find that some treatment that we've been using for years is actually largely ineffective, or frankly worse than no treatment at all. Statistics save lives. Yet, there are countless instances I can think of from my own practice when I intuitively knew the best course of treatment for a patient based not on the dry statistical analysis at the end of the latest study, but on my years of clinical experience. I could have read the same studies as a younger, more inexperienced clinician, but the patient would still not have fared as well.
So what am I saying here? We skeptics need to find the best ways to integrate statistical findings with personal experience. Throwing a snowball onto the floor of the Senate does not allow one to proclaim that one has disproved human-induced climate change. Personal experience should not be discarded out of hand, but should be viewed...skeptically. Conversely, statistics don't always give us the whole story, either. Reading the numbers usually isn't a powerful enough tool to sway the emotions. And we humans are, at base, emotion-driven beings.
Comments
Post a Comment