In SuperFreakonomics, we write about shark attacks as anomalous events that tend to make a lot of noise and therefore persuade many people that they aren’t such anomalies.
A reader named Michael McDonald sends along encouraging evidence that at least one person is thinking clearly on the matter. A young girl named Caitlin Dubois gets nibbled on by a shark in New Smyrna Beach, Fla., and won’t even consider staying out of the water:
“I like swimming in the ocean,” she said. “It’s a freak thing, and a one-in-a-million chance that I would get bitten by a shark. So it really wouldn’t happen again, I don’t think.”
Here’s what Michael wrote:
Could not help but read [this] and smile that this 10-year-old girl grasps the irrationality of being afraid of something that 1) has overwhelmingly long odds of ever happening to us (shark bite); and 2) has already happened to her. Somehow this little girl understands what so many adults in a similar situation would walk away with the opposite conclusion (the mental certainty that they are somehow “prone” or at least very likely to experience another bite). Maybe our next generation isn’t as troubling as our bias leads us to believe….or, more likely, this girl is untainted as of yet by adults’ fear-mongering.
Surely we don’t know enough about her yet to start a Caitlin-for-President movement, but still, it is tempting.
and 
. You can also watch it on 
I wouldn’t be surprised if she were scared anyway, though, even if she knows it’s irrational.
I admire her courage but there’s still an error in her reasoning. She seems to think because the rare event has already happened to her, there is no chance of it happening again when the probability really is exactly the same as it was before she was bitten. Still, she shows a bit more of a grasp of the reality of the situation than most adults. As far as Caitlin for President, I don’t know but if she were in the Colorado Senatorial race (or even, maybe Gubernatorial), she’d have my vote, far superior to the current pack in reasoning skills and lack of bias.
If Caitlin–God Forbid– was eaten and killed by a shark, she would not give the same smarty comment.
You are looking at a survivors bias. Talk instead to the parapalegic whose life is permanently altered and is unable to swim because they are missing legs and arms.
This is a great story, and thanks to Michael for pointing it out. I couldn’t help but notice, however, that Michael seems to be suffering from a bit of irrationality himself. If successive shark bites are independent random events (which I assume they are), it doesn’t make any sense to add “and 2) has already happened to her.” Having already been bitten by a shark neither decreases nor increases Caitlin’s odds of being bitten again, just like flipping heads 400 times in a row neither decreases nor increases my odds of flipping heads on the 401st attempt.
Rudiger, if the situation involved a car accident rather than a shark attack, would you be saying the same thing? I really don’t understand how her reasoning has anything to do with a survivorship bias. Maybe it would be a survivorship bias if she was saying something like “when I get bit again, I’ll probably live, since I survived the last attack”.
I think the point behind (2) is that those to whom a highly-improbable event has occurred would typically overestimate its likelihood, whereas Caitlin has managed to avoid such a biased adjustment to her estimate of the probability.
It’s not clear from her comment that she is using the gambler’s fallacy (that because it happened it’s less likely to happen again). The base rate of shark attacks is so low that her comment could simply mean the chances have returned to the base rate (rather than actually being lower post-shark attack). The difference between the base rate and zero is quite small so it would be hard to distinguish the two beliefs based on casual language (“it really wouldn’t happen again, I don’t think”).
Her comment is impressive in either case because she’s resisting the human tendency to assign too much meaning to personal experiences (especially very memorable, emotional experiences like shark attacks) in predicting future events.
Brian -
I think Michael would have been much better served by saying, “… yet 2) …” instead of “… and 2) …” I like to think that’s what he meant.