I was beginning to wonder if I was alone is questioning the drunk walking assertion in SuperFreakonomics. But the WaPo review nails it:
... the authors theorize that it is more dangerous for a tipsy person to walk any given distance than it is for that person to drive. That would be interesting, if true, and certainly useful information for anyone who has ever stumbled out of a downtown Washington bar a few blocks from home.The problem is that Levitt and Dubner don't actually have the foggiest idea whether it's safer to drive drunk than walk drunk, as they claim. As my colleague Ezra Klein has pointed out, they don't have data on how many miles are walked under the influence, and so they just assume that people walk drunk in the same proportion that people drive drunk. In calculating the rate of deaths from walking drunk, then, they have the numerator (the number of drunk pedestrians killed each year) but not the denominator (the number of miles walked drunk).
Here is the climate geoengineering analogy:
It is like a family declining to save for college [ed. reduce emissions] because their 10-year-old Little Leaguer with a decent arm may end up getting a full baseball scholarship [ed. the 18 mile volcano tube will work].
Finally, this entry hits close to home:
"Superfreakonomics" is, like the first book, written in a sprightly, easy-to-digest manner. A reasonably quick reader could finish it on a coast-to-coast flight, with time left to watch a movie. But the feeling at the end is about the same as the one after reading a Dan Brown novel or eating a bag of Cheetos. You finished the whole thing but didn't walk away feeling particularly proud of yourself.
I was 100 pages in to The Lost Symbol with SuperFreak arrived in the mail. I ate a $0.99 bag of Cheetos last night (crunchy, not puffed). Sorry, but I kindof like that not-particularly-proud-of-myself-feeling.
*Not me (I hope). Experimental actually: John List, according to SuperFreakonomics, for running the dictator game 1000s of times in field settings and finding that people aren't as altruistic as lab experimentalists say they are and negating the Nobel Prize in economics awarded to experimental economics (i.e., Vernon Smith and Daniel Kahneman).