A SuperFreakonomics Video Preview

Yesterday, HuffingtonPost ran this cast of characters you can expect to hear about in SuperFreakonomics. It was accompanied by a video preview in which Dubner’s juggling practice pays off and Levitt wonders why college students, if they’re so altruistic in the lab, never give him money in the subway:


Heavy D

SD, can I be the first person to comment?

charles

I liked it. I bet you're going to make some people angry. Fun.

Marcos

Wait, so the book is out...? :O

Webster Hubble Telescope

Everyone should also understand how wrong Levitt and Dubner get the topic of oil depletion, AKA Peak Oil.

They do not understand the fundamentals behind peak oil, which ultimately surprises me since oil depletion analysis is nothing more than probability and statistics, which they purport to be authorities on. I myself consider it no more difficult than bean-counting.

This is what they have said in the past:
" So why do I compare peak oil to shark attacks? It is because shark attacks mostly stay about constant, but fear of them goes up sharply when the media decides to report on them. The same thing, I bet, will now happen with peak oil. I expect tons of copycat journalism stoking the fears of consumers about oil induced catastrophe, even though nothing fundamental has changed in the oil outlook in the last decade. "
http://freakonomics.com/2005/08/21/peak-oil-welcome-to-the-medias-new-version-of-shark-attacks

At one time, I thought Freakonomics was all about looking at the statistics and basic math underlying a premise and trying to debunk or support that premise. They have often been able to do this by demonstrating how that almost certain correlations between cause and effect were simply anomalies that could not overcome the null hypothesis.

But here Freakonomics says “I don't know much about world oil reserves.” So, with that, how can they predict anything, one way or another, on how things will turn out?

They don't even mention that the USA has gone through its own peak oil in the 70's. This could have formed the basis for a classic Freakonomics statistical studies from that well-understood set of data. Economists largely ignore this topic, yet I thought these guys were FREAK economists, willing and able to buck the tide. Yet they don't do that.

Please, why don't you cover this topic properly?

Read more...

Sugato

Wait! Wouldn't you want to eat *more* cows so that there are fewer of them left to emit the methane?

DaveyNC

No, if we eat more kangaroos then ranchers will stop raising cattle and start raising 'roos. More Rooburgers equals fewer Mooburgers.

Would a Rooburger for a kid be called a Joeyburger?

Ordering the book now.

Clayton

You may put yourself in more danger by walking drunk, but I doubt more innocent victims have been killed by drunk walkers than by drunk drivers.

ceolaf

What poor poor analysis.

OK, walking drunk a mile is 8 times more likely to result in your own death than driving drunk a mile.

But what about causing other people's death? Shouldn't we take that into account?

What about non-lethal injuries?

And what about the cost in property damage? Like what about damage to your own car? Isn't property damage MUCH more likely than lethal injury?

So, if you only taking into account the risk of your own death, you might come to one answer. But you'll also be ignoring most of the issues, including the most likely ones.

SoCalStudent

The stat, "walking drunk is 8 times more dangerous than driving drunk", is supposed to make people think twice about walking or driving home when their intoxicated?

I would bet that study was flawed. You easily can pull such data from the end effect... e.g. polling the patients in a hospital emergency room on a weekend night, asking them if they were walking or driving home. I doubt that a researcher polled all the people walking home that DIDN'T get into accidents. Or that they checked the mean blood alcohol percents of both walking & driving groups and those that were or were not in an accident....

People were are too falling-on-their-face drunk to drive, i.e. are physically incapable of driving, usually attempt to walk home, if within walking distance. This one example alone shows how the sample can have bias error.

Perhaps my intuition is wrong on this, but I've been drunk plenty of times and know that when I'm within walking distance of home --> I tend to drink more because I know I don't have to drive back. It is safer to walk than to drive. I'd be a great danger to many if I drove when I'm obviously drunk.

In short, I thought the "8 times more dangerous" stat quote was obviously flawed and was of the type that you guys debunked instead of propagated.

Read more...

Jeff

Kudos for pointing out the problems inherent in consuming beef, but why such an impractical solution!!!!?

People might hear that and go huh, but do little. You could instead replace some beef meals with vegetarian meals, or even chicken, and do real good.