February 10, 2006

Freakonomics

by Brian Turner

If you're new here, you may want to subscribe to my RSS feed. Thank you for visiting!

Freakonomics, by Steven D Levitt & Stephen J Dubner, opens up claiming to be an enigma.

It states there is no unifying theme - then begins with the assertion that information is controlled by experts who misuse their position.

And then Freakonomics seems to do precisely that.

There’s a lot of interesting information in Freakonomics - but it’s little more than one man with a few ideas (Steven D Levitt), who is intent on “proving” superficial ideas with statistics - which are interpersed with sycophantic dribblings about him by another (Stephen J Dubner).

While Levitt is sometimes keen to criticise flaws in other data collection and analysis techniques, he fails to open his own claims up for criticism. After all, he has statistics, and statistics can “prove” things.

Well, actually, they can’t. There’s a claim of Steven Levitt being a “rogue economist” - but the study of economics is not about science - it’s about guesswork. Or, more to the point - “lies, damn lies, and statistics”.

So what we have here is simply one type of statistician making comments on social issues.

That means - as with all things statistics - a slew of presumptions, generalisations, and conclusions built on sand.

There is some great information in this book - and it is a genuinely interesting read.

But at the end of the day, Levitt proves little else other than his is skilled at using statistics to prove his own points. I’m sure there’s a great political career open for him if he’s ever interested.

If I sound somewhat critical, it’s because I trained in the physical and social sciences, and long ago learned that statistical analysis at best provides a superficial view of the world. It’s a valid view, but made alone, it lacks the necessary context and detail to justify far-reaching conclusions made from them.

The trouble in this instance is that Levitt chooses social issues relating to crime and parenting as a focus for Freakonomics - and then expects to reduce these complex subject areas into single instances of cause and effect - a trait many statisticians are crippled with.

I have no idea why the publisher thought it a good idea to include Dubner’s ramblings about the greatness of Levitt throughout the book - but it’s a complete waste of space. If a publisher needs to include a salesman in a book to continually try and sell you its value, perhaps that speaks volumes about the quality of the book itself.

Overall, I read Freakonomics hoping for some interesting information on marketing issues. Instead I ended up with something that’s little more than pop sociology.

EDIT: The original treated both men as joint authors of Freakonomics - now corrected.

Discuss this in the Internet Business forums

Story link: Freakonomics

Add to Bookmarks: