Tuesday, 14 July 2009

Alex's Book Club - Freakonomics by Levitt and Dubner

Well, that was disappointing.

I was kind of looking forward to Freakonomics for a long time. The basic idea of someone sifting through data to find out the truth's about modern life that other people overlook is an enticing one, and I read so many little glowing reviews that I figured once I got round to it, it would be a very satisfying read. Sadly it was mostly trivial and annoying, with a few nice bits of insight but almost nothing that raised the proverbial eyebrow. Or, my eyebrow should say, nothing very proverbial about it.

I'd recommend it if you haven't really... y'know thought about the world much before. In that case I reckon it'd probably make you "hmmm" and make you "ooooh" and all that. But really it's pretty common sense stuff backed up with figures. It's a pop-economics presentation of Levitt's work I guess; he's a hot young thinker fella and someone had the smart, bankable idea of putting some of his famous theses into an easy to read book. I'm not sure I'd really argue against any of the points he makes here. It's the book itself that really pissed me off.

Because even when I think about it being written specifically to be easily accesible, I think of shortcomings. Later in the book they go (wisely) into the difference between causation and correlation, and throughout the book they chastise other writers for confusing the two. But some of the points earlier in the book are much too quickly or flippantly expressed - it feels like they're substituting their own pat, conventional generalisation for the currently accepted pat, conventional generalisation. And in doing that it sounds like they're mistaking correlation for causation. I'm sure they're not, I'm sure Levitt's covered his ass on that, but reading the book there were a few times when I thought - but why is your reading of the data "causative" while their reading is "correlative"?

And their claim that there is no underlying theme. Well, as others point out, there kind of is and it's about incentives and what people do to get what they want. But even so the chapters just feel so messy, just a random series of statistical patterns that stick with the same general theme, but at times switch clumsily to explore a new area of that theme. The constant two bit oratory too - "crack is like nylon stockings! and here's why!" I found it very grating.

Oh, also the final chapter on kids names. Man, I just read Steven Pinker's The Stuff of Thought where he covers the same topic, and it was amazing. This was essentially saying "Poor kids' names and rich kids' names are different! And they change in trends!" Pinker did a much better job of trying to analyse why this happens.

But, by far the most annoying thing in this bloody book is the stuff about Levitt himself. I'm sure he's all that they say and more, and a helluva guy, but I was sick of being told what a genius he is the first time it came up, by the end of the book he was perilously close to making it onto the list of people I want to punch in the goddamn face. Seriously, he's the co-author of the book and yet there are numerous excerpts from other publications describing him as a genius, a demi-god, yet humble and possessed of a mind and spirit unlike anyone else. I'm putting two fingers down my throat and making comedy wretching noises right now. It must be the co-author Dubner who wanted to put that stuff in, and man, it's horrendous. There is no justification for padding out your book with chapter breaks that tell your reader how amazing the man whose ideas they're reading is. Don't tell me he's amazing, show me! Blow my socks off with data based revelations that will shake my view of the world!

Freakonomics certainly didn't do that for me. I don't recommend it, and if you want to read it, I'd say check out the Digested Read of it on The Guardian first.


  1. I couldn't agree more! I was pretty excited by the idea, but just ended up feeling 'meh'. Oh, and wanting to kick both authors in the balls. Gladwell covers similar territory with more coherence and flair, but with less intellectual clout (I guess, given I'm not an economist or whatnot). Have you read Fooled by Randomness or The Black Swan by Nicholas Nassim Taleb? They're both much deeper treatments of work in statistical science, although they still have the kick-the-author-in-the-balls-ness.

  2. I've got one of Taleb's books. Best 70% off purchase at Junkudou ever. You can borrow it if you feel like reading it. I expect my Andrew WK in return though.