Quote:
Originally posted by sebastian_dangerfield
My first reaction to Gladwell and Leavitt was “huh, two guys trying to prove their points by engaging in exactly the endeavors they criticize.” But that’s a dishonest dismissal. Gladwell acknowledges the contradiction by saying the book is meant to entertain, and that, because people are so accustomed to having the reasons behind any decision vetted from every conceivable angle and spoon fed to them, he needs to cite tons of examples to prove his point. I haven’t read all of Freakonomics, but from what I have, I don’t think Leavitt’s saying “statistical studies are generally bullshit.” Isn’t he saying most statistical studies are bullshit because they’re no done completely enough or they’re manuevered to reflect the bias of the group paying for the study?
|
I think that's one underlying implication of Leavitt's book, although he'd just as readily attribute the errors to ignorance. His basic thesis is that people mistake correlation for causation - either willfully to promote their own agenda or simply out of ignorance.
While I agree with the underlying premise, he then sprinkles the book with countless examples where he does
exactly what he's being critical of: correlation instead of causation, or using irrelevant measures (such as time vs. distance in measuring airflight - although I kind of think that one may be an elaborate joke on his part).
