June 11, 2010

Excess Kurtosis

Last Christmas, my father gave me Outliers by Malcom Gladwell. As a Gladwell devotee, he must have felt compelled to evangelize. He described Gladwell as a “contrarian” and a “fresh thinker.” With apologies to my father, it’s a complete crock.

Outliers is absolute trash. The book’s introduction details an immigrant community in Pennsylvania that mysteriously displayed a much lower incidence of heart disease than the rest of the country. The mystery had doctors flummoxed. Gladwell intends the allegory as an example for the rest of the book. Instead of finding concrete reasons why the community has a lower incidence of heart disease, researchers gave up their exhaustive methods, and simply came to the conclusion that people were healthier because they talked to each other more. There existed a set of circumstances that scientists couldn’t explain in the 1950s, so they instead turned to mysticism. This example should be a beacon of scientific failure. Like a toddler that takes on a toy that’s a little too adult for him, they saw pieces that didn’t fit together and inferred magic in the overall construction.  Far from decrying this  conclusion, Gladwell sees it as the open-mindedness of those researchers, and announces his intention to emulate it.

So Gladwell’s opus on success is about as comprehensive as the Q volume of an encyclopedia. In part, Outliers fails because Gladwell is so manifestly dumb as to argue a point that is universally conceded (and indeed definitionally mandated)—that cause and effect are correlated—and arrogant enough to regard his own tripe as novel. More glaringly, however, Outliers offends my sensibilities because Gladwell, despite reprinting the dictionary definition of an outlier in his introduction, doesn’t seem to understand the statistical concept in the slightest.

A population doesn’t have any outliers. Samples have outliers. Outliers are data points that are excluded from the sample because they skew the statistical accuracy of a group that is selected to represent the population. If you take a sample of five people, and one of them happens to be Bill Gates, their average net worth is going to well exceed the average of the entire population, no matter who the other four members are. Bill Gates is, in the context of this sample, an outlier.

Gladwell, however, never defines a sample. Instead, he takes the most exceptional people in the population of both all present and historical people. His own meager experience guides him to anecdotal illustrations of both the unassailably successful and the disappointments to their own potential. Incidentally, both sets are outliers (one positive, one negative) of Gladwell’s sample of people with extraordinary potential. The opportunity for selection bias enormous. In a display of dizzying statistical and logical nonsense, Gladwell actively seeks to define a population exclusively by its outliers on one side of the spectrum. This would be infuriating enough, but Gladwell never defines the spectrum. He doesn’t define success. He doesn’t define potential, outside of a high IQ score. He doesn’t look for hard data anywhere except birthdates of Canadian hockey players. Instead, Gladwell rails against the social scientists who do believe that success can be quantified, instead favoring the idea that fate is fickle and life is in the tea leaves. Why do some people with high IQs succeed wildly and others fail miserably?  

You might as well ask why Daryl Strawberry is never going to be a Hall of Famer. We know why; he snorted away his talent and spent the best years of what should have been among the best career in baseball history strung out. Is there any reason why we should be similarly surprised at the lack of academic success of a man from Montana with a 195 IQ that chose to spend the bulk of his intellectual peak as a bouncer?

Gladwell is a simple iconoclast. He doesn’t have a legitimate argument to make; he simply wants to tear down the accomplishments of the massively successful and provide an excuse for the failures. His arguments are fallacious, his logic filled with non-sequiturs, leading arguments, and selective inclusion. I realize that Gladwell is emblematic of the style of writing, particularly in business books, that is exceedingly popular. It shouldn’t be. 

No comments:

Post a Comment