I Have An Above-Average Number Of Feet
In The New York Times, Barry Gewen reviews a book on stats abuse, The Numbers Game, The Commonsense Guide to Understanding Numbers in the News, in Politics, and in Life, by Michael Blastland and Andrew Dilnot:

It's hard to resist a book that tells you that most people have more than the average number of feet. Or that researchers have found that Republicans enjoy sex more than Democrats do. Michael Blastland and Andrew Dilnot delight in bringing such facts to our attention -- and then in explaining them away.

Because of amputations, birth defects and the like, the average number of feet per person across the human population is slightly fewer than two.

...Most of us, Mr. Blastland and Mr. Dilnot observe, expect numbers to do too much. We like their precision and want to believe that statistics can tell us all we need to know about the world. But precision comes at a price: before you can count something, you have to define what it is you're counting, and often that's not as simple as it sounds.

Unemployment statistics, for example, conceal a host of decisions. How much can someone work and still be considered unemployed? How hard does a person have to be looking for a job? The Thatcher government changed the definition of "unemployed" either 23 or 27 times. ("There is some disagreement" about the precise number, the authors blandly write.)

Sampling is another headache. Most of the numbers we need involve populations too big to count one by one. As a result, we sample. But no matter how sophisticated the statistical techniques, they are still prone to error. Any American has only to think back to the polls during last year's primary season.

And that's with communicative human beings. Mr. Blastland and Mr. Dilnot describe Britain's efforts to count its hedgehog population (the National Hedgehog Survey) because of worries that this shy animal was in decline. They also discuss the Herculean -- but gravely important -- task of counting the fish in the sea.

"Uncertainty is a fact of life," they say, even if it's a given of human nature to look for meaning where there isn't any (see under: religion). They devote an entire chapter to chance to explain why the public sometimes sees a pattern where there is no such thing.

In 2003 the villagers of Wishaw, England, convinced that a recent rise in the incidence of cancer in their area was caused by a nearby cellphone tower, proceeded to lynch the tower, or, more accurately, pull it down in the dead of night.

What the villagers didn't know, the authors say, is that "cancer clusters" occur naturally, just as a coin tossed 30 times will probably produce at least one sequence of four heads or four tails. Tattoo this on your arm: a pattern doesn't always mean a plan. Throw some rice in the air and you will most likely see patterns in the way it lands.

Statisticians even have a name for the phenomenon: it's called the Texas Sharpshooter Fallacy. "The alleged sharpshooter," the authors write, "takes numerous shots at a barn (actually, he's a terrible shot -- that's why it's a fallacy), then draws his bull's-eye afterward, around the holes that cluster."

Actually, I read that first in a story about WWII, which, if I recall correctly, didn't take place in Texas. But, their book sounds pretty good and pretty necessary.