Brooks on Books: The Numbers Game

Not to be confused with the excellent book of the same title about baseball statistics by Alan Schwarz, this "Numbers Game" is a brief but illuminating hardcover by two Brits who have spent a lot of time exposing the misuse and ignorance of numbers and statistics in over there, especially in the media and the public sector.This edition, just published recently, takes the original U.K. bestseller and updates it a bit, with some modifications for us Yanks.

I enjoyed this book, but let me analyze one number myself: The $22.00 cover price is too high for a smallish 200-pager, even a hardback. Wait for the paperback or be a cheapskate and borrow it from the library like, uh, some people.

The idea of a text that delves into math and stats might sound either dreary or intimidating to many readers. Fortunately, authors Michael Blastland and Andrew Dilnot make the material accessible with plain but not condescending language and pertinent real examples to illustrate their points. One thing that is a little odd about their approach: Many examples come from British data and situations, and though the writers provide comparisons to U.S. equivalents when they can, it is a bit disorienting sometimes. Reading this at times made me feel like I kind of understood that whole argument about the SAT being "culturally biased." I have to admit that trying to get a handle on the British health system or money system or whatever took an extra few seconds.

But I may be making the book sound like homework again. It's quite entertaining. Each chapter examines a different aspect of how numbers are misused and/or manipulated: Sampling, Percentages, Comparisons, Chance, and more. They talk about concepts like regression to the mean, or the tendency of really high statistical phenomena to come down to match the average. A lot of this stuff will be familiar to anyone interested in baseball statistical analysis or even anyone who didn't sleep through, say, Statistics 200 in college (Now, THAT might be a small number).

To give just one example, Blastland and Dilnot show how the press will report the risk of cancer, say, going up by X % if you eat bacon, and the reporting makes it look like an epidemic, but the numbers actually indicate 5 out of 100 men get colon cancer and if all of them ate a few pieces of bacon a day, the number would go up...to 6 out of 100. The risk of adding bacon to our diet is small (though, as is stated, doctors advise just staying away from bacon, and the authors aren't arguing that point), but the newspaper might say that it adds 20% chance of getting cancer. Without the larger context, percentages and other numbers are meaningless, and they can be (and are) manipulated to create dramatic "news."

It's tempting to throw your hands up and say, "What's the point?" and ignore stats and data altogether, but the author pointedly urge us NOT to do that. They emphasize the need to look for that larger context, to ask questions and analyze the numbers instead of just accepting the way they are often presented. They never trash numbers themselves, but rather the misuse of them, and their recommendation is not to ignore, but to look more closely.

"The Numbers Game" is a little slim for 22 bucks, but it's a useful effort for those interested in getting a basic handle on practical interpretation of numbers. Any of us who follow the news (hopefully a lot of us) can get a lot out of it.