Proofiness - Charles Seife [120]
5
After the deception was exposed by comedian Jon Stewart on The Daily Show, Sean Hannity promptly admitted that he had made an “error”: “We screwed up. It was an inadvertent mistake, but it was a mistake nonetheless.” Then, smirking, he thanked Stewart for watching the show.
6
And 36 percent of readers will actually believe that statistic.
7
Assuming that the “truth” exists, which isn’t always so obvious. These are atomic-scale distances, and the concept of “length” stops having a fixed meaning when you’re looking at measurements this precise. The atoms in the pencil are constantly moving, so the pencil is expanding and contracting—so the pencil no longer really has a given length for more than an instant. After a certain point, nature conspires to keep you from finding out the “true” length of a pencil by making the truth itself rather fuzzy. But the general point still holds: there’s always a level of error in any measurement that obscures reality.
8
We’ll get to the pollster’s favorite slippery convenience, margin of error, in chapter 4.
9
A great deal of scientific effort goes into increasing the precision of measurements to get closer to the truth by adding an extra decimal place or two. For example, in the past two decades, scientists’ estimate of the age of the universe went from “about 15 billion years” to “14 billion years” to “13.7 billion years.” This seemingly subtle change represents an extraordinary—almost revolutionary—advance in our knowledge about the cosmos.
10
It’s also very common in the scientific world, thanks to a phenomenon known as “publication bias.” Peer-reviewed journals cherry-pick the most exciting papers, selecting them for publication. This means that papers with spectacular results are published in high-profile journals while less sexy ones (including negative results) are relegated to lesser journals or aren’t published at all. Publication bias distorts science, making new drugs, for example, seem more effective than they actually are.
11
Despite my singling out Al Gore for cherry-picking, there are unambiguous data that show that global warming is occurring. It’s just that sea levels aren’t going to rise twenty feet anytime soon.
12
There’s a third kind of cherry-pick here, in fact. Even if you accept that math and reading scores are important, schools teach a lot more subjects: writing, science, history, and more. Data from these disciplines show either mild improvement or, in some cases, decline, particularly in upper grades. Concentrating on reading and (particularly) math as indicators of a school’s improvement is only looking at part of the picture.
13
Budgets are always subject to proofiness. The people making them have a vested interest in making expenditures seem tiny; those criticizing them are trying to make those same numbers seem large. Though the details are too wonky for this book, it’s worth mentioning that in the United States, the government keeps certain expenditures (like Social Security spending) off the official budget, counting them separately. Because of this, politicians can cherry-pick, ignoring off-budget expenditures when it suits them.
14
In cases like this, it’s often better to use a construct known as the median to figure out what a typical salary should be. To calculate the median, you line the numbers up from lowest to highest and pick the one in the middle. Here, the median salary would be $1, clearly a better representation of “typical” than the mean would be.
15
The other half prevent cancer, of course. As columnist Ben Goldacre once put it, newspapers “will continue with the Sisyphean task of dividing all the inanimate objects in the world into the ones that either cause or cure cancer.” Unfortunately, this is a fruitless endeavor, as some compounds, such as red wine, both cure and cause cancer, at least if you trust the headlines.
16
Sharp-eyed