Beware Statistics

You know the old saying about three kinds of lies, Lies, Dirty Rotten Lies, and Statistics. (Okay, the second one uses a four-letter word). Statistics are tricky (or dangerous) because you can truthfully describe something and still mislead. In high school I read a book, How to Lie with Statistics  by Darrell Huff. It’s a pretty good intro to the topic. I ran into an article the other day that can illustrate the trickiness of statistics two ways. First the passage in question:

The researchers found that compared to participants who watched TV less than 2.5 hours each day, deaths from a pulmonary embolism increased by 70 percent among those who watched TV from 2.5 to 4.9 hours; by 40 percent for each additional 2 hours of daily TV watching; and 2.5 times among those who watched TV 5 or more hours.

Sounds dangerous, doesn’t it? A 70% increase! Here’s the misleading part: What’s the percentage of getting the embolism in the first place? It turns out to be roughly one in a thousand. So a 70% increase of that is still less than two in a thousand. Not so bad now, eh? The trickiness is that

even a large increase of a small amount is still a small amount.

We’re not finished. Let’s look at absolute numbers instead of percentages. The population of the US is about 319 million. That makes that one in a thousand to be roughly 319,000. Add 70% to that and it’s more than half a million! How would you like to pay the medical bills for half a million people? The principle here is that

if the numbers are large enough, even a small percentage is still pretty big.

For an engaging book that goes into these and other principles for understanding statistics, I suggest How not to be Wrong: The Power of Mathematical Thinking by Jordan Ellenberg.

So what does all this have to do with expository writing?

When you write, make sure you don’t mislead your readers.