Statistics: Be Afraid. Be Very Afraid.
David Spiegelhalter, a distinguished British statistician and the Winton Professor of the Public Understanding of Risk at the University of Cambridge, is a fan of numbers. But, like a carton of eggs, they should be handled with care. Unfortunately, the media often splatters them all over the place.
Take this headline for example:
Eating bacon every day boosts the risk by cancer 20 percent? Sounds alarming, doesn't it? Now, let's look at the statistics behind that claim in a slightly different way.
The study in question examined a particular type of cancer: pancreatic. Pancreatic cancer is quite deadly, but thankfully, it's rare. Approximately five out of every 400 people will develop it. Now, let's apply that 20% increase...
“If all of those 400 all stuff their gob every morning with a great big greasy bacon sandwich, that five would increase by 20% which means going up to six,” Spiegelhalter explained to the BBC's The Infinite Monkey Cage. “Told in that way, it’s a complete non-story.”
The framing of a story about risk, health, and death makes a big difference as to how many people will read it – the scarier the better, generally. Accuracy often suffers in the process.
Let's look at another headline, this one from February:
Yowza. Stay inside, stock up on ice cream, and blast the air-conditioning!
But lost in the rush to predict dire consequences from global warming, were two key nuggets of information: 1. There will be more deaths anyway as the population of the United Kingdom grows older. 2. The reduction in cold-related deaths actually dwarfs the rise in heat-related deaths!
Take a look at the original graphs from the research showing off heat deaths and cold deaths. Notice anything misleading? (Hint: take a look at the vertical axes.)
The axes on both plots had been arranged to make the numbers look more comparable than they really are!
On his blog, Spiegelhalter remade the graphs and combined them to more accurately reflect the data.
Looking at these graphs leads one to a different conclusion. "Overall, for any individual in the UK, the risk of a temperature-related death is expected to fall steadily due to climate change," Spiegelhalter wrote.
Anytime you read an alarming headline where assessments of risk are involved, be skeptical. There could be a more nuanced story buried in the details.