Cautionary Tales with Tim Harford - BONUS: Storks, Smoking and the Power of Doubt

Episode Date: September 16, 2020

It's easy to mock statistics or cast doubt on them... but we do so at our peril. Undermining our trust in facts and figures can cause great harm, and even death. We should guard against it.Tim Harford looks at how the seeds of doubt are planted in this mini-episode of Cautionary Tales to celebrate the release of his latest book.“How To Make The World Add Up” is out now in much of the world, while listeners in US/Canada can pre-order it under the title "The Data Detective" - ahead of its release in early 2021. Learn more about your ad-choices at https://www.iheartpodcastnetwork.comSee omnystudio.com/listener for privacy information.

Pushkin. Hi, Tim Halfert here with two pieces of news and a brand new mini-corsionary tale for you. News headline number one, we're back in the studio recording a new monster season of 14 episodes of cautionary tales. There will be wonderful actors, hilarious calamities and tragic catastrophes. Corsionary tales, season two, stay tuned. Newtide line number two. I have a brand new book coming. If you're listening in the US or Canada, it's out early next year and it's called The Data Detective. It's never too early to pre-order, so please consider doing that. But if you're listening anywhere else in the world, the book

is out right now, and it's called How to Make the World Add Up. On the surface, it's a book about statistics, but really, it's a book about how to think clearly about the world. And if you want to know what the book is really about, gather close. I'm Tim Haferd, and this is a special mini episode of Corsion Retails. of cautionary tales. You can make an estimate of how many children have been born into a dutch or Danish family by counting the stalks' nests on the roof of their house. In statistical terminology, it would be said that a positive correlation has been found to exist between these two things.

These are the words of Darrell Huff, the author of the most popular book about statistics ever written, How to Lie with Statistics. But of course, Storks do not actually deliver babies no matter how strong the correlation may be. Huff explains. Big Houses attract big and potentially big families, and big houses have more chimney pots on which stalks may nest. Of course, sometimes it seems you can prove anything with statistics, especially things that aren't true. Darrell Huff's little book is a masterclass in skepticism. He exposes mistakes, half-truths, and outright lies. Why, he explains?

The crooks already know these tricks. Honest men must learn them in self-defense. Darrell Huff made statistics seem like a game, or perhaps a stage magician's trick, all good fun, but never to be trusted. And I worry that we're starting to trust nobody. We're starting to believe that lying with statistics is all anyone ever does, but skepticism can easily curdle into cynicism and cynicism can be corrosive. After all, in 1954, the very same year that Huff published How to Lie with Statistics, two researchers, Richard Doll and Austin Bradford Hill, produced one of the first convincing

demonstrations that smoking cigarettes caused lung cancer. They couldn't have figured that out without statistics. Even to notice the dramatic increase in lung cancer rates in the mid-20th century required a statistical view of the world. So what Darrell Hough saw as a trick, which a doll and Austin Bradford Hill saw as a tool and a tool that has saved many millions of lives. Two visions of statistics had emerged at the same time and they very quickly came into conflict because as the evidence on smoking and cancer emerged the tobacco companies fought back. When their product was so dangerous, so demonstrably dangerous, how on earth could

they persuade people to keep smoking? In a stroke of twisted genius, they realised the answer. Scientific evidence was the enemy, so they had to give people reason to doubt that evidence, without attacking it head on. They chose to emphasise uncertainty, the fact that experts disagree, and to call for more research before anyone did anything hasty. They manufactured doubt. And it turns out that doubt is a really easy product to make. A couple of decades ago, two psychologists, Carrie Edwards and Edward Smith, conducted an experiment to test how people felt about all sorts of politically fraught questions, including abortion rights, smacking children, and the death penalty for under 16s. They invited people to produce as many arguments as they could in favour of or against the conclusion, given three minutes per topic.

Of course, people found it easier to produce arguments that backed up what they already believed. Lots of researchers have found that. But more strikingly, Edwards and Smith found that disbelief seemed to flow more fluidly than belief. It's much easier to argue against positions you dislike than in favour of positions you support. Doubt has a special kind of power, and it is the easiest thing in the world to suggest that you should doubt what the statistics are telling you.

This is why the simplest move in any argument today is to encourage people to believe nothing. Doubt is a powerful weapon. Now statistics are a natural target, but honest statistics need to be defended because while it's easy to lie with statistics, it's even easier to lie without them. And there are plenty of things in this world, from the link between smoking and cancer, to how to deal with the coronavirus epidemic, that we simply can't begin to understand without statistics. In a complex world, they are an essential tool. In the spring of 1965, a US Senate committee was pondering the life or death matter of whether to put a health warning on packets of cigarettes. An expert witness appeared,

and he wanted to throw doubt on the scientific evidence. So he turned to the topic of stalks and babies. It turns out he said that households with more children also have more stalks on their roofs. That old story about babies being delivered by stalks wasn't true the expert went on. Of course it wasn't. Correlation is not causation. And similarly, just because smoking was correlated with lung cancer didn't mean not for a moment that smoking

caused cancer. Do you honestly think there is a casual relationship between statistics linking smoking with disease, as there is about stalks, asked the committee chair. The expert witness replied that the two seemed to me the same. The witness's name was Darrell Huff. He'd been paid by the tobacco lobby to do what he did best. Weave together with the examples, some statistical savvy, and a certain amount of cynicism to cast doubt on the idea that cigarettes were dangerous. He was even working on a sequel to his masterpiece, although it was never published.

The sequel's name was How to Lie with Smoking Statistics. with smoking statistics. I understand that statistics can seem intimidating and they are often used deceitfully, but we mustn't give way to cynicism or feel helpless under the deluge of numbers. Understanding statistical claims doesn't require a deep technical background. If you're able to think, then you're able to think sensibly about statistical claims. Which is why I've written a book about how to think more clearly about numbers and about the world. If you're listening in the US or Canada, that book is called The Data Detective, and

it's out early next year. And if you're listening anywhere else, the book is called How to Make the World Add Up. And it's out now as a book, an ebook, and as an audiobook read by me. If you like what I do here on cautionary tales, I really think you can love the book. The world is an amazing place, and these days, whatever you're interested in, the chances are you'll understand it much better through the lens of statistics. Don't be cynical. Don't assume it's all a lie or a trick. Don't be afraid to pick up this statistical telescope and gaze around. This mini episode of Corson Retails was written and read by me Tim Harford. The producer was Ryan Dilly and the music and sound design were by Pascal Weiss. Corson Retails is a Thank you. you