Tim Harford: “If You Can Make Sure You're Not An Idiot, You've Done Well.” - Freakonomics
"Nightingale, I should say, she wasn’t a professional statistician... she’s a woman in a man’s world. She’s trying to challenge the British medical establishment and the British military establishment. She’s telling them they’re doing it all wrong. She was basically saying, “We need massive public health reform because the hospitals are breeding grounds for disease.” And the chief medical officer at the time, a guy called John Simon, was saying, “Well yeah, disease is bad, but there’s nothing we can do about it.”...
LEVITT: What I hear you saying is that there are two ways in which arguments can go wrong. One is that the facts can be off. And the other is that the interpretation made, given a set of facts, can be wrong. I think that if we divide problems into those two pieces, everything becomes much simpler. If people don’t agree on facts, then we should go and evaluate the facts and figure out what the facts are. If they don’t agree on the interpretation, then I think that is a much easier problem for the human brain to tackle than the problem which is: how do I take storytelling and facts and everything all mixed together and try to parse out the importance? And I think that’s actually in the background in your book, that’s really lurking in your book and your own thinking...
HARFORD: I talk about the illusion of explanatory depth, which I love. The illusion of explanatory depth is basically: if you ask people, “How well do you understand how a zipper works on a scale of, say, zero to seven?” Most people will say, “Yeah, six, I understand it pretty well.” And then, you say, “Oh, great, here’s a pen, here’s some paper. Use diagrams, bullet points, whatever. Just explain to me exactly how it does work.” And then, they realize, actually, I don’t really know how it works. The illusion of explanatory depth says just asking people to lay out the facts may help them to understand that maybe they don’t actually know the facts. Maybe they don’t understand the thing that they’re arguing about. It turns out that if you use a similar tactic for say policy choices — so, you say, “Just explain to me how a cap-and-trade system would work.” People who are willing to die in a ditch over whether cap and trade is a good response to climate change or not, it turns out they don’t really know how it works. When you ask them to explain it, they start to realize, “Oh, I don’t completely understand this. Maybe I should moderate my political views. Maybe I shouldn’t be so critical of people who disagree.” This process of laying out the facts, which I think is worthwhile in and of itself, there’s this bonus which it actually gets people to reflect and be a bit more humble about the limits to their own knowledge.
LEVITT: The other piece that I think is really important for lay people understanding data that I didn’t see you cover in the book that I want to mention is thinking hard about the incentives of the people who are putting forth the argument. And being suspicious of any argument in which the incentives are such that the creators of the argument could benefit in any way...
HARFORD: I feel that people have received that message over and over again. I think people are constantly being told to be suspicious of the motives of the people who are telling them things. And I think we may have gone too far because although it’s true, I think it’s bred a lot of cynicism. A lot of people worry that we believe anything. And actually, what I worry about is that we believe nothing at all, that we’re completely skeptical of everything and we just think, “Well, they’re all lying to us. It’s all fake news.” So, that’s what was very much on my mind in writing the book. And actually, it’s interesting because Freakonomics is a book that doesn’t make that mistake. So, Freakonomics right from the start, is a book that says, “Hey, let me tell you something really interesting about the world using this data.” But most books about data actually don’t do that. Most of the books about data that I’ve got on my shelf are written by eminent economists, statisticians, explaining all the different ways in which data can be used to lie to you. And of course, it’s a really engaging way to talk about data, but there is this worry that I have that people hear that message over and over and over again. And in the end, it becomes an excuse to just go, “I can’t believe any of these people. I don’t trust any of the experts. I’ll just believe whatever my gut tells me, whatever I feel should be true. And I’m not going to look at any evidence because you can’t believe any of it.”...
LEVITT: The National Rifle Association has been extremely successful at limiting the collection of data around guns, and that has really hamstrung the academic research into it. In fact, one of the most clever papers ever done on guns was done by my good friend, Mark Duggan. He was simply trying to figure out how he could determine how many guns were in different places. He had the incredibly clever idea to go to a different data source, which is magazine data. So, there’s enormously carefully collected data on magazine circulation because that’s how advertising payments are done. He used purchase of handgun magazines as a proxy for purchase of handguns. And what was very difficult and clever about the paper is he actually showed that over time, the changes in the number of guns correlated very, very highly with his measures of magazine subscriptions. He used that as a proxy and actually was able to say interesting things about guns. Things that can’t be measured, it’s very difficult to regulate or control them. I think the N.R.A. has understood that for a long time. And they’ve been very, very effective at making sure that guns can’t be measured. If you think about the economy, imagine that we couldn’t measure incomes or we couldn’t measure G.D.P. That’s the equivalent when it comes to guns. We just don’t know how many guns there are in different places and how that changes over time. And so, it’s really hard to study the problem and certainly extremely hard to get a causality...
HARFORD: The story begins in 1975 when this German teenager called Vera Brandes walks out on the stage of the Cologne Opera House bursting with excitement because, a few hours later, Keith Jarrett is going to be on that stage improvising. He’s a great jazz musician. He’s going to be sitting at this piano. And he’s going to be just playing whatever comes into his head. And all this has come about because Vera is the youngest jazz promoter in Germany. She’s 17-years-old. She just loves jazz. And she’s managed to score this amazing coup of getting Jarrett into the Opera House to play this late-night concert. When Jarrett actually comes on the stage to check out the piano, immediately, it becomes clear that something has gone wrong and there’s been a mix up. They’ve brought out a rehearsal model. The keys are sticky. The pedals don’t work. It’s too small. It sounds tinny. It’s just a bad piano. And Jarrett says, “Well, I’m not going to play.” But it turns out there’s no way of getting a replacement piano on the stage in time; it’s not possible. The tickets can’t be refunded because of the way the concert’s been set up. This teenage kid is about to be ripped apart by 1,400 people who show up for a concert and there’s no concert.
And so, Jarrett takes pity on her. And although he’s a real perfectionist, although he likes things exactly the way he likes them, although he feels the piano is completely unplayable, he just thinks, I’ve got to do it because I’ve got to help this girl out. He, a few hours later, walks out on stage, sits down at this piano that he knows is unplayable, and begins to play. And instead of the musical catastrophe that he expects, it’s a masterpiece. The concert was recorded supposedly to provide documentary evidence of what a musical catastrophe sounds like. But in fact, once it was remixed, it sounded great. Many people think it’s his best work. It’s easily his most successful work. So, the concert has been released as The Köln Concert, best-selling jazz piano album in history. And it only got played because Jarrett felt he’d been backed into a corner and he couldn’t let this girl down. He thought, “This is terrible. It’s a bad piano. It’s going to be a bad concert.” But he was, of course, forced to play in a different way and to improvise in a different way. So, he stuck to the middle of the keyboard, which made it sound very soothing and ambient because the upper register sounded terrible. Because it was such a small instrument, it was quiet. So, he was pounding down on the keys to try to create more volume. So, there’s this weird tension. He was playing this nice ambient-y music, but he was really hammering it hard and playing with a lot of energy. And there’s just something about that that worked really well. It’s how I begin my book, Messy. The book that’s really all about how disruption and challenges and weird stuff that’s ambiguous and messes around with us can actually lead to a problem-solving response...
A strike on the London Underground that shut down half of the London Underground for 48 hours. And when researchers looked at the data, they found that tens of thousands of commuters had changed their route because of the closures. And then, at the end of the 48 hours, they never changed back. They discovered a better way to get to work and all it took was this perturbation to the system...
LEVITT: So, I’m really interested in the subject of persuasion... you also have insights for everyday people about how to make an argument persuasive...
HARFORD: I would go for a memorable story. Stories are not so threatening. You’re not attacking anybody. You’re giving them something they’ll immediately find interesting. And they’ll follow the story along and they’ll be curious, and it starts to open their minds. So, if you’re talking to people in terms of stories, you’re lowering their instinctive psychological defenses that basically say, “This guy is challenging my sacred beliefs. And I’m champion of all that’s right.” Stories get people into a more open-minded frame of mind...
LEVITT: Stories have beginnings and middles and ends. So, what’s interesting is I used to teach a lecture in my course on data to the undergrads where I talked about storytelling with data. And it always left me feeling a little bit strange because it wasn’t very powerful. And then, one year, I just sat back, and I thought about it. And I thought, “Wait a second, if a good story has all of the elements we just talked about, analysis of data never leads to a good story. There’s almost never a person involved that you can identify. There’s almost never any intrigue or uncertainty. There’s not a twist at the end.” And I completely redid the lecture and now, the way I start the lecture is by telling a great story. And everyone laughs. They think it’s a great story. And I talk about what made it a great story. And then, I say, “Let’s take some examples with data and how we would turn them into stories.” And it becomes really clear to everyone that you cannot tell great stories with data. And I really came to the conclusion that when it comes to data, you just should completely abandon the idea of telling stories, that you should use data and just explain the truth."
Too bad the British medical establishment didn't "trust the experts"
In my experience, asking people to explain just makes them get upset since they realise they have no idea what they're talking about, and then they block you
People only care about conflict of interest when it threatens their views
