Freakonomics Radio Live: “We Thought of a Way to Manipulate Your Perception of Time.”
We learn how to be less impatient, how to tell fake news from real, and the simple trick that nurses used to make better predictions than doctors. Journalist Manoush Zomorodi co-hosts; our real-time fact-checker is the author and humorist A.J. Jacobs.
* * *
This is a bonus episode of Freakonomics Radio Live. It’s the non-fiction game show we call Tell Me Something I Don’t Know. This was recently recorded in New York. If you’d like to attend a future show or be on a future show, visit freakonomics.com/live. We’ll be back in New York on March 8th and 9th, at City Winery; and in May, we’re coming to California: in San Francisco on May 16, at the Nourse Theater, in partnership with KQED; and in Los Angeles on May 18th, at the Ace Hotel Theater, in partnership with KCRW. And now, on with our show.
Stephen DUBNER: Good evening, this is Freakonomics Radio Live! Tonight we’re at Joe’s Pub in New York City and joining me as co-host is Manoush Zomorodi. Manoush is the host and creator of the podcasts ZigZag and Note To Self. She’s the author of the book Bored and Brilliant: How Spacing Out Can Unlock Your Most Productive and Creative Self. Manoush, we know you grew up in Princeton, N.J., the child of not one but two psychiatrists.
Manoush ZOMORODI: Indeed. Everybody went like this: “Aww.”
DUBNER: Well, afterwards they’ll come to you with their problems, presumably. We know that, before getting into the cutting edge world of podcasting, that you reported for legacy media companies including Thomson Reuters and the BBC. So tell us something we don’t yet know about you, Manoush.
ZOMORODI: My big break was I was a breaking news producer for the BBC and I was sent with a correspondent, a real grownup reporter person to report from Mount Etna, which was erupting. The volcano is going off, that was pretty cool. We’re on TV. And then he’s like, “Right, see you later. I’m going back to Rome for my child’s birthday party.” And I was like, “All right, bye.” I couldn’t get a flight out until the next day.
DUBNER: Was that a Roman accent?
ZOMORODI: That was Brian Barron actually. So I went to sleep in my lovely hotel. So they wake me up at four in the morning, and they are like, “The volcano is erupting again.” And I was like, “Yeah, but Brian left.” And they were like, “You, go to the volcano and report.” So I was on the morning news reporting from an erupting volcano and never looked back. 2001, been a reporter ever since.
DUBNER: Congratulations. We are so excited that there was danger happening and Manoush got to be there —
ZOMORODI: Have you ever seen lava flow, really? It goes like this.
DUBNER: Just so you know, this is radio. I’d like to describe what Manoush was doing. She was holding up her hand and moving it very slowly. Did it change your life in any way other than career-wise?
ZOMORODI: Yeah, because I thought of my own capabilities completely differently. So that changed everything.
DUBNER: Okay, Manoush, very, very happy to have you here tonight.
ZOMORODI: Thank you.
DUBNER: Thank you for coming to play Tell Me Something I Don’t Know with us. Here’s how it works: Guests will come onstage to tell us some interesting fact, or idea, or story, maybe a historical wrinkle we don’t know. You and I can then ask them anything we want. And at the end of the show our live audience will pick a winner. They will vote on three simple criteria: No. 1, did the guest tell us something we truly did not know? No. 2, was it worth knowing? And No. 3, was it demonstrably true? And to help with that demonstrably-true part, would you please welcome our real-time fact-checker, the author of four New York Times best-sellers and counting, including The Year of Living Biblically, A.J. Jacobs.
A.J. JACOBS: Thank you Stephen.
DUBNER: So A.J., it’s been a while since we did one of these shows together. I assume you’ve just been sitting at home waiting for us to call. Have you been working on anything at all?
JACOBS: Well that and I was able to squeeze in— I have a new book. It’s called Thanks A Thousand, and the idea is I went around the world and thanked a thousand people who had even the smallest role in making my morning cup of coffee possible. So I thanked the farmer who grew the coffee beans, and the trucker, and the logo designer, and the man who made the zarf.
DUBNER: What’s a zarf?
ZOMORODI: Bless you.
JACOBS: Well, thank you. A zarf is, I learned, the official name for that little cardboard sleeve that goes around a coffee cup.
ZOMORODI: No way. Come on, that’s not the word for it. Zarf?
JACOBS: Zarf. Yes. And, just so you know, it has a long and glorious history. There were zarfs in ancient China made of gold and tortoise shell so—
ZOMORODI: Wait, A.J. wins. I’m sorry. That fact was amazing. That’s so cool.
DUBNER: Do we know where the word comes— Is it the sound the first person made when they grabbed a cup of hot coffee without a zarf?
JACOBS: That is that is a good question. If you give me 30 seconds, I can give you an answer.
DUBNER: We’ll get back to it, by the end of the show you can tell us the etymology of zarf. A.J., delighted that you are joining us as well. It’s time now to play Tell Me Something I Don’t Know. Would you please welcome our first guest, Julie Arslanoglu. Julie, welcome. I understand you are a research scientist at the Metropolitan Museum of Art. I’m guessing that’s pretty fascinating work and very promising for our purposes tonight. So, I’m ready, as are Manoush Zomorodi and A.J. Jacobs. What do you know that’s worth knowing that you think we don’t know?
Julie ARSLANOGLU: I have a simple question: What do antibodies have to do with art and art conservation?
DUBNER: Antibodies being the protein in blood that attacks the bad guys?
ARSLANOGLU: It’s the protein that your body produces to recognize an other. So every living organism has these.
ZOMORODI: Is it something that you apply?
ARSLANOGLU: We apply antibodies to art. Yes.
DUBNER: Oh, you apply antibodies. You’re not looking for antibodies in the art.
ARSLANOGLU: No we’re not.
DUBNER: Let’s ask about you. Is your background purely art, and/or art conservation, or do you have some kind of bio-chem background?
ARSLANOGLU: I have an organic chemistry graduate degree.
ZOMORODI: So are you potentially part of the preservations staff at the Met?
ARSLANOGLU: I’m a research scientist within the conservation department.
DUBNER: Does the Met have a lot of you, or are you flying solo?
ARSLANOGLU: No, there’s 12 of us.
ZOMORODI: There’s 12 of you?
DUBNER: Where do the antibodies come from that you apply?
ARSLANOGLU: We purchase them from commercial sources.
DUBNER: What do they belong to? Where do they come from?
ARSLANOGLU: Well. The way an antibody is created is, you take something that you want to study, some protein from an organism. It can even be a small molecule. You inject it into an animal that’s called the host. The host creates antibodies against the ‘other.’ You harvest those, and then you inject those into a second animal and create antibodies to that first antibody and use that as your reporting system. So the whole idea is that when you have one thing that you want to recognize, you create an antibody to that.
DUBNER: That is super fascinating. So this is some version of antibody dating or— I don’t mean like smooching, dating. I mean like carbon dating, dating.
ARSLANOGLU: It could be timed theoretically, but it’s really complex, because you have a couple of problems. One is that normally antibodies are made against what are called native proteins. So these are the proteins that come off freshly from an organism. So let’s talk about collagen. If you extract collagen from let’s say bovine skin, you can create an antibody for collagen, and you can create an antibody specifically for bovine skin. But these are going to be in their native state, meaning it’s the way the protein is extracted from the tissue or the organism.
My issue is that the proteins that are extracted, are extracted to be prepared to use for artwork. So if it’s going to be a glue or if it’s going to be a binder for a paint, meaning a paint is usually ground-up minerals and you have some sort of adhesive that holds the whole thing together. So, if you prepare this material, you are going to heat it to extract it. You’re going to do something to it to prepare it for an artistic way of using it. Then you’re going to mix it with these inorganic mineral pigments which have cations that react with proteins.
DUBNER: Ions from cats?
ARSLANOGLU: Cations means positive charged.
ARSLANOGLU: And then you’re going to let this stuff dry. So a protein normally lives in an aqueous environment. Now you are going to remove the water and then you’re going to let it age for 500 years, a thousand years. So stuff happens to those proteins. Art is made up of materials, materials continue to react. The way they react is really complex because we don’t have a really clear knowledge of the conditions that art was exposed to. And art continues to be treated over time. So if it enters a museum, for example, it might be consolidated with additional animal products, like sturgeon glue, or animal glue. It might have a synthetic polymer added to it to create a more cohesive surface. So you have all these things mixed together, they continue to react.
DUBNER: So you’re doing this kind of proactive-ish, historical-ish, detective work for what purpose? For restoration, for proving the provenance or history of something? What?
ARSLANOGLU: So, at the most basic level what we’re trying to identify is: what materials are used to create the artwork. And why this is important is a few things. One is, you have a sort of a lexicon of how art was created. So for example, egg tempera being used in the Italian Renaissance. Well, was every painting in the Italian Renaissance made with egg tempera, or did they use other protein based binders? So you can inform that lexicon, you can create a timeline that informs the art historians about what the materials actually were being used.
But more than that, when you look at a piece of artwork, it’s a combination of all the chemistry that’s going on. So when you mix a binder like a protein with a mineral pigment, as the light passes through it, you’re going to get a certain amount of saturation of colour. You can get some— An amount of gloss, and it all depends on that combination of which pigment in which binder. And if you add an oil to it, it changes everything. So looking at an artwork, understanding what you see, why does it look the way it does right now? And did it look that way originally, or has it changed? So that’s one of the most basic reasons for doing that.
ZOMORODI: Oh my God. If you had been around in high school, I would have actually liked biochemistry. I was the artsy kid who was like, “I don’t understand.” And I got a C- in chemistry you guys, it was my worst grade. But if you had —
DUBNER: You got no sympathy for that, I think what that means is that most of them got worse than C-.
ZOMORODI: Maybe. I found it really difficult, but also because there was no applicable usage. Like what you just described just lit up my brain and explained to me why I like certain paintings and others, and that’s amazing! That’s so cool.
ARSLANOGLU: There is a really strong connection between art and science. So the way it looks, why it changes, the mechanics of the film. As it changes and gets older, it actually increases in stiffness and you get cracking. And you can explain all this with chemistry and engineering. This is a really strong connection with STEM and STEAM, there are universities that really pull us into their chemistry classes to teach folks who aren’t quite so keen on chemistry.
DUBNER: Are there any applications of this method or something similar beyond artwork? Is it used in archaeology, etc., etc.?
ARSLANOGLU: Absolutely. One of the earliest uses of antibodies was actually paleontology. They used it to identify the collagen in dinosaur bones.
DUBNER: This was how long ago?
ARSLANOGLU: This is like late 70’s, early 80’s.
ZOMORODI: Is this process used in order to, not only understand the ingredients that have been used to make the art, but potentially to preserve it in some way? Or clean it, I guess.
ARSLANOGLU: Well, we’re trying to use a combination of the antibodies and something called mass spectrometry to look at the molecular structure of the protein. So what we found out is that when you mix different protein binders — like collagen, or whole egg, or milk products — casein — with different pigments, you actually will change the conformation of the protein. These combinations change the conformation of the protein. So you’re getting some sort of structural change in the protein, and we’re trying to look to see how that affects the longevity of the paint.
DUBNER: A.J. Jacobs, Julie Arslanoglu from the Metropolitan Museum of Art has been telling us about antibodies and the use thereof in restoration and learning more about art for many applications. I’m sure you know an awful lot about this.
JACOBS: Of course. Even before I came. Yeah. Well, it all checks out. Julie gets an A+ for accuracy, not a C+. No offense.
JACOBS: Cheap shot. But actually it led me to this list of the strangest ingredients contained in paint. There was Indian yellow, apologies to those eating right now, but according to legend this was made from cow urine. And not just regular cow urine. This was cows who were fed only mango leaves which makes for a gorgeous pee, apparently. And Scheele’s Green, which was also lovely, but poisonous. It was made from arsenic, and according to legend it’s what killed Napoleon Bonaparte. He had green walls in his room. So there you go. Cow urine and arsenic. \That’s art.
DUBNER: Well A.J., thank you and Julie thanks so much for coming to play Tell Me Something I Don’t Know. Would you please welcome our next guest David Reitter. So, David Reitter it says here you’re a professor of Information Sciences and Technology at Penn State and that your research has been particularly focused on what makes us intelligent — those of us who may or may not be — and why we make mistakes. So that sounds great turf for us. David, tell us something we don’t know please.
David REITTER: So I live in this college town, and I guess it’s a bit divided, which you see when you drive around there. There’s the people that lived there all their lives, and these townies have all the time in the world. And then there’s people like me who are a little impatient, and we’d like to drive our fast cars to where we need to go, right? How can you make me a little more patient?
DUBNER: We could start by not calling them townies, because I don’t think they like that. But that’s just a hunch. You’re saying that you and your uber-educated class of people, you got a lot to do. You’re rushing to do research and to give aid to floundering students like Manoush who are getting C-‘s.
ZOMORODI: So are you doing research into ways to get people to not be like you?
REITTER: Yeah. I think there’s a lot of work that’s been done in behavioral economics that has found out that we’re all a little impatient, right? So I’ll give you an example: Do you like chocolate?
ZOMORODI: I love chocolate. Can you please just tell me what you’re going to — I’m sorry, just kidding. That was my demonstrating my impatience.
REITTER: I’ll run a little experiment to you, okay? Would you like two pieces of chocolate or one?
ZOMORODI: Of course two.
REITTER: All right. Now I’ll attach a little bit of time to that. Would you like two pieces of chocolate in a week from now—
ZOMORODI: Oh are you doing the marshmallow test on me? It’s that what you are doing on me?
REITTER: That’s exactly right.
ZOMORODI: I’m a tech reporter, so I know about this stuff, let me tell you, alright? We can’t wait for anything any more, because we have instant gratification.
REITTER: So in a way we all know that we’re impatient. Right? The question is, can you do something about it?
DUBNER: So I have to say, even though the way you presented your dilemma you sounded a little — I don’t want to say arrogant, but like, “I have a problem with the slow people.” Right? That’s what you said. But it’s interesting that now when you’re searching for a solution, your solution is not to make the slow people faster, you do want to ameliorate your impatience.
ZOMORODI: That’s a good point.
DUBNER: You are identifying that. So you want to know how you could become more patient while driving. Have you tried listening to a fine podcast?
ZOMORODI: I think that’s a really good idea. Or smoking weed. Have you ever been behind someone who is like, “Whoa, we’re going so fast, dude.” And it’s like 20 miles an hour. That was also high school, could explain the C-. Just going there.
REITTER: All of these would work.
ZOMORODI: Okay. But you’re trying to look at it more of a cognitive behavioral, sort of talk yourself into being more patient thing? Because that’s what you do?
REITTER: It’s not talk therapy.
DUBNER: Is it a technological intervention of some sort?
REITTER: It is.
DUBNER: Oh, I once learned that the best way to make people on a train feel that the trip is shorter is just by putting in really good fast WiFi. So if you do that in your car then you can watch Netflix while driving.
ZOMORODI: Okay. What if it’s autonomous vehicles and therefore no one’s driving.
REITTER: How boring.
ZOMORODI: I mean the vehicles are all just driving themselves and they’re all at the same speed.
DUBNER: Then you can smoke dope and watch Netflix and it’s not a problem.
ZOMORODI: Exactly, amazing.
DUBNER: Why don’t you tell us what you did? And I’m curious to know, is this an experiment that you did in the field or in a lab? Because we care about those distinctions.
REITTER: Neither. We run experiments on thousands of people that are somewhere in the world. So our research is not based on American undergraduate psychology students but people from all over the planet of different ages that like to do our experiments.
ZOMORODI: Who live in towns where there are half the people who are really snooty and think that they’re really busy and have to go somewhere really fast and the rest of the people are like, “F— you, we just are trying to get to the grocery store.” Like that?
REITTER: Exactly. We thought of a way to manipulate your perception of time by giving you something that you might already know. A countdown. Like the countdown you see on old time movies before the movie actually starts. Or a progress bar when something is really really slow on a computer. So we show people a countdown and then we give people a test of their impatience. Now we manipulate the countdown. We were interested in what happens when that countdown is fast versus when that count down is really slow. The countdowns always take the same amount of time, 15 seconds. But I count down “15, 14, 13, 12,” or I can countdown like this, “5, 4, 3—” Are you impatient yet?
DUBNER: Yes, very. So you’re saying that if the numbers are going faster, even if the duration of time is identical, we experience it faster.
REITTER: That is correct. We’re happier with the whole game. But most importantly we make better decisions during the impatience test that follows.
DUBNER: What is the decision-making tests that you use in this experiment?
REITTER: Well so, this is kind of fun. My collaborator Moojan Ghafurian came up with this beautiful experiment, where we bring in Cookie Monster. Now Cookie Monster is probably the most impatient guy we know.
ZOMORODI: Now you have my attention.
REITTER: And your job is to host your friend, the Cookie Monster. And you’ve got a jar of cookies sitting in your living room and Cookie Monster. Now the question is, for you, how often do you check on Cookie Monster to make sure he hasn’t started eating cookies yet, or you catch him the right moment when he starts eating cookies. So if you do it right, you only check once, right after he starts eating cookies. And that’s really really hard. And we found that the people that saw the slow countdown, checked earlier and more often. The people that were more impatient made worse decisions. And that was in the time that followed watching the countdown, not during the countdown.
DUBNER: So let me just see if I understand. If you were to make a prescription would you say that, for instance, traffic lights should have attached to them a countdown clock from, whatever it is, 30, going fast? Is that the idea?
REITTER: They should be going fast, and they should even be speeding up. We found that—.
DUBNER: But you’re talking about literally mounting a countdown clock where it’s visible at a traffic light or an intersection or whatnot, is that the idea?
REITTER: And for pedestrians that’s already being done.
DUBNER: Well we have that in New York, fairly recently. Is your research connected in any way to that? Our walk and don’t walk signs they used to be the hand.
DUBNER: Now you get a countdown clock and I think it starts with like 30 and goes really fast. And I see old people running. Which to me seems potentially counterproductive. I don’t know. But do you know anything about that and whether it’s working, safer, etc?
REITTER: My guess it will be safer. I don’t think it’s meant to manage people’s impatience in that sense. It manages people’s timing. So they don’t block the intersection, anything like that.
DUBNER: Is impatience necessarily, however, a trait to be lessened or dampened? Because I would consider myself a fairly impatient person, which I know has its downsides in some cases, but I think there are also upsides. You quit things faster when they’re not working out, which that may not suit everyone, but there are those who argue that that can be a good thing. And I’m curious whether impatience is actually the thing that you’re fighting or was that just a viable mechanism to try to figure out how to manipulate people’s perception of the events?
REITTER: When Etna is breaking out, do you really want to be patient?
DUBNER: He played the volcano card.
REITTER: We’re truly interested in how we can change people’s perception of time and how we can affect people’s decision making. And you can use these countdowns in both directions. You can make somebody more patient or you can make them more impatient.
DUBNER: Oh, that could be handy.
ZOMORODI: So, when you’re doing this research, you couched it in the fact that this is happening in your town. But are there better use-case scenarios were you are trying to fix that problem in particular? Or were there other problems that you were actually trying to sort out?
REITTER: So our experiment that we designed is meant to be very much like many decisions that we take in real life. And these are decisions such as how often do I inspect a crumbling bridge? Or when do I decide to renovate it? Because every week I don’t renovate the bridge, I get more use out of it. Right? Or similarly, a police precinct deciding how often to patrol an area. Or simply, again, you’re driving and you have to make quick decisions on how to gather information about the things around you. Has that cyclist moved, or is he still in my blind spot?
So making all of these decisions is really something that’s very, very commonplace. Timing decisions are very, very important to managing risk, managing our safety. So of course this is applicable in the context of driving as well. And if you can’t put up countdowns on your traffic lights, perhaps you could listen to some fast music before you get in the car, you’d listen to some slow music while you’re going. Or a slow podcast.
DUBNER: There are no slow podcasts, only slow podcast listeners. A.J. Jacobs, David Reitter has been telling us about how to essentially manipulate away our impatience, which is fascinating. What more can you tell us on that?
JACOBS: Well, I’ll just get right to it. It does check out. Actually I was losing the train of thought a little, so I—
DUBNER: Smoked some dope.
JACOBS: No, my kids are in the audience. But I did research what is the longest traffic light in America, according to The New York Times. It’s in New Jersey, an impressive five minutes and 28 seconds.
ZOMORODI: Where? Where in New Jersey?
JACOBS: West Milford, New Jersey.
DUBNER: Is anyone surprised the longest traffic light is in New Jersey? Let’s be honest. But wait a minute, that can’t be right.
JACOBS: That’s what it says. This is the paper of record. You can listen to the Gettysburg Address almost three times in five minutes. So that’s a good use of your time.
ZOMORODI: So what I want to do is, I want to apply what you’re saying to Twitter. Instead of people reflexively retweeting or responding with outrage, what if they were like, “Countdown, here we go. I need 15 seconds before I can respond.” Do you think that would work to make people stop tweeting stupid s— basically?
JACOBS: I love that.
REITTER: I love that.
DUBNER: Hey David Reitter, thank you so much for coming to play Tell Me Something I Don’t Know. And would you please welcome our next guest, Jeff Nosanov. Jeff is a consultant for NASA. He formerly worked at NASA’s Jet Propulsion Lab, the coolest spot in the NASA universe. He’s also got a law degree. Yeah, no clapping for that. But check this out. He’s also got a very rare master’s degree in space and telecommunications law.
Jeff NOSANOV: That’s right.
DUBNER: Which is odd and nifty.
NOSANOV: It is. It is. It’s been a strange journey.
DUBNER: So Jeff, welcome to our stage. Thank you so much for coming. What do you have for us tonight?
NOSANOV: Well my question to you is what is the most useful mission that NASA has done?
DUBNER: Well I would say, if the moon landing had not been faked, that would have been it.
NOSANOV: You got us, you got us.
ZOMORODI: Okay. Are there humans involved in this mission?
NOSANOV: No. Great way to narrow it down. So there’s a human side and the robotic side. So I’m asking about the robotic side.
DUBNER: The most useful mission?
DUBNER: It’s going to be something about gathering information with a big telescope.
DUBNER: Is it a flyby? Is it — you fly by Mars?
NOSANOV: No it’s not. It’s an Earth orbit.
ZOMORODI: So it’s not the cute Mars lander? That guy?
NOSANOV: No. No.
DUBNER: Does it have to do with the location of space minerals?
NOSANOV: In a sense.
NOSANOV: Sort of.
DUBNER: Minerals and gases. Antibodies?
NOSANOV: No. No. That would be — if that had ever happened that would be the answer, but we’re not there yet.
ZOMORODI: Is the mission ongoing?
NOSANOV: Sort of.
JACOBS: This is going to be very hard to fact-check.
DUBNER: Have you ever said either yes or no in your life?
NOSANOV: Yes, yes I have.
DUBNER: Why don’t you tell us?
NOSANOV: Sure. So, the mission that I think is the most important and most useful is the Kepler space telescope. Do you have a penny up here by any chance? A penny in your pocket?
DUBNER: I don’t have a penny in my pocket.
NOSANOV: Okay, so hold up your finger and look at— Look at the ceiling kind of, in the shape of a penny. If you make a telescope that looks through that—
DUBNER: If I had a penny I couldn’t see the ceiling, so why did you want us to use a penny?
NOSANOV: If you imagine a cone that goes from your eye through that penny, and out in space, and you look through that telescope you will find thousands and thousands of planets that are similar in a lot of ways to the ones we have here in the solar system. And that’s just in a penny-sized slice or section of the sky. And that’s not even looking that far. That’s just staying within our own galaxy.
So, since we were all kids the number of planets has gone from 9 to 8 to about 3,000. And to me that I think is the most important and useful mission because it truly places into an unimaginable perspective everything else that NASA does and that, really, humans do. And every other field of science including minerals and gases—
DUBNER: Now, I believe Kepler was recently retired, but it was a massive success, wasn’t it? It was up there something like three times as long as originally planned?
NOSANOV: This is a really very much accomplished mission, which is showing us that the galaxy at least, and the universe by extension, is full of planets, more than there are stars. And that to me, the philosophical conclusion there, is that it’s almost impossible that the conditions that make Earth unique are unique.
DUBNER: So that is fascinating. It resonates and I think it’s an interesting answer, that Kepler is the most useful mission. Do you have a larger point about NASA and usefulness though? Because it’s a big point of contention.
NOSANOV: Well what initially drew me to the podcast was the idea of hidden economies, and the idea that for every mission that you read about in the paper or that you see photos from, there’s hundreds of other missions that are designed, evaluated, and rejected.
DUBNER: What is the actual rejection ratio, would you say?
NOSANOV: Oh, a hundred to one. If not more.
NOSANOV: And that’s because it’s that much harder to actually build and fly something to another planet.
DUBNER: Are the rejections primarily for a lack of technical or engineering ability?
NOSANOV: That’s where the hidden economies come in. There’s really four factors that really matter. There’s risk, cost, science return, and technology development. And that has changed over time, the weighting of those has changed over time. So in Apollo, everything was off the charts but they did it anyway. Now there’s a larger focus on minimizing risk and maximizing science return, which makes sense. But if you tell them you’re going to discover life somewhere and your technology needs another 10 years, you’re going to get rejected for multiple reasons. Science — we’re not ready to make that conclusion yet — too risky, and too much technology work.
DUBNER: When you said the four dimensions on which a proposal is assessed, and you talked about risk. What is meant in that context?
NOSANOV: Well, there’s a number of components. There’s actual technical risk, like can we look for life somewhere? Does the scientific evidence support a look for life, a search for life? And I believe that there is sort of a larger philosophical debate that goes around in the top floor of NASA headquarters about do we really want to support a mission that, if it’s successful, we have to declare for all time that we have discovered life somewhere else?
DUBNER: Why would that be a burden?
NOSANOV: Announcing you’ve discovered life somewhere else will permanently change human history. It will throw, in my opinion, countless ideologies into internal conflict. I hope I see it, but I can see that it will be hard for someone to sign off on.
DUBNER: Overall what share of the collective missions would you say are driven by scientific concerns, and not political or economic concerns?
NOSANOV: Well, formally the answer is about 10 percent, which is the Science Mission Directorate of NASA that sends out these robotic spacecraft. If a human spaceflight expert were here, they would probably point out the tremendous advancements that come sending humans into space, and those are all true. But the human side of NASA has from the beginning been associated with political gamesmanship. And that doesn’t take away anything from it, but the pure science, the “What are the rocks on Mars made of,” that’s only about 10 percent.
ZOMORODI: So but does that change now that it may not be up to you to decide? Because Elon Musk might do it for you?
NOSANOV: Elon Musk is actually doing a great service for those of us who try to get missions off the ground because he’s building the delivery truck. And we haven’t had a really good delivery truck at a cost that’s sustainable for a while.
DUBNER: If you talk about risk being a barrier, is working with a private firm like SpaceX, is that essentially a way of kind of offshoring some of the risk for NASA?
NOSANOV: So, and this is this is worth noting in its own right, so space launch is no longer really considered a risk, it’s a cost. You don’t really have to say, “Well they might blow up.” And I think it’s worth noting, as a species we’ve reached a point where we can say, “Yeah putting stuff into space is no longer the hard part.” So it’s not really a risk. It can help with the cost though. The SpaceX rockets are still a little— They have a slightly shorter history of success than some others, but we certainly propose to use them whenever we can because they save us money on other stuff
DUBNER: So I hate to ask you to reduce an extraordinarily complex and fascinating set of ideas into essentially a headline, but I am curious to know what’s your problem? What is the thing that you want to happen? Do you want NASA to take more of a different kind of risk?
NOSANOV: So I have a naive answer and a realistic one. The naive one is that I would like people to march on Washington demanding more funding for space science missions. More realistically I would like —
DUBNER: Did you hear those deafening cheers from people that were ready to go march?
NOSANOV: Thank you. You can send them this recording. More realistically I think I would like to see the re-emergence of a scientifically, confident, literate, encouraging society across the board.
ZOMORODI: Wait which one is naive, did you say?
DUBNER: Can I ask you one last thing before we turn it over to A.J.? The Kepler observatory I believe was built by Ball Aerospace, which is a subsidy is a subsidiary of Ball Corporation, which until 1993 I believe used to make Ball mason jars. So I’m curious if that’s the root of the NASA problem, somehow.
NOSANOV: Ball did build part of it. And that’s part of what we did to reduce cost.
DUBNER: They built the jams and jellies for the mission?
NOSANOV: I think they built the main spacecraft part and the telescope came from somewhere else. So NASA doesn’t build all of those spacecraft anymore. When possible, we use contractors and vendors.
DUBNER: A.J. Jacobs, Jeff Nosanov, a NASA consultant, has been telling us a lot of interesting things about what he feels are the slightly wrong headed philosophies behind NASA. Keeping in mind he has a little bit of a horse in the race as a consultant who wants to get his projects going. How much of what he said was totally false, A.J.?
JACOBS: About 40 percent.
JACOBS: No. In my extensive research it did check out. And I’m a big fan of the Kepler telescope. It found over a thousand planets, and I actually looked up what the planets were called. They’ve got some wonderful names. There’s Kepler 560-B and Kepler 438-B. There’s a crowd favorite of Kepler 841-B.
JACOBS: So you guys need some creativity, I think.
NOSANOV: Once we know a little bit more about them, other than there’s one there, it will be easier to name them. I think.
JACOBS: There are a lot of Roman gods and goddess left.
NOSANOV: Yeah that’s true.
JACOBS: And I also looked up the original Kepler, that it’s named for. He’s Johannes Kepler, 17th century astronomer, and it turns out, appropriately enough, he had money problems. So NASA, he had money problems. And he had to supplement his astronomy work with astrology. He was the astrologer to the Holy Roman Emperor. Which I think is just super sad because it’s like Stephen Hawking reading tarot cards. I mean it’s the great scientist of the period.
NOSANOV: So I did not know that.
DUBNER: Jeff, thank you so much for playing Tell Me Something I Don’t Know.
* * *
DUBNER: Before we get back to the game we have got some FREAK-quently asked questions written just for you, Manoush. Are you ready for it?
ZOMORODI: Okay. Yeah.
DUBNER: Manoush we know that your latest podcast, ZigZag, spends a lot of time talking about the blockchain. So for those who still don’t get it, can you explain the blockchain in 30 seconds or less, please?
ZOMORODI: 30 seconds. Okay. Think of it as Google Docs, right? If you have a document in Google Docs, if you change it, everyone sees the change right? And when you go back in, if somebody else changed it, you see it as well. Think of that, but with no Google being in charge. Pretty cool, right?
DUBNER: That was phenomenal, and it was 17 seconds.
ZOMORODI: I could keep— Do you want me to keep going?
DUBNER: Give me 13 seconds more on it.
ZOMORODI: Okay, great. So also, blockchain, it really truly is like a necklace of computers across the world all linked together. When the change gets made, it goes across the entire necklace. You can have private ones, like a bank can have its own blockchain, or you can have more public ones, like Ethereum, which anyone can join the Ethereum blockchain, as it were. And you can layer— Keep going?
DUBNER: You’re way over your 13 seconds. Well, let me ask you one more question on blockchain. Most people who know a little bit about blockchain are alternately kind of enamored and petrified, especially when it’s attached to a currency, which tends to be very volatile. Tell us your one very-favorite totally non-currency potential application of blockchain.
ZOMORODI: Saving journalism. That’s the weirdo experiment that I’ve been part of. Anybody heard of Civil here? [Silence] Great. Okay, so Civil is the blockchain startup that my former executive producer at WNYC, we quit our jobs to join this weirdo startup and we are documenting the entire process of trying to get our heads around what blockchain is, how it could potentially save journalism. The idea with Civil is that there would be a network of trusted, verified, little media publications and people would be able to sort of pay for them as they go, or tip, or vote if somebody is putting nonsense on there, fake news could get voted off by staking their tokens. Spoiler, the token sale failed this week. All documented on our podcast, ZigZag.
DUBNER: So that sounds like a great project. Let me ask you this, since you have an interesting relationship with technology. Sometimes very pro, sometimes much less pro. What is your personal strategy for creating and managing computer passwords?
ZOMORODI: Okay, I’ve never told anyone this.
DUBNER: It’s just me here.
ZOMORODI: I write messages to the tech platforms about how I really feel about them. You’re supposed to be like random strings of words is a better way to write a password? So I will write like, “F.U. Facebook” and then like —
DUBNER: So your whole string of passwords it’s just a litany of your feelings towards the tech companies that you engage with?
ZOMORODI: Yes, correct. Or if I like the tech company, it’s a message to myself reminding me of why I like it.
DUBNER: Give me a for instance. In other words, just tell us your Amazon password.
ZOMORODI: I’ll give you another one. One is like a running app and it was like words of encouragement to myself.
DUBNER: Oh that’s so cute.
ZOMORODI: Kind of sweet right? Anyway I also use a password manager, which you all should do so that that you’re not keeping your passwords in places that are not—
DUBNER: Does it run on blockchain?
ZOMORODI: No, it does not run on blockchain.
DUBNER: And finally, Manoush Zomorodi, you’ve worked in many media forms, radio, TV, books. Why in your view is podcasting superior to all of them?
ZOMORODI: Because of the listeners, right? No, I’m serious. It’s the truth. I’ve been a journalist for a long time now and only when I started doing podcasts would people write me the most personal incredible emails. Hug me when I met them at events. It’s a relationship that I have never had before with people I don’t know based on sharing of stories and information.
DUBNER: And you’re comfortable with this level of forced intimacy?
ZOMORODI: Have you listened to my shows, Stephen? I’m pretty comfortable with a lot of things. Yes.
DUBNER: Ladies and gentlemen, Manoush Zomorodi. Alrighty then. Let’s get back to our game. Would you please welcome our next guest Scott Highhouse. Hi Scott, it says here that you are a professor of psychology at Bowling Green State University in Ohio. And I understand you have a riddle for us of some sort. Yes?
Scott HIGHHOUSE: There was a study in 1959 that showed that psychiatric nurses in a mental hospital were just as good at predicting patient readmission as were the expert psychiatrists. So what were the nurses doing?
ZOMORODI: Listening to the patients? Oh, I feel bad. My mom and dad are going to listen to this. Were they actually — because there’s new technology out there that is analyzing voice and can predict when someone is going to have a psychotic break. Were they actually listening to the way that they spoke?
DUBNER: You just loved to stomp on her enthusiasm, didn’t you? Why don’t you tell us because I have a feeling that the story behind the answer is interesting.
HIGHHOUSE: They took each patient’s folder and placed it on a kitchen scale and the heavier folders predicted readmission more than the lighter folders.
DUBNER: So is this what people talk about when they talk about Occam’s Razor? The simplest theory is more likely to be correct? Or is it something different than that?
HIGHHOUSE: I don’t disagree with that. I think the general principle is that expert intuition is not very good when it comes to making predictions, particularly about people’s behaviors and their performance.
ZOMORODI: So are you also saying that this could not be replicated in this day and age due to digital files?
HIGHHOUSE: Right. You’d have to be more creative. I think maybe the length of the file or something.
DUBNER: So let me ask you this. You’re a psychology researcher.
DUBNER: Is this story that you just told us related particularly to the work you do? Or does this expert intuition idea travel across domains?
HIGHHOUSE: Yes, my area is industrial organizational psychology and I’m interested specifically in hiring and interviewing. We know that intuition is a derailer. We knew way back that admissions officers for universities who knew the G.P.A. and the S.A.T. score, screwed things up when they added their holistic judgment about the students. And we find the same thing with job interviews. Expert interviewers, experienced interviewers in H.R. are actually worse than a layperson who uses structured questions that are job related and behavioral in nature.
DUBNER: So, can you give an example of a good interview question and a poor one that’s more fishing for intuition?
HIGHHOUSE: Yes. A traditional interview question would be, “Why do you want to work here?” “Tell me about yourself.”
DUBNER: And those are boring sounding, but you’re saying they work.
HIGHHOUSE: No, those are more intuitive. “What is that drives you,” and things like that. A more structured question would be job related and behavioral. So, “Tell me about a time when you encountered conflict at work and what did you do about it.” Or, “What would you do in a situation where someone tried to undermine you at work?” So those very behavioral questions, are very specific and they ask about what would you do or what did you do.
DUBNER: Of all the domains in which all of us engage all the time — so workplace, dating and mating, an example like you gave in the medical field where you’re trying to assess someone’s prospects or assessing anyone’s prospects — where do you find intuition is most heavily relied on, and therefore most damaging?
HIGHHOUSE: Oh, goodness. I do know that — maybe it doesn’t answer your question directly — but intuition is good in some areas like wine tasting and art appreciation.
DUBNER: But when you say it’s good—
HIGHHOUSE: Well those are studies based on an agreement with experts. So—
DUBNER: But the experts you just told us are full of —
HIGHHOUSE: My area is prediction, remember. We are trying to predict future performance on the job. So in areas where they look at agreement with experts on aesthetic judgments, intuition seems to work well, and the more you think about, “Is this de Kooning a good painting?” the farther away you get from expert judgment.
DUBNER: I find really interesting the idea that in a job interview, but I’m guessing in any context, if we asked for specific behavioral response, whether it’s theoretical or real from history, I mean that makes a lot of sense. A.J., Scott’s telling us that intuition is to be leery of, at least in some cases and that experts tend to have a lot of it and make some bad decisions. What more can you tell us about that?
JACOBS: Well, yeah, both my intuition and the data support that intuition is terrible. It’s a terrible predictor of future. I actually kind of got sidetracked because I’m a fan of the old fashioned ways of predicting the future. And maybe you can tell me how successful they are. There’s Bontroscopy, which is predicting the future by the sound of thunder; Haruspex, which is predicting the future from the livers of sacrificed sheep; and Myomancy, predicting the future by the movement of rats and mice.
DUBNER: Do you use a lot of sheep livers at Bowling Green?
HIGHHOUSE: No, but there are areas of employee-hiring where they look at handwriting or, many years ago, bumps on the head. And none of those were very useful.
DUBNER: But if you got your bump in a very dramatic way, it could tell you something about the person.
JACOBS: And, just so you know, Zarf, the coffee cup sleeve, and I know you’ve been waiting — from the Arabic, Zarf for vessel.
ZOMORODI: Thank you.
DUBNER: A.J. Jacobs as always going way above and beyond the call duty. A.J. thank you and Scott Highhouse, thank you so much for playing. It’s time for our final guest of the evening. So would you please welcome her, Rada Mihalcea. Rada is the director of the Artificial Intelligence Lab at the University of Michigan. Rada, the floor is yours.
MIHALCEA: I have a timely topic. How can you increase your odds of finding out if a news article is true or fake?
ZOMORODI: Does it involve A.J.?
JACOBS: I am available.
DUBNER: I wonder if what we just heard from Scott Highhouse should weigh into it in some way, which is distrusting intuition. Does that have anything to do with it or no?
MIHALCEA: To some extent.
DUBNER: Cagey answer. Let’s say, more than just discounting intuition, seeking out firm behavioral or structural elements like punctuation or typography.
MIHALCEA: Getting closer.
ZOMORODI: Okay, so the second season of ZigZag is trust and information, is our theme. And we just did something with the Knight Foundation looking at how misinformation and fake news essentially spread on Twitter before the 2016 election and post. And actually the crazy surprising finding was that all that fake information, the millions of tweets went back to just a few dozen sites. So it was far more centralized than people thought before.
The other fact that was really interesting was that 95 to 97 percent of the information coming from those news sites was true. It was a very small amount — it was that 3 to 5 percent that was nonsense that really got pumped out across Twitter. So, counterintuitively one might say is the genesis with somewhat reputable sites or sites that are well established?
MIHALCEA: And I think that would add to the challenge in fact, because you cannot really rely on the source.
DUBNER: You work in language and I.T. So your answer has something to do with technology computing. Yes?
MIHALCEA: That’s true.
DUBNER: Okay so are you in possession of a pretty good method or algorithm to determine fake news? Is that what you’re saying? You have in your pocket something useful?
MIHALCEA: Right. So your best bet would be to bring along a computer. It turns out that computers are better than people at detecting deception. What we found with our algorithms for instance in courtrooms, we can spot witnesses who are lying about 75 percent of the time. Which is quite a bit better than what people would do at the same task. So people do a little bit better than random at 55 percent. In fake news people are better. They get fake news about 70 percent of the time.
DUBNER: And computers do what?
MIHALCEA: And computers were 76. So people are still behind the computers at detecting fake news.
DUBNER: Okay so what are the computers actually doing though. Is it a text analysis, is it finding inconsistency in mood or language? What’s happening?
MIHALCEA: So computers are basically learning from data. They are learning from collections of lies and truths what are the attributes of those. So basically we program the system to look for certain features or attributes like sequences of words, or relations between words or the semantic type of the word.
DUBNER: Can you give me an example of a phrase or a sentence or even a word that would indicate fakeness?
MIHALCEA: So one of the aspects of language that computers would pick on is the use of personal pronouns. Liars would tend to use less often first person pronouns like I, me, myself, we, and instead would use more often he, she, they. Psychologists would explain that by saying that liars would want to detach themselves from the lie. Another one which I think it’s somehow counterintuitive is the use of words that reflect certainty. Liars will tend to more often use words such as always, absolutely, or using exaggeration.
ZOMORODI: The best!
MIHALCEA: The best, there we go. Unlike the truth tellers. Truth tellers would more often use hedging, like “maybe, perhaps, probably,” something like that.
DUBNER: Interesting. So, cynical question, by publicly discussing this research both in general and specifically, aren’t you just making it easier for the purveyors of fake news to get better?
MIHALCEA: Not necessarily. I think the clues that computers tend to pick on are not intuitive for humans. So if I were to ask you how many times I said “I,” you probably have no idea, because you don’t look for those little words that actually make a difference in deception detection. So it is still hard. Even if you want to prevent others or a computer to detect deception, it’s actually hard.
ZOMORODI: Are you using deep learning I’m assuming with processing all this information?
MIHALCEA: We do use deep learning in other projects but not in this particular one.
ZOMORODI: What are you using?
MIHALCEA: We are using machine learning. The reason being that deep learning works very well when it has a lot of data. So you need a lot of lies, a lot of truths.
ZOMORODI: We know some places.
DUBNER: Can I just ask: how surprised should we be that this is the kind of task that computers are better at? I mean isn’t the list of things that humans are better at computers than getting really short? I don’t mean to degrade the value of this kind of identification, but I guess it’s just not so surprising that a computer, an algorithm would be better than an intuitive, emotional, impatient human being, because we know that we’re bad at those things right?
MIHALCEA: Well, yes and no. I think there are certain tasks where we are still better. Like for instance, writing. If you were to write a novel, people are still much better. So there is still a fair number of applications or ways in which we are much better.
ZOMORODI: So, I think the question is, could you use the technology to parse Brett Kavanaugh’s testimony?
MIHALCEA: We could and we are planning to. So we are working on—
ZOMORODI: Oh! You heard it here first! And Dr. Ford, of course. I mean, right?
MIHALCEA: Of course, the whole dialogue.
DUBNER: A.J. Jacobs, Rada Mihalcea from the University of Michigan is telling us that computers are getting pretty good at detecting deception. What more to add?
JACOBS: It is looking good. These are early days for this technology. But I mean we desperately need it. And I looked into a little of the history of truth detection devices. Because the polygraph tests that measure your pulse, and your skin, they are not that reliable. The American Psychological Association says to be very skeptical. Though, in their defense, polygraphs do have a very cool backstory. Because one of the inventors of the polygraph was William Marston, the man who created Wonder Woman, the superhero. And Wonder Woman’s lasso of truth, that is 100 percent scientific and reliable. So that’s the secret.
DUBNER: A.J., thank you. And Rada Mihalcea. Thank you so much for playing. Can we give one more hand to all our guests tonight? It is time now for our live audience to pick a winner. So who’s it going to be?
- Julie Arslanoglu, with using antibodies to answer art questions?
- David Reitter, with how to manipulate away our impatience?
- Jeff Nosanov, with rethinking risk in NASA and space?
- Scott Highhouse, with how intuition is often wrong?
- Or Rada Mihalcea, with detecting deception with computers?
DUBNER: Okay, the audience vote is in. Once again thanks so much to all our guests presenters. And our grand-prize winner tonight, you could chalk this up to a little recency bias, but I don’t think so, for telling us about detecting deception with computers, Rada Mihalcea, congratulations. And Rada, to commemorate your victory we’d like to present you with this Certificate of Impressive Knowledge. It reads, in full, “I, Stephen Dubner, in consultation with Manoush Zomorodi and A.J. Jacobs, do hereby vow that Rada Mihalcea, told us something we did not know for which we are eternally grateful.” That’s our show for tonight. I hope we told you something you didn’t know. Huge thanks to Manoush and A.J., to our guests, and thanks especially to you for coming to play “Tell Me Something …
AUDIENCE: I Don’t Know!”
Tell Me Something I Don’t Know and Freakonomics Radio are produced by Stitcher and Dubner Productions. This episode was produced by Alison Craiglow, Harry Huggins, Zack Lapinski, Morgan Levey, Emma Morgenstern, Dan Dzula, and David Herman, who also composed our theme music. The Freakonomics Radio staff also includes Greg Rippin and Alvin Melathe. Thanks to our good friends at Qualtrics, whose online survey software is so helpful in putting on this show, and to Joe’s Pub at the Public Theater for hosting us.