Question: Describe your "invisible gorilla" experiment.
Christopher Chabris: Dan Simons and I were teaching a course at Harvard University on research methods in the Psychology Department. This is a course where students have to design their own psychology research projects and carry them out. And as part of the course, we also thought it would be a good idea to have some group projects that everyone could participate in. Dan had the idea to look at some work that had been done in the 1970’s, a somewhat famous experiment by Ulrich Neisser, one of the pioneers of cognitive psychology who had some people sort of play basketball passing balls around on an empty basketball court, and managed to get the groups of people to overlap by filming them with special mirrors. And then while people were watching these basketball players, they’re supposed to count the number of times the ball gets passed. About halfway through the video, a woman carrying an umbrella walks through the basketball court
The surprising finding of his experiment was that many people who were counting the basketball passes didn’t see the woman with the umbrella walking through and didn’t remember her being there at all. The funny thing about the experiment though, was that since it was filmed using the special technique with mirrors, everyone was sort of invisible and transparent, and you could see through them. It was a very unusual looking display. It’s the kind of thing that nowadays you would do with digital video editing. Back in the 1970’s, you used mirrors.
Some people had sort of dismissed the finding that people missed this very obvious and salient thing, like a woman walking through a basketball game because it was a strange-looking visual display. So, we decided to try to do a new version of this experiment where all the action was live, and we had six people passing basketballs around, sort of choreographed to not run into each other and to not throw the ball in each other's face, and so on. And that was a bit of a challenge insetting up this experiment, but that’s what we have students in the course for. We managed to get it right.
And then we had a woman walk through carrying an umbrella, just like Neisser did. What was a little more amusing for us though was that one of the other professors in the department happened to have a gorilla suit lying around in his lab and we thought it would be fun to have someone walk though wearing the gorilla suit and see whether that was noticed.
And we had first thought that people would notice the gorilla walking through the basketball game because there’s no longer this sort of transparency through display and the gorilla actually stayed on the screen for nine full seconds in one of the versions of our video.
But when we ran the experiment and our students went out and tested people on the Harvard campus, we found that about half of the people did not at all notice the gorilla and, in fact, were very surprised they hadn’t noticed the gorilla. There were actually two findings from this experiment. One, you can miss very salient things, like a gorilla walking right in front of you, and two, that you’re shocked that you could miss it. Most people seem to have the intuitive idea that they’re going to see this kind of thing and they’re really surprised when they find out that they don’t.
Question: What does this experiment demonstrate to us about selective attention?
Christopher Chabris: What this experiment shows is that when we’re paying attention to something, basically doing a task that demands our attention such as counting the passes of the basketball in this case, or really any other kind of really attention-demanding task that we do, we can seriously overestimate our ability to do other tasks at the same time and especially to notice and handle unexpected or surprising things. We think that we’re going to notice unexpected things that come into our field of view and we think we’re going to pay attention to the things we should pay attention to, but in fact, when we’re focused on one task, we’re noticing and paying attention to a lot less than we really think.
Question: What implications does this research have for real-world multi-tasking?
Christopher Chabris: This experiment implies a lot about our behavior in every day life. For example, when we are driving and talking on a cell phone at the same time, we get the feeling that we are actually driving just a well as when we are not talking on the phone. That’s part of why we do things like talk on the phone while we are driving, or send text messages or read our email, which are actually much worse. But even talking on the phone while driving depletes a lot of your attention.
Talking on the phone while you are driving is like counting the basketball passes in our experiment. You can still drive, but what you’re going to have a problem with is noticing unexpected things. And when you think about it, those are the really important things to notice when you are driving. It’s one thing to just drive down an open highway at night, it’s another thing to drive in the suburbs when someone might be pulling out of their driveway right in front of you, or someone might be stopping short right in front of you and you won’t have the reaction time to be able to stop and not crash into them, or even worse, someone pushes a baby stroller out in front of the road. So, we’re actually a lot worse at that than we think we are and we should, in fact, put down the cell phone while we are driving and perhaps even zip it up in a brief case or a purse or something like that.
Those are some of the implications for everyday life. But it’s not just driving; there are many other situations like this. For example, the security screeners at airports... we sort of have the impression the are paying attention that’s in our suitcase, and our luggage and so on, but actually everybody has sort of an inflated impression of how much they are paying attention to, and we may not be as safe as we think we are even with trained and dedicated people doing that. Lifeguards at swimming pools. Radiologists looking at medical scans for anomalies and so on. When you are looking for one thing and paying attention to one thing, it is easy to not notice other things and to not realize that you are not noticing those things.
Question: What is the funniest reaction you've gotten from your attention experiments?
Christopher Chabris: A TV news magazine that covered our research actually recreated the experiment for themselves in their own studios and they had a bunch of volunteers come in and look at the video tape and count the passes and missed the gorilla and they filmed the reactions of some of them. And my favorite one of all was someone who said, “That bear didn’t go through there, did it?" ...
Did you notice that what I’m wearing actually changed several times during the course of this interview. Some people might have noticed some of the changes, but it’s probably unusual to notice all of those changes.
Question: What are we talking about when we talk about intuition?
Christopher Chabris: There’s been a lot of talk over the past few years and some popular books about the power of intuition and how if we trusted our intuitions or went with our guts more often, we would actually wind up making better decisions and the world would even be a better place. Some researchers have talked about how our intuitions and the intuitive judgments we can make, the snap judgments, are actually accurate a lot of the time. And a lot of that is true, but what we noticed when thinking about our gorilla experiment and doing research for this book is that there’s a whole category of intuitions that are actually systematically wrong and in very dangerous ways. And those are the intuitions we have about how our own minds work.
So our gorilla experiment shows that we intuitively think that we pay attention to and notice much more than we actually do, and that can have tragic consequences. If you get into a car accident or... one of the examples in our book is a nuclear submarine that surfaced right into a Japanese fishing boat because, in part, the captain of the submarine looked up in the periscope, didn’t see any boats around and surfaced right into one that he didn’t see. That was actually probably right in front of him.
So there’s a case where intuitively, we think something and it can really get us into big trouble. We realize as we were thinking about the book and putting together the ideas that this category of intuitions is very broad, it’s not just about visual attention. It’s also about memory. We think we remember things much more accurately than we really do. It’s about confidence. We think that people who are confident are also more skilled and accurate and knowledgeable. It’s about our own knowledge. We think that we know and understand more about the world than we really do. And so on. There’s sort of like a whole sort of set of these intuitions which can really lead us astray if we’re not aware of them. And that’s what we decided to write the book about.
Question: Is your research in dialogue with recent books that tout the value of intuition?
Christopher Chabris: We are, in a way, taking on the impression that a lot of people have from books like, "Blink," by Malcolm Gladwell, and others in that category, which is sort of an uncritical belief in the power of intuition and snap judgments and so on, and the idea that you should rely on them whenever possible. We sort of are, in a sense taking on the cult of intuition or the myth of intuition as we call it in the book.
I should also say that we’re very pleased that Malcolm Gladwell actually read our original Scientific Journal article about the Invisible Gorilla study and talked about it in one of his New Yorker articles which helped it get more publicity and ultimately we went into it becoming more widely known and as famous as it is.
Question: Why do we overestimate the accuracy of forecasters?
Christopher Chabris: In the area of confidence, we have a lot of trouble properly interpreting the level of confidence that other people express.
A thing we do is we attach too much significant to how confident someone seems and appears and acts. In another sense, we prefer confident people more than less-confident people, even when they’re less accurate and less knowledgeable. So for example, a Dutch psychologist named Gideon Keren did a very clever experiment where he said simply to people, here are two weather forecasters, and they have given the forecast, percentage chance of rain on four consecutive days. Forecaster A says, there’s a 90% chance of rain on Monday, a 90% chance on Tuesday, a 90% chance on Wednesday, and a 90% chance on Thursday. Forecaster B says, there’s a 75% chance on Monday, 75% on Tuesday, 75% on Wednesday and 75% on Thursday.
In fact, after those four days, it turns out to have rained three out of those four days, that is 75% of the days. Which weather forecaster do you think was the better forecaster?
Now, the right answer is Forecaster B because Forecaster B said there was a 75% chance of rain, and he was right. On 75% of those days it actually did rain. He’s what we call perfectly calibrated. But a majority of subjects in this study preferred Forecaster A, and the only difference between Forecaster A and B is that Forecaster A is more confident. Forecaster A is sort of more decisive in his forecasts, but less accurate. Yet we still prefer that forecaster, even knowing their track record.
And here we actually know what their track record is. So we should really be able to say, "Look, this guy got it right, he’s the right one," but instead, even knowing their track record, we go for the one who’s more confident. And this is an unusual case actually because you’ve been given all the information you need to know to make the right decision.
If you think about people who forecast the stock market, or who forecast political elections, or who forecast, you know, trends in shopping and design and so on. We don’t know what the track record of those people are when we pick which ones we like better. So even more seduced by confidence and we don’t realize that we’re paying so much attention to that and we think we’re making a better decision than we really are.
Question: How do the fields of psychology and neurology intersect?
Christopher Chabris: My research is in sort of two areas. One is traditional cognitive psychology and the other is cognitive neuroscience, a more modern field that uses technologies like brain scanning to figure out how what’s going on in the brain explains how our minds work. And these are exciting developments and exciting technologies that enable us to learn those kinds of things and I’m all for them. But, we do point out in our books sort of one unfortunate side effect of some of this new development and we call these side effects "neurobabble" and "brain porn." And "neurobabble" is sort of the adornment of explanations about the mind with references to the brain. A lot of people think that if you can say something about the brain, then you must know more about how the mind works. And it really doesn’t work that way. The brain and the mind are not entirely the same thing Just seeing a picture of the brain with some colorful blobs showing this is where the brain is active while we’re doing some particular task, doesn’t necessarily tell us a whole lot more about that task and how our minds accomplish it.
Likewise, a lot of talk about the brain is often used to sort of... as an add on to other arguments to make them seem more convincing as sort of like an all-purpose you now, reference to make your argument more convincing. To say it has something to do with the brain and show a picture of the brain and so on. Advertising does this quite a bit as well. You can see the brain, if you look, in all kinds of advertisements where when you think about it, it’s really not that relevant. And I sort of am a little bit troubled about that trend, while at the same time I think that it’s exciting to discover more about the brain just as long as we know what the knowledge means and that it doesn’t mean that traditional psychology is somehow obsolete and replaced by neuroscience and so on.
They are two topics of study that definitely relate to each other quite a bit, but the one doesn’t replace or really make the other any more credible.
Question: What is a "flash-bulb memory?"
Christopher Chabris: A "flash-bulb memory" is a memory that forms kind of as though a flash is going off in a camera and you imprint a picture in your mind of what’s going on at a particular moment in time. That was a term that was devised and coined in the 1970s by the late social psychologist, Roger Brown, who did a study on people's memories of the assassination of President Kennedy. And he found that people had extremely vivid memories of where they were when they heard about it, what they were thinking, doing, who they were with, what they did next. And he found also for other significant events, like the assassination of Martin Luther King, people had formed similar memories.
And he concluded on that basis that highly significant events sort of imprint themselves into your memory and you’re going to always remember them. It turns out that... and this is kind of a natural belief, and this shouldn’t surprise us. For example, we all probably have a pretty good idea where we were when we heard about the terrorist attacks on September 11, 2001. I have a good memory of it. I actually heard about it on "The Howard Stern Show," of all places, when I woke up, I was in graduate school, so I woke up late those days and I actually tuned into the Howard Stern Show, and that’s how I found out about it.
At least I think that’s how I found out about it. That’s how I remember it. When people have actually done clever studies on flash-bulb memories to look and see whether our intuitive beliefs about how accurate they are match up with their true accuracy, they find out that the flash-bulb memories are not really any more accurate than ordinary everyday memories.
One especially interesting study was done by Ulrich Neisser, who had also done, long ago, the study that inspired our gorilla study. He actually, after the Challenger space shuttle exploded in 1986, the very next day he went to a class full of students and had them write down all this information. Where they were, who they were with, how they heard about it, and so on. And then followed up several years later before they graduated college and had them recall it again. Found out that their recall, years later, did not really match what they had written down the day after, but their confidence was extremely high. They were sure that that’s exactly how they remembered it. They had no doubt, whereas, of course they couldn’t tell you what they were doing the day before the Challenger exploded or the day after the Challenger exploded, just like all of us probably have no memory of what was going on, on September 10, 2001. So, the thing about flash-bulb memories that really makes them sort of deceptive is that they’re no more accurate than ordinary memories, they’re subject to the same kind of distortions that just happen in time to all of our memories, but we’re more confident in them because they’re so vivid and they’re so detailed and we sort of place, really, unwarranted faith in them.
Question: If these kinds of highly vivid memories aren’t reliable, are any memories reliable?
Christopher Chabris: Memory is not a complete fraud, we do remember some things. It’s not as though everything in our memory is a distortion and inaccurate and so on. One thing that’s been learned from a lot of research on memory over decades is that memory for the jest of something, for the main idea, is much better than memory for specific details.
Memory for details can change and fade over time. Memory for sort of main ideas, themes, emotional experiences, probably nobody really misremembers how they felt on September 11, but they might misremember where they were, who told them about it, and details like that. Especially the farther away they are from the epicenter of events. But they’re not going to forget how they felt on that day as easily.
So, sort of the overall message often comes through, but the details can change and fade over time. The problem, of course, is that we don’t realize that. So, we can get into big arguments over the details of memories, who said what to whom, when did they say it? Exactly what was said before that, what was said after that. Think about how many arguments you’ve gotten into over the course of your life where that’s what’s going on and those are kind of silly arguments because nobody can really trust their memory as much as they claim to in the heat of the moment.
Question: How do errors in film continuity relate to illusions and memory errors?
Christopher Chabris: A surprising fact about memory is that our memories can be pretty weak even for things that just happened a couple of seconds ago, and really a timeframe when you would think that memory should be pretty good. It’s one thing for memory to fade after a few years, but it’s another thing to get completely erased after a couple of seconds. And a great example of how this can happen is shown in films every day. Every movie that you watch has what are called "continuity errors," and there are catalogues of these on the web. You can and type in any movie name you want and find all the mistakes that the film editors made. Now, sometimes they realize they were making those mistakes, but they knew that most people wouldn’t see them. and in fact, when you watch movies, you hardly ever notice a continuity error because you’re paying attention to the plot and the action and the characters and so on and you don’t notice that, for example, in the “Godfather” there’s a glass of wine on the table in one scene and when the camera comes back to it, it’s gone, and then when the camera comes back again, it’s back again. But those are really sort of failures of memory.
You looked at that scene and you felt like you were taking it all in, and then when you came back to it from a different camera angle, or from a cut, you didn’t bother to match up your previous memory to what was then on the screen afterward, or you didn’t even store as much detail about it in the first place as you thought you did. You might not have actually even stored the information about the wine glass even though you paid attention to it at the time.
So it’s one thing to pay attention to things and notice them, but it’s a whole other thing to get them into memory and this phenomenon of continuity errors and how many changes we can miss as a video cuts from one angle to the other, from one scene back to another, illustrates another aspect of the illusion of memory.
Question: Can you perform a quick memory test for us?
Christopher Chabris: One of the nice things about memory as a subject of study for cognitive psychologists is that it doesn’t require a lot of high technology to do an experiment. So, we can actually do a simple memory experiment right now that you can test yourself with. Now, I’ll just read out a series of words, and I want you to listen to the words and try to remember them, and then after a little while, I’ll ask you some questions about the words. So, we’ll get started. I read them about, oh one every second or so; bed, rest, awake, tired, dream, wake, snooze, blanket, doze, slumber, snore, nap, peace, yawn, drowsy.
So that was a simple list of words and what I’m going to do now is just ask you about some of the words that may or may not have been on the list. You can just think to yourself, now was that word part of the list of not.
Let’s start with, snooze. Think about whether that was part of the list. Now, we’ll try apple. Was that part of the list? Next, house, was that part of the list? How about dream? Was that part of the list? And what about sleep? Was that part of the list?
So we just did about the lowest technology possible memory experiment, a list of words that you have to remember and answer questions about it. This experiment was actually originally done by researchers named, James Deese and also Roddy Roediger and Kathleen McDermott. And what they found was that people were, of course, pretty good at realizing that which words were inside the list and which words were not in the list. But there’s one kind of mistake that they make pretty systematically and that’s falsely saying that the word "sleep" was part of the list. So, the word "sleep" wasn’t in the list that I just gave you. The list was bed, rest, awake, tired, dream, wake, snooze, blanket, doze, slumber, snore, nap, peace, yawn, and drowsy. All those word have to do with sleep, but "sleep" was actually not on the list. Yet a lot of subjects in this experiment falsely think that "sleep" was there. And they're usually pretty confident that sleep was on that list also. It’s not sort of a marginal thing.
So, even in a simple situation like this, just trying to remember a few words, our memories can play tricks on us.
Question: Can being aware of these kinds of cognitive illusions help us to be less selective in our attention?
Christopher Chabris: The illusions that we talk about have one main thing in common. Most of them are overestimations of our cognitive abilities or, another way of putting that is an under appreciation of some inherent limits on our cognitive abilities. We don’t appreciate the limits on how much we can pay attention to it once and therefore we think we can multi-task better than we can. We don’t appreciate the limits of what we actually store in memory and therefore we think our memories are more accurate than they are. How do we solve this problem?
Well, one would be raising those limits, or taking them away entirely. Somehow enabling our brains to pay attention to more things at once or our memories to store things more accurately and retrieve them more accurately, and so on. That might be technologically possible in some sense and in fact many technologies that we’ve invented over recorded history are designed to increase our cognitive limitation. Why do we write things down? Because oral tradition is not as good a way of conveying information as written history. We have lots of devices that enable us to do things like that better.
So, technological solutions may help us to increase some of those limits, but there’s a fundamental problem that sort of gets in the way of that which is that we have these illusions, or sort of false intuitions about those limits in the first place, so really what people can do and what they ought to do is think a little bit more about how their intuitions about the mind might be wrong. Sort of take to heart some of the examples that we talk about and not be so certain of themselves and not be so angry at other people who might be suffering from these illusions that affect all of us rather than being malicious or having some other bad motives.
Question: What are the implications of these types of cognitive errors for the criminal justice system?
Christopher Chabris: The American legal system is one which makes a lot of sort of assumptions and common sense... and uses a lot of common sense beliefs about how the mind works in its own procedures. For example, eyewitness testimony is thought to be an extremely high valued form of testimony. If someone says, "I saw that guy do it," that’s about the strongest testimony you can get. In fact, other kinds of evidence is sort of sometimes derisively labeled circumstantial, as though that’s somehow worse than eyewitness testimony.
But what research over decades by many psychologists have shown is that eyewitness testimony is in many ways, less than it’s cracked up to be because eyewitnesses are just as fallible as the rest of us in terms of how their memories work. And they’re just as unlikely to notice things; they’re just as likely to distort them in their memory over time. There have been many cases were eyewitness testimony convicted someone, they went as far as death row, and were exonerated by DNA evidence. And it was found that someone else did it. In cases where the witness was incredibly confident on the stand and this is a place where memory and confidence illusions interplay. If someone thinks their memories more accurate than they are, and they’re really poised and confident and convincing on the witness stand, then you’re in trouble if you’re the defendant who didn’t do it. And the synergy between those illusions can really be powerful in the case of the legal system.
The legal system in my opinion, needs to sort of come to grips more with the way the mind actually works and the kinds of beliefs that we have about the mind which aren't actually true and maybe rely a little less somehow on common sense and a little more on scientific evidence about how the mind works.
Recorded on May 13, 2010
Interviewed by Austin Allen