The Science of Everything Podcast - Episode 20: The Reliability of Memory

Episode Date: June 29, 2011

An analysis of the various biases that effect the accuracy of human memory, including a discussion of spatial memory distortions, the impact of schema on memory, source monitoring failure, and the misinformation effect. I also discuss the interesting research regarding false memories, eyewitness memory, and flashbulb memories. If you enjoyed the podcast please consider supporting the show by making a paypal donation or becoming a patreon supporter. https://www.patreon.com/jamesfodor https://www.paypal.me/ScienceofEverything

You're listening to The Science of Everything podcast, episode 20, The Reliability of Memory. So in this episode, I'm going to discuss the, well, the reliability of human memory. I'm going to focus my discussion around a number of different biases or inaccuracies or ways in which memory has been shown to be distorted or even falsified. And these different effects can sort of be. categorize in different ways. So the type of things I'm going to talk about include spatial memory distortion, the impact of schemers on memory, the misinformation effect, a very interesting one, source monitoring failure, false memories, flashbulb memories, and I'll also talk a bit about

eyewitness memory. This discussion, I think, will show that memory is not as reliable as we like to think that it is, and so we should always be careful when we're sure that we can remember something that we're sure happened. It may not have. have been as we had thought. Okay, so let's get into it. First I'll start with discussing spatial memory and distortion. There is a field of psychology called Gestalt's psychology, which kind of, it's a bit hard to define, but it sort of focuses on how people perceive and interact with broader frameworks or environments of stimuli rather than more traditional experimental psychology where you just have, you know, like you count how many words from a list they can't remember, Gestalt's psychologist

will kind of focus on sort of a bigger picture, more qualitative approach. And in terms of memory, the Gestalt psychologists produce some interesting results where they showed that reproductions, when people were shown shapes and then had to reproduce them later from memory, the shapes tended to shift towards more familiar typical forms. So, for example, if it was initially sort of an amorphous blob sort of shape, over time, the person would draw it from memory becoming closer and closer to a circle, or closer and closer to a square,

sort of looked like a square originally and so on. The other effect, though, is whereby if there was a certain particular characteristic of the original feature that stood out, that effect would become exaggerated in subsequent copies and subsequent recalls. These two effects were called sharpening and levelling. So, just to go over that again, sharpening involves an exaggeration of certain selected features of the original figure, particularly features that stand out. Leveling, on the other hand, is kind of the opposite, and involves weakening of features

that don't particularly stand out. And leveling tends to shift the memory, or your reproduction of the figure, to a more standardized typical figure. And it was found that these changes were progressive, meaning that later reproductions were more exaggerated, or either more sharpened or more leveled, later reproductions than in earlier reproductions from memory.

So these effects got worse over time, basically. Originally, the studies were done with shapes, but later on they progressed to using human faces, and they found very similar results there, so that, for example, subject, were more easily able to associate a name with a caricatured version of a face, that is a face with exaggerated features, like if the original person had a long nose, the caricatured had a really big nose, or if they had messy hair, then the caricature would

have a really insane hair and so on. Carricatured versions of faces were much easier for subjects to recognize than non-characatured versions of the faces, and also easier to recognize than so-called anti-charicatured versions of the original faces. Anti-characatured means that you downplay the distinctive features, that you make them less distinctive, whereas caricatured you make them more distinctive. So that indicates that people's memory of the faces was shifted in the direction of caricaturing the distinctive features of them. Similar effects were observed in another study, which involved estimated distances and remembered

distances. It was found that people were sort of roughly right at sort of remembering the distance between two points. You know, if it was bigger, they remembered it the distances being bigger, or small distance they remembered it as being smaller. But there were many different ways that they could be distracted, or that their memory could be altered or distorted from the actual distance. For example, estimated distances tended to increase with a linear function of the number of intervening points between a start and end position. This was referred to as clutter. So for example, an intervening point could be a turn on the road or a building there or something like that, just any intervening point or milestone or something like that that you reach on the

journey from point A to point B. The more of these intervening points there were, the longer the estimated distance was. So that would indicate that if you travel just along a long desert road for 10 kilometers, your estimated distance for that would be much less than if you traveled 10 kilometers through a large city, for example, lots of buildings and other things passing you by. So related to that, there was also an overestimation of distance for routes that contained lots of bends as opposed to just straight roads. That's very similar to the intervening points or the clutter effect. And there was also a tendency for people's memories of the routes to normalize to 90-degree angles. So that means that there might have

been a slight turn or a very sharp turn or a sort of almost 90 degree angle but not quite right. And in people's minds, like when they were asked to draw maps of where they went or talk through the route or something like that, angles between roads were generally normalized to 90 degree angles, which sort of makes sense because that's the simple way of thinking about things. But it's not just that that's how they described it, is that they actually remembered it that way, too. They remembered roads as being sort of neater and more 90 degrees than they actually were. Another very interesting phenomenon is the perspective effect, which is another example of a spatial distortion of memories.

Subjects who imagine themselves in New York, okay, so you've got subjects in there imagining themselves to be in New York, and then they were asked to judge the distance between New York and Pittsburgh in this particular study. And then in version, or a second trial, they were asked, different subjects were asked to imagine themselves to be in San Francisco, and then to judge the distance between New York and Pittsburgh. it was found that those subjects who imagined themselves to be in New York thought that it or imagined or remembered that the distance between New York and Pittsburgh was greater than those subjects who imagined themselves to be in San Francisco. So clearly when you're in New York,

the distance from New York to Pittsburgh is relatively much greater than is the distance between New York and Pittsburgh compared to the distance between both of those points and San Francisco. So to the people in New York, it sort of looks like Pittsburgh is further away than to the people in San Francisco see New York and Pittsburgh, because San Franciscans are so much further away from New York and Pittsburgh. But, of course, the actual distance between New York and Pittsburgh doesn't change in either case. But people's memory or judgment of the distance did change,

depending on where they imagine themselves to be. So all these effects I've just discussed, sharpening of exaggerated features, leveling of non-distinct features, the clutter effect of intervening points making a route seem longer, and also the perspective effect of judging distances differently depending on where you are. All of these come out of the category of spatial memory distortions. Basically, people are trying to remember spatial information about roots and distances and so on,

or shapes and faces, and these effects demonstrate that people aren't particularly good at getting their details accurate. They have these specific biases that tend to push them in one direction or the other. So let's move on to schema distortions. Schema theory is sort of a psychological theory which holds that what people remember is the result of an interaction between input information, so that's sensory information, and pre-existing schemas. So schemas are sort of just generic knowledge or expectancies or beliefs about how things work, about what exists, about things like that. How you interpret things, how you see the world is working and so on. It's a schema. It's like a background belief or a framework, a structure to understand things. So according to one theory, there are four basic types of schema processes that occur during encoding of memories.

That is when you... So you take in your sensory input, and then that is encoded such that you form memories, and you remember stuff. You don't remember everything that you take in is sensory input, but you remember some stuff. The process of forming those long-term memories is called encoding. And according to this theory, there are four basic types of schema processing errors or biases that can occur. These are selection, abstraction, interpretation, and integration. So let's go through them. Selection. Selection basically means that the information that you choose to remember,

or not even necessarily choose to remember, but just that you do or you tend to remember, is information that is consistent with the schema that you are currently operating with. So if you're expecting to see things or to receive input about a sports game, the information that you pick up and select to remember, or select to encode, will be information relevant. to a sports game. And if something completely different happens, like a person in a monkey seat walks through the picture, if you've seen that video of the people throwing balls around and

the person in the monkey seat walking past, that's a classic example of selection, because when you're watching that video, you're told to count the number of passes between these people throwing balls backwards and forwards. And so that's your schema. Your schemer is ball throwing. That's what you're focused on. That's what is, not just your focus, but it's what you're expecting to see. That's what is your framework for understanding, the video you're watching. You don't expect to see some guy in a griller costume and you're not focusing on that and so it just completely, completely passes you by, you don't notice it, or at least many people don't notice it. That's sort of similar to what's happening in this

selection effect, that you, the information still is, you still see it or hear it, and it still reaches your sensory organs and still sort of enters the brain, but it's not really encoded because it's not connected to the active schema or the background knowledge that is forming the backdrop and helping you to understand whatever it is you're observing or whatever you're doing at the time. So the basic idea is there is that information that is more central to the activated schema will be recalled better than information that is less central. Another example might be if you're in a biology class and the lecturers gives some sort of peripherally related historical background to some of the figures involved or discoveries or

whatever, you're perhaps less likely to remember that if your schema for that lecture is biology. The other stuff just kind of doesn't fit, the historical stuff just doesn't kind of fit within that schema, and so unless it's particularly memorable for some other reason, you are less likely to encode it. So that's selection. Next one is abstraction. Abstraction refers to the fact that people tend to remember basically the gist of messages rather than the specifics or the verbatim content of messages or of experiences. And so that's sort of fairly intuitive if you read it something, an article, a book or whatever, you can generally give the gist of what happened, an overall summary, but you can't remember specific details or passages.

That's fairly common sense. The trouble with this, though, is that when people are asked to remember details about an experience or something they read or whatever, what they try and do is they reconstruct those missing details because the details aren't there. They haven't encoded them because of this abstraction principle. They just encode the gist.

So when they're asked for the details, people tend to just reconstruct them using the general information that they did encode. And particularly, they take... So they take some information that they remember from, say, the book they read, and then they fill in the gaps with information from the schema that they used to interpret or understand that text or that experience or whatever. So if you're at a sports stadium, say, and you're focusing on the event,

and then later someone asks you what the person next to you was wearing, you might engage your sports stadium schema and think, well, what do people normally wear at these sort of events? And then you'll use that information to provide an answer to the question and what the person next to you was wearing. Now, of course, sometimes that can be conscious. You can do that deliberately, but a lot of the time, what we find in this memory research is that people don't realize that they're doing that.

And I'll mention this later on when we come back to the source monitoring failure, but people tend to confuse or have difficulty distinguishing between what's an actual memory, what was encoded from the event, and what is general information that's actually coming from their schema, just sort of general background information that they have, that they're then using to reconstruct specific details. It's very difficult for people to distinguish between the two, particularly if they've sort of discussed the event many times or thought about it lots of times. They might have thought about schema-related information, and that becomes linked to the initial event, and they can't really distinguish the two.

So something that didn't actually happen, but it's consistent with what they expected to happen, can actually be remembered as having happened. And we'll talk more about that later on, but that's the general problem of abstraction, that you only remember the gist and reconstruct the details from the schema that you're using to interpret the gist. Okay, so that's the second one. The third bias here under a schema distortion is interpretation. So interpretation basically follows right on from what I was saying with abstraction. It refers to the making changes or additions to the memory based on schema knowledge. So interpretation and abstraction kind of go together. Interpretation is really the problematic aspect that results from abstraction.

So because you only abstract the gist of a memory, you have to interpret other details. if you need those details later. And you don't always realize that you're doing that. And so that's where the interpretation step comes in. The fourth bias here is integration. Integration is an interesting one because it involves the fact that people integrate their experiences of an event or something that they've experienced into a unified whole. Generally, people don't have just isolated incidents or bits and pieces.

Sometimes they'll do, but that's an uncomfortable experience. We like to have sort of coherent narratives or coherent unified stories to explain one, event or memory, so that we tend to integrate things. In doing that, or as mentioned before, you can integrate things that are merely interpretations with things at actual memories, and so that's problematic. But in doing so, we can also come up with big explanations or narratives, which we've kind of imposed on the memory, and that has been implicated in explaining things like hindsight bias, where everyone, where, if you ask someone about something, how predictable was 9-11, for example, or how predictable was some war before it happened or whatever. People afterwards will

say, yes, it was extremely predictable, I knew that was going to happen all the long, I knew this team was going to win, I knew she was going to do this, or whatever. But before the fact, people don't say that. One potential reason for that is because people are integrating the knowledge that they have, including memories of what they knew beforehand and memories of what they knew afterwards, they're pulling it together in a coherent story, and having trouble differentiating one thing from another and producing therefore the hindsight bias. Okay, so that's some of the biases that occur in memory as a result of a human use of schemers. You might think that it'd be better to avoid schemas if they produce all these problems, but we really can't avoid doing it because we

need to have that background information to bear on experiences or things that we do, because otherwise we wouldn't know what to focus on, we wouldn't know what we're doing, or the point of it is, and we have so much sensory information that we have, we have a very, that we have a available to encode, it's impossible for a story member at all, or we'd go crazy. We have to pick out important parts, and that's what schemas do. They help us to know which parts to pick out. But as said, they have these unfortunate side effects. And so there's just a couple of other studies based on schema effects I'd like to discuss. As mentioned, the trouble with schemer effects is that they can, the way that you view an event, or the way that you sort of interpret it and

bring facts together can affect how you actually remember it. It's hard to distinguish between your interpretation of the event and your actual memory of the event. And so in one study, subjects had to describe a character from a book or some story to a hostile audience, that is an audience who disliked the target character. So the subjects knew that the audience disliked this character, and then the subjects had to describe that character to the audience. And then afterwards, the subjects were asked to rate that character themselves, like how much they liked them or something like that. And it was found that subjects who

described, had to describe the character to a hostile audience, rated the character less positively than subjects who did not have to describe it to a hostile audience. So this is consistent with other studies and theory which show that the communication goal that these subjects

had, they had to communicate about the character to a hostile audience. That goal introduced a bias or a perspective on the initial memory, which is that this character was not such a good person. obviously when you narrate their story to a hostile audience, you're not going to try and make them sound too good,

even if you want to appear unbiased, and presumably that's what the subjects are trying to do, still, you're going to consciously or subconsciously err on the side of not making them appear too good, or at least the fact that this is a hostile audience is going to be in your mind when you're relating the story. And so that sort of becomes built into your schema. Later on, when you're recalling the character

and what they actually did and your impressions of it, that hostility from the audience sort of gets built into your memory of it, and it's difficult to distinguish one from the other, and so there's a tendency for people to then rate the character less positively. So that's sort of an example of an interpretation or an integration effect, whereby what you did with the memory actually affects the way you remember it itself, the actual memory. And so there was another study that was done to, based on this idea, that subjects were asked to write a letter of complaint about one of the two story characters,

and then it was found that subjects who were asked, subjects will then later sort of tested about these characters, and it was found that the characters that you wrote a negative story about, you then tended to remember more, the subjects tended to remember more negative things about that character than other characters that they didn't have to write a letters and complaint about. Another study that I read about involved two groups of subjects both listening to or reading the same story, and then one group had to then narrate that story just as factually as possible, and a second group was asked to narrate the same story, except to try and make it entertaining.

So they were both narrating exactly the same story, just one was asked to do it factually, the other was asked to do it in an entertaining way, but to still keep it factual, just make it entertaining as well. And it was then found, and then afterwards, then both groups of subject were given a test on their memory for the original events in the original story, and it was found that those who had to relate the accurate account

had a better memory for the original events of the story than those who had to relate the entertaining account. Once again, an example of how your schema can affect your memory because the people who had to narrate the entertaining story sort of viewed the whole thing in an entertainment scheme or focused on interesting aspects of it or to exaggerate parts of the story and so on, whereas the people who had to narrate the straight accurately

through a different schema of focusing on the specific details of events and so on. And so that affected their memory-related. later on. Though another interesting part of this study is that neither group was very accurate. Even accuracy participants, who remember were the more accurate of the two groups, recalled only 58% of the story events, and it wasn't a particularly long story either. And I think when they were finally test on that, was only a few days after they'd done the, after they'd read the story initially, and they'd practiced it or sort of recalled it and thought about it several times since then. So even after a few days

in a number of recall periods, they only remembered 58% of the initial story events. not particularly good. Also, there was a significant rate of intrusions, that is, new events that weren't in the original story, from both groups of participants, both the entertaining and the accuracy groups. So that alone, even aside from all the schema effects, that alone doesn't do too much to encourage confidence in the validity of human memory. Okay, but that's enough about schemers. I'm now going to move on to discuss the misinformation effect, which is what a large portion of memory research focuses on. The misinformation effect basically demonstrates that leading questions or inputs of apparent provisions of what is apparently reliable information

after the fact can bias people's memory of what happened before the fact. So for example, one classical study, participants watched films of automobile accidents and then they were asked how fast they thought the cars were going before they crashed, except in different versions of the study or different, the words were changed slightly so that sometimes it asked how fast were the cars going before they contacted each other or how fast were the cars going before they crashed into each other or how fast were the cars going before they smashed into each other and so on. So each of those words connotates sort of a different speed and different severity of the crash. And perhaps as you might expect, when the words, when words like crashed and smashed were

used, people remembered the cars as having been moving faster than they did when words like contacted were used. And there was also people made more claims of having seen broken glass when these high-speed words were used, even though there was no broken glass in the videos at all. And many, many such studies have been done showing that how misleading information or apparent information or leading questions about a memory can lead to biases

or even completely false memories. And it's been found that repeated exposure to such suggestions increases the incidence of false memories or of distortions in memories, and also that the more authoritative the provider of these suggestions or information is, the more likely people are to feel prey to them. The more likely it is that people will fall prey to the suggestions. Interestingly, even if subjects are told directly and specifically that they have been misinformed,

they still persist in claiming that they remember witnessing whatever has been suggested. So, for example, I don't know if it was done in this particular study, but certainly studies like it, basically what would happen is you would ask, was their glass on the road, broken glass on the road from the car crashes, when there was definitely no broken glass on the road. The person never saw broken glass on the road, but you ask them, was there broken glass on the road? A decent portion of people say,

oh yeah, actually, I do remember it was here or there from this car or that car. And then the conductor of the experimenter would say, interestingly, it's been found in certain studies when you set up the experiment a certain way that you can provide participants with some information which will cause them to misremember something. For example, you could ask,

which car did the broken glass come from? or what, on which road was the stop sign, for example, when there was no stop sign and there was no broken glass at all. And then many of them will give you an answer saying where the broken glass was or where the stop sign was or whatever. And then later you say, well, actually, no, you were misinformed. You specifically tell the subject that there was no stop sign, there was no broken glass. The subject will persist that, no, I remember the broken glass, I remember the stop sign, even though you could specifically tell them they were misinformed. A number of studies have shown that.

And so that indicates that people are, in a large portion of these cases, they're not just going along with what the experiment is saying. They actually take that information the experiment provides them and sort of incorporated into their memory, and they actually remember seeing these things that they never did see. A related phenomenon is the so-called social contagion effect, whereby something that one witness remembers, or one person remembers, can contaminate someone else's memory so that the second person then thinks they remembered

what the first person reported seeing. So in one very interesting study, two participants, one of whom was a Confederate, so they were in on what was going on, viewed a series of scenes, and then they engaged in a collaborative recall task where they had to recall or remember as many items from the scenes as they could, and they did that together. However, some of the items that were recalled by the Confederate were false, so they deliberately provided false items that they said had occurred in the scene. Then what happened is the actual participant in the experiment were then later tested individually to recall as much as they could, and they were told to only recall the items that they had remembered from the scenes, not the items that the Confederate had recalled from the scenes. Of course, the participant didn't know they were Confederate,

but only the ones that they had remembered. And there was a couple of interesting results from that. Well, first, that many participants recalled things that the other person had seen and that they had not originally recalled, but thought that they had. But even more interestingly, they recalled a number of the false events that never occurred in the scene at all, that had just been provided by the confederate during the collaborative recall part of the task. And even when they were specifically informed, once again,

even when the people were specifically informed that, well, some of the other participant's responses might have been wrong or were wrong, they still remembered those things, those incorrect items as having been in the original scenes. They remembered it being there. They remembered seeing it. They could describe details about it. And where was this coming from? Well, it was presumably coming from some other memory that was getting confused with what they had actually seen.

The fact that the other subject had mentioned this incorrect item had been incorporated into the memory, but they didn't remember that it was the other person who said that. They attached the memory to the original scenes that they had seen. All right, so that's the misinformation effect. It's a recap, it's been found that when you provide information, true or false, about something that the person witnessed, and then later test them on it, they'll often incorporate that false information into the original memory and think they remembered things from that, in relation to that false information, even when they

definitely did not. And this is related to, this misinformation effect is related to what's called source monitoring failure, which is the next thing I'm going to talk about, which basically means that you remember something, but you can't remember where it came from. So you fail to monitor the source of a memory. And this has been very well demonstrated that people can't remember stuff, but just hopeless remembering where it came from. And that can be really problematic because you can, it's found that there's basically no difference between the neural activity that occurs when someone, for example, see something and when they just imagine something, or when they remember something that they actually happen, versus when they remember something that they

just actually, that they only imagined. So it's very difficult for people to distinguish these things. So some examples of this. When Ronald Reagan was president, he, um, he publicly told a story of a great act of terrorism by a US pilot, I think in World War II or the Korean War or something like that it was supposed to be. But reporters couldn't find any record of it until someone found that the scene that he had described bore a very close resemblance to a scene from a 1940s war movie called A Wing and a Prayer. And so it's thought that Ronald Reagan probably remembered this scene from a movie, but failed to remember where he'd remembered it from and thought that it was a real event that someone had told him about. There are a number of interesting cases of unconscious

plagiarism, whereby, for example, Helen Keller, there was the famous deaf mute, wrote a short story and then later discovered that it was actually very similar to one that was read to her years before, but she hadn't remembered it. George Harrison became involved in a famous lawsuit about his song, My Sweet Lord, which used a line from a song called, He's So Fine, and George Harrison claimed that he didn't remember hearing it from that song. He just thought of it himself, but it was, once again, thought that he had unconsciously included the line, so he'd remembered it.

The line was in his memory, but he didn't remember that it actually had come from another song. He thought he just thought it up himself or whatever. Of course, lack of intention is not a defense against copyright infringement, so he still lost that lawsuit, unfortunately, for him. Another case I read about a famous historian whose book contained a large number of plagiarized sections, which the author said was that they had confused the handwritten notes that they had written with, so handwritten notes that the author had actually composed with passages that they copied from their sources. You'd think people would be able to recognize their own wording from someone else, but this is another example of source monitoring

failure that the person writes down the words and remembers that the words are familiar, but then thinks that they compose them themselves and forgets that they actually read them somewhere else. And it's been found that certain things can make it much more likely that you're going to forget the original source of a memory and confuse it with something else. So the more times you imagine an event or a memory, or think about it, the more likely it is that you'll confuse your imaginations or remembering from the actual event itself. Repeated questioning or rehearsal of the details of an event also can tend to confuse, lead to source monitoring failure. So the way you confuse questions

of people asked or things that you thought about or things that you said with what actually happened originally. Encouraging people to relate an event to their particular lives or other people's lives also tends to increase source monitoring failure, presumably because once again you're thinking about it, you're making neural connections and things like that between

different memories and so it becomes more difficult to distinguish between what your thoughts are and what your imaginations are versus what the actual original memory was. And the really troubling aspect of that is that repeated questions questioning, repeated imagination, and even presentation of false evidence are all methods that are commonly used in police interrogations. And they're also all great ways, as the studies I've been discussing, have shown, of inducing false or biased memories.

And so that has very important implications for eyewitness memory, which I'll discuss more a little bit later. And of course, this source monitoring failure also has implications. For example, it's very easy to remember something that really did happen, but then things. think that it occurred in a different place or a different time. You take one bit of a genuine memory and tack it onto another and confuse where it came from. Or another classic one is that you remember someone said something but you forget who it was and then you apply it to someone else who didn't actually say it. You remember the information but you can't remember the source. All right, so that's source monitoring failure and now I'm going to move on to a very interesting

aspect of the research which are false memories. So far I've mostly been talking about memory biases, so spatial memory biases and distortions, distortions caused by schemers, misinformation effects causing people to distort the original memories, and then, of course, source monitoring failure. But all of those are basically true memories, or memories are originally true, that are being kind of warped or falsified or biased. False memories are just complete confabulations, or maybe there's a little bits of true memories that are taken from here and there, but they're combined in a completely new way, so that the person thinks that they did something or that they remembered something that never happened at all. And that would be kind of the ultimate in memory failure, when you're just remembering things that are completely. wrong. And you might think that, okay, we can have these various subtle biases here and there,

but we're not going to just think that ridiculous things happen that never happened at all. Well, the evidence shows in a number of psychological studies, well, many psychological studies, in fact, that that can happen. False memories are not that hard to produce, even in normal, healthy adults. All right, so let me talk about some of these studies. In one study, both college students and children were shown short excerpts from a movie depicting a boy's experience at a summer camp. Students were then asked questions about the film afterwards, and they were told that they had to answer every question, even if they just had to guess, even if they didn't know the answer,

they just had to make a guess. Then a week afterwards, so a week after watching the film and being questioned about it, the students were then asked whether these various different items that were listed off were contained in the film or not. And the results showed that all age groups were more likely, about twice as likely, in fact, to misattribute items that they had generated as false responses

than they were to misattribute control items, which were just random things thrown in there by the experimenters. So what that means is that if the person... So remember, when the subject was originally answering questions, they sometimes had to fabricate answers or just guess answers because they weren't allowed to say, I don't know. But then a week later, they were twice as likely to... If that item that they were being asked about

was something that they had mentioned, or they had confabulated or they had just made up or guessed a week before, they were twice as likely to think it was in the original film than if it was just some random item. So if originally they had guessed, yeah, there was a shoe in the film. Then a week later, they were twice as likely to say, yes, there was a shoe in the film, than a chair or something else that they hadn't guessed.

So here, a false memory is being formed about what had appeared in the film. Basically, it seems that people are confusing their imaginations or their original answers with an actual memory of what was in the film. And what's most interesting about that is the effect stayed basically the same, even when participants were told that the experimenter from the previous week had made a mistake and had asked them about some things that were not actually in the film. So it's like in the original week, the experiment asks you, maybe not, is there a shoe in the film, but some question about the shoe in the film?

And then you make up some answer, implying that there was a shoe in the film. Then a week later, another experiment asks you, was there a shoe in the film? And you say, then you remember something about a shoe and say yes. Then the experimenter says, well, actually, the first experiment made a mistake, and they asked you about some incorrect, some things that weren't actually in the film. And then they ask you again, so was there a shoe in the film? And you say, well, yeah, there was. Okay, the experimenter might have made a mistake about something,

but it wasn't about the shoe because I remember the shoe, even though there was no shoe, and the subject is confusing their memory with their answer or the thoughts that they were going through their mind when they were giving their original answer. Even more interesting are the number of different studies that have asked people to imagine fictitious childhood experiences and then have actually managed to transform those imaginations into real memories.

So, for example, in one study, they encourage people to remember an event or different events from their childhood, which had never happened. They didn't encourage them to remember it. They just encouraged them to imagine as if it had happened. And then they asked them to think about it and imagine it over a course of a couple of sessions over the course of a number of weeks. And by the end of it, a large portion of participants were claiming to remember some of these

events that they initially had only begun, had only just imagined as if they were happening. So just by telling people to think about something and imagine it didn't happen, they were able to convince some subjects that, in fact, that event had happened in their childhood. Another very famous experiment used the so-called lost-in-a-l-mall technique to implant false memories. Basically, they obtained some details, and often particularly photographs, were particularly useful in helping to implant false memories, of people when they were, say, age five or six or something, and the experimenters led the subjects to believe that they had obtained descriptions of past events

from their family members, from the subjects family members. So some of these events were real and some of them were made up. And then what they basically did was they asked the subject about each of these events to describe them and remember them and say what happened. And remember, some of these events were true and some were false. And the particular one that's most famous is that they would ask people, oh, do you remember when you were five or six and you were lost in shopping more and got really upset and then you were ultimately rescued by an elderly person?

And they particularly made sure that this had not happened to any of the subjects, this and in similar studies which did similar things. So they knew that none of these, and they talked to the subject's family members and so on to ensure that it did not happen, although there was no evidence that it did. None of these people had had that experience. And yet something like 25% of the participants

were led to believe wholly or partially that they had, in fact, had that experience. And many of them were able to provide intricate details and explanations about the events and sensory details and how they felt and so on, even though the event never actually happened. And just to make sure that this effect was, Okay, well, being lost in them all's not something that's so unreasonable.

So they tried to replicate it with things that were increasingly unlikely or even later on impossible, and they still got the same effect. They still got around 25 to 30% of subjects to recall, often in considerable detail, an event that never happened. So examples of events they could convince people that they... Remember, this was like 20 to 30% of subjects they could convince having been engaged in having had this event happened to them. These included being hospitalized overnight, having an accident at a family wedding, having been nearly drowned and then rescued by a lifeguard, having been victims of a vicious animal attack, and probably in, what's my favourite, having met Bugs Bunny at Disneyland.

Now, meeting Bugs Bunny at Disneyland is impossible because Bugs Bunny is a Warner Bros. character, and so it would never be at Disneyland. And yet, once again, about one third of subjects were able to report memories that they had met Bugs Bunny at Disneyland. And particular, once again, this rate was achieved after multiple suggestions and exposure to things like fake advertisements of showing Bugs Bunny at Disneyland. So once again, the fake advertisements, you might say, well, if they're a fake advertiser, if they're showing fake advertisements to the subjects of Bugs Bunny at Disneyland, that's

kind of giving misleading information. So maybe it's not so unreasonable that these people think that Bugs Bunny could be at Disneyland. But remember, they weren't just asking people, is Bugs Bunny at Disneyland? They're asking them to remember having met them at Disneyland. And so providing fake ads, if memory is reliable, shouldn't make any difference to that. All the fake ads do is make it seem somewhat. plausible that this memory could be true. And then that's all the subjects needed then to,

a bit of plausibility and suggestions and thinking about it and repetition in order to come up with the false memory. And even more interesting than just the proportions is the details that many of these people came up with to explain what happened in their false memories. So for example, of people who recalled a personal encounter with Bugs Bunny at Disneyland, 62% said they shook his hand. Forty-six percent remembered hugging him, and many others remembered touching his ear or tail and hearing him speak, what's up, dark? So people had all of these specific details about a false memory that never happened, and in this particular case, the experimenters knew never could have happened was impossible. And this paradigm has been sort of tried out in many

different circumstances of different events. I believe it's even been done with things like alien abductions, and they can implant memories about that as well, just with very, very unlikely or even impossible events. And once again, warning people about the fact that they were being potentially or actually exposed to misleading information doesn't really reduce this effect at all. They still claim to have remembered these events. And in fact, when the studies are over and the experimenters say, well, actually this is false, it never happened, people continue to, many people at least who had the false memory, continue to insist that no, it did happen, I remember it, I remember it. And it's very difficult to convince them that it didn't actually happen. So these false memories

that people are coming out with, they're not just sort of vague recollections or minor things, sort of saying, oh yeah, might have happened. They really, a large portion of them, really remember it and really convinced that it actually happened to them. Particularly once again, when you provide them with specific details, photographic evidence, stuff like that, even if it's falsified, they'll still remember something that never happened. And the more times you ask them about it, the more times you ask them to think about it and try and remember it happening, the more likely it is that they'll come up with something. Perhaps the most interesting study of all. I think I said that about the Bugs Bunny one, but this is maybe even more interesting.

It's a study where a person was presented with false eyewitness evidence that they had committed a crime. Well, it wasn't really a crime in this case. It was just damaging a computer console during a study. But the person was basically accused of having done this, and of course none of them did, so they all said no. But then the experimenters came up with a false witness who said that they had seen them do it and then gave some details about it. And once that had happened, some reasonably large proportion, I think of what exactly was, of the participants in the study, then sort of confessed, and they said, oh yeah, I do remember having damaged the keyboard, even though they never did, just because they had someone else there that was saying that they

had seen them do it, that once again seemed to trigger some sort of memory to that they had done it. That has very important implications for false confessions, which have been implicated in being involved in many false conviction cases. Often we only know about these when DNA evidence, later becomes available and it's found definitively that the person who was jailed for 10 years wasn't, didn't not actually commit the crime. And so we've now transitioned into eyewitness memory and I'm just going to talk a little bit about that. Police lineups to identify a subject are particularly egregious really. They're just horribly, horribly biased way of getting, of identifying a suspect because it's found that people, what they tend to do is identify the

person from the lineup who most resembles their, what they remember about the subject. And It's even found that, so you'll have a lineup that has the correct suspect, that has the actual perpetrator there, and then the subject often identifies the correct perpetrator, okay. But then you remove the correct perpetrator from the lineup and do it again. Well, you don't literally do it for the same study, but you do comparisons of when the correct perpetrator is there and when they're not there. And basically, it seems that if you just remove the true perpetrator,

people would just pick someone else from the lineup, the person that most closely resembles the actual perpetrator. And also, there's a number of meta-analysis which have been done which show that there's very little, if any, correlation between the confidence that someone expresses in their witness identification or in their criminal identification,

in the lineup or something, and whether they're actually right or not. So some will say, yep, I'm certain it was him, and they're probably just 50-50 of being right or something, or no better than chance of being right, regardless of how sure they say they are. A very interesting example of this, and there are many, as I said,

but this is a particularly well-known one, is Jennifer Thompson's case. She was a college student who was raped in the 1980s, and she, throughout the rape, she later said that she studied the perpetrator's face in great detail. She was determined to memorize every detail of their face, scars, particular features and so on, so that she'd be able to recognize them later and ensure that they were convicted. And so then later on, she was taken to a lineup and asked to identify her attacker, and she identified her attacker and said, yep, that's definitely him, I'm certain of it.

I studied his faith in detail. I remember that's him. He was the one. And then she testified against him at the trial, and she was positive it was him. No doubt. She wanted him to rot in jail and even wanted him to have the death sentence,

if it all possible. He was sentenced to prison. But later on, they're a bit unsure because after he'd been in jail for like a year or something, they heard about another guy who was also in the same prison, who was bragging about having committed the rape and having gotten away with it. And so they did.

some kind of retrial, and they brought this, the guy who had been claiming to have committed the crime, they brought him into the courtroom and said, asked Ms. Thompson if she recognized him, and he said, no, I've never seen this guy before in my life. The other guy did it. It was definitely him. I don't know who this guy is. Forget about him. And so, uh, the original decision was upheld. And about 10 years later, new DNA evidence became available. And lo and behold, it turned out that the person she had recognized as definitely being her attacker was definitely not her attacker. And the person who had been bragging, and the person who had been bragging about having committed the crime, did in fact commit the crime, it was him.

And she had seen him in the courtroom and said it was definitely not him, I've never seen him before. Even though she had studied his face during the incident and was certain that she had made the correct identification, but she was wrong. Now, of course, this is only one case. It doesn't prove the general point, but it's illustrative of the general point, which is proven in many other studies and matter analysis and things like that that have found this happens again and again, it's very easy to get people to misidentify, well, people, and to confuse memories and things like that. So police lineups are particularly unreliable. Now, the last thing that I want

to talk about are flashbulb memories. This is, I left this to last, because in a sense, it's the most interesting and most chilling experimental finding about the reliability of memory. Flashbulb memories, for those who don't know us, are supposedly, there's a bit of debate about whether they really exist, but the idea is that they're particularly vivid memories that you, that you have, particularly visually vivid, about a specific event, usually some kind of catastrophe or a very newsworthy event, and you remember exactly where you were at the time and place and everything else, exactly how you found out about said event. Classy examples are like the explosions of both of the space shuttles, assassination of John F. Kennedy, 9-11 attacks, stuff like

that. Death of famous celebrities is another one. A number of studies have been done to see, well, are these flashball memories as accurate and as people are, as people are, as people think they are, given how vivid the memories often are. So, in one study, this study did the O.J. Simpson trial, back in, I think, the 90s this was, but studies have been done with other events as well. But this particular study related to the O.J. Simpson trial verdict, and they contacted a number of subjects, a few dozen, just, I think, a few days after the original verdict, and then they contacted them, and then asked them questions about where they were and what they were doing

when they'd found out about the verdict and whether they agreed about it and what they thought about it and things like that. Then they contacted the same people once again, or as many of them as they could, 15 months afterwards and 32 months afterwards, which is two and a half years afterwards. So basically a year and a bit, and then two and a half years after the original verdict was announced, and then asked them to complete the same survey again about how they found out about it, what they remember about it, and so on, what they felt about it, and how confident they were in these memories. Now, the very interesting finding, only 10% of subjects had major distortions in their memories after 15 months. By the way, the way they determined,

and how many distortions or how big the distortions were as they got a couple of independent coders who code announces from like no distortion to extreme distortion and then they compared them for reliability and so on and so the methods they used are fairly well established in these sorts of studies and are pretty reliable anyway only 10% of subjects had major distortions in their memories

of what happened and where they were after 15 months so that's not very many but after 32 months so only one year and a half more on So after about three years as opposed to one year, 40% of subjects had major distortions in their memories, which means basically they told completely different stories about where they were and what they were doing. Instead of I was driving home,

it was I was at home watching television with my parents or whatever. Completely different stories. Interestingly, those who had low accuracy scores or large distortions, the ways that these low accuracy scores came up in the two different timeframes was very different. So in the initial 15-month follow-up, most of the people who had low accuracy scores did so because they said they couldn't remember. And so once you took those out, everyone who did think they remembered was pretty accurate. And that was still 15 months after the original event.

However, two and a half years after the event, so that's 32 months later, more than five times as many of the low accuracy scores were that low because of major distortions as opposed to failing to remember. So basically that initially, most people remembered correctly what they were doing. like 90% of people correctly remembered where they were and how they felt, and most of those who didn't just couldn't remember. Whereas in the 32-month period, only around 60% of people could remember reasonably well, although even many of them had minor distortions, but only around 60% of people could remember reasonably well

what they were doing, and most of those who couldn't said something completely wrong. They did not say, I can't remember. They came up with something different and reported that very confidently often as being their memory of what happened. And the overall rates of not... remembering actually declined from 20% to 6% in from the 15 to 32 month period. So basically

lots of people or a significant number of people said, I don't know or I don't remember to different questions in the 15 month follow-up, but virtually no one did in the 32-month follow-up. And also, after 32 months, there was almost no correlation between a subject's accuracy in their memories. That is comparing what they say to what they said in the one that was in the survey they felt they filled that after only a few days. virtually no correlation between that and the confidence that they expressed in their answer. So the basic takeaway from that study is that after 32 months, so two to three years after an event, regardless even if it was a particularly notable flashbulb event-like, as this example was,

many people, a significant minority of people, perhaps about a third, have major distortions in what they think they remember, even though they report being very certain or very confident in those memories. And once again, as with previous studies, people who had major distortions reported significant details and intricate aspects about their experiences which were known to be wrong because they were different to what they originally said. In addition to the, say, around 30% who had major distortions, about another 30% had minor distortions. So only a minority of people after 32 months were able to remember to a high degree of accuracy where they'd actually been. And that's with a flashbulb memory. You compare that to what an ordinary memory would be like and it would probably be, if you were able to, if you were able to, and less accurate. By the way, it's also disturbing that almost no one remembered having filled

out the original survey two and a half years previously, which is somewhat interesting. Okay, so some conclusions from all this. I've discussed distortions in spatial memory, distortions that result in, that result from our application of schemas to understanding memories and to put them into context. The misinformation affects how provision of misleading or false information can distort and bias our memory, how we find it difficult to to remember or distinguish the source of a memory, including whether that was a dream or an imagination or an actual true memory. That's source monitoring failure. How it's very easy to implant false memories, even very rich, detailed false memories, even about events that are impossible to have

actually happened, like Bugs Bunny appearing at Disneyland, how shockingly inaccurate eyewitness memory can be, particularly police lineups, and how flashbulb memories as a coup de grace de Gras example are not as reliable as many people think they are. Also, just because a memory report is expressed with confidence, with lots of rich detail and lots of emotion does not necessarily mean that it's true. And so I just want to conclude with a quote, a very nice quote from Elizabeth Loftus, who is a very famous researcher into the reliability of memory, particularly focusing on eyewitness memory, and she has said, quote, it has been said that we are the sum of our memories, but after three decades of research on memory in general and memory distortion in particular,

it makes sense to consider the reverse of this statement. Memories are also the sum of what people have thought, what they have been told, and what they believe. Thus, our memories are shaped by who we are and what we have been led to believe. We seem to reinvent our memories, and in doing so, we become the person of our own imagination. End quote. So that's all for this podcast. It's been quite a bit longer than I had expected, but lots of interesting research to discuss. I would appreciate any feedback you might have.

My email address is Fods12 at gmail.com, F-O-D-S-1-2 at gmail.com. Thanks for listening, and I'll talk to you next time.